“From your site and the other information I’ve found on the net, I think we should scan our photo collections in TIFF, at 600 dpi, using your naming convention / workflow. You don’t cover TIFF versus other formats in your articles, but I see you are using that format and there seems to be general acceptance that it is the best format for archiving.
What do you think of the PNG format? “
Wellington, New Zealand
Peter, that is a great question. And you’re right, up until now I have not covered what I feel is the best file format(s) to save scanned photos with. But, as you astutely noticed, I did sort of allude to my personal choice in a couple of my posts. Especially in some of my images I used in my 3-part “naming convention” series you brought up called What Everyone Ought to Know When Naming Your Scanned Photos.
I think your question actually deserves a slightly more complex answer than I could normally get away with. Had you simply asked, “Which do you prefer for scanning photos, the TIFF or PNG format?”, I would feel comfortable quickly answering you that in my humble opinion, the TIFF format is by far more superior for the purpose of scanning photos. But, since you brought up your interest in “archiving” your photographs, I want to make sure I elaborate a bit more to explain why our personal goals of scanning need to be considered when making the final decision which file format to save our master image files.
PNG vs. TIFF Files – So Which IS Better?
PNG files (Portable Network Graphic) were invented to replace the GIF format (Graphics Interchange Format) that was quite popular at the time for images on the internet. They really weren’t intended to be used for professional-quality photos. Currently on the internet, you will primarily see JPEG and PNG files followed by the lingering GIFS.
PNG files:
- can only be saved compressed (“lossless” – reduces file size)
- can hold an alpha channel
I use PNG files here on the SYEL website a good portion of the time because they have small files sizes and I love being able to use the alpha channel. This is a fancy way of describing a “compositing” process where the image can have a transparent background so you can lay it on top of another image and use it as the first image’s background. It’s very cool! And in case you’re wondering, no, JPEG files don’t do alpha channels.
TIFF files (Tagged Image File Format) on the other hand:
- can be opened with almost every image program (it’s an extremely common format)
- can be saved compressed or uncompressed
- can store “layers” within (great for use with Adobe’s high-end Photoshop for example)
- can hold all color, color depths and color groups (like RGB and CMYK)
- can save 16-bits per channel scans (your 48-bit scanner setting)
- can store IPTC metadata (captions etc.)
In my opinion, one of the biggest reasons not to use the PNG format for photo scanning – and more importantly archiving photos – is because it’s a form of compression. It’s a form of compression called “lossless,” which as you might already know, means that your photo will be saved without having any of the data (details) of your photo “tossed out.” However, the image is still undergoing a very complex process using a mathematical algorithm to compress the file so that it will take up less space on your hard drive.
So let me explain why this should be important to you.
Do You Have Basic or Advanced Goals?
To keep it simple, if you are about to scan your entire photo collection, you are probably going to fall into one of the following two categories. First, you might be the type of person that just wants to get a good-looking scan of all of your photographs as quickly and easily as possible. If you feel this might describe your needs, I like to say that you are someone with basic goals.
Now don’t take that the wrong way. I don’t mean to say basic in a negative way. The fact that you are even wanting to scan your photos already says a lot about how much you care about your collection.
When Peter brought up archiving, to me it implies he’s stepping into the shoes of someone in the second category. Someone with advanced goals is a person I feel is invested from the very beginning to make the decisions that will insure everything they do throughout the workflow of scanning, correcting and managing their photos, will preserve every last bit of image quality (detail) they reasonably can. Those with advanced goals also are willing to sacrifice more of their time and possibly money – which in many cases is not that much more – for the ability to create and store these higher quality images.
Let me quickly add, having advanced goals is by no means exclusive to those seeking the level of perfection a lifelong professional in the graphic arts field would produce. You don’t need to be an expert to have advanced goals. We all have our own personal limitations and expectations for our collection.
Settings panel when saving a TIFF file in Photoshop CS5. (Mac) Notice there are three compression options or none
What I Really Think of Compression
I believe in my heart, that if you are going to really go at archiving your entire photo collection – I mean really give it your all – you should scan them with adequately high DPI settings, save them “raw” without lots of filters and computations altering the outcome of the master image during the scan, and save and archive them uncompressed. I know this might be asking a lot of you, but the sacrifice has its rewards.
If you are considering compressing your images, to save hard drive space or ease the processing load on your current computer, just know you are taking steps toward the side of people with more basic goals. I just want to make sure you understand this.
Many may disagree with me, but I consider compression, even the lossless kind, to be an act of compromise to the overall integrity of your collection.
The hugely popular JPEG file format (Joint Photographic Experts Group) also uses compression. And it’s even more severe because it’s a form of lossy compression. Lossy means your file size will typically be even smaller than that of lossless, but at the cost of reduced image quality. And once the detail is gone – it’s gone. If you’re lucky, your eye won’t be able to see what’s now missing.
Even worse, every time you re-save a JPEG as a JPEG, you are “tossing out” even more information in addition to the information you got rid of before. It’s compounding detail loss!
Non-destructive image managers will however automatically protect you from this additional image loss when you make edits to your photos inside. This is why I highly recommend everyone start using one. (Check out Use 1 of These 4 Photo Managers If You Care About Your Photo Collection)
Additionally, and I know this is really more theoretical, the second you bring to the table the option of compressing your masters, you introduce a greater possibility for file corruption. You’re putting more trust in that mathematical formula to “destruct” your photo now and “reconstruct” it later than you may want to give it. Some formats have so many variant types of compression, who is to say that “shareware” graphics program you found one day off the internet to compress your images will be around 20 years from now? What if their proprietary scheme doesn’t open correctly in the version of Photoshop you buy in the future? Then what are you going to do?
The Best Format to Save Your Scanned Photos
You are the safest saving your master images in the most common, nonproprietary format that you possible can. And if you are swinging towards the advanced side of goals for your collection, you want to save your files uncompressed. Yes, uncompressed files are huge, but so are the capacities of average priced hard drives on store shelves today.
The two most common formats that I feel safe recommending to you are JPEG (high-quality setting) for those with basic goals and TIFF (uncompressed) for those with advanced goals. It’s also the two formats all of the major scanning services feel safe saving your precious memories in as well.
Wow! And I bet Peter thought this would be a simple answer to a simple question! Leave it to me – just get me all wound up and it all comes out!
I hope that all made sense. If I left out anything you are wondering about, just ask me in the comments below. I will be glad to answer them for you!
Thanks for that question Peter from over there in beautiful Wellington, New Zealand! And if you have a question of your own you would like me to cover in an article like this one, ask away using my contact form. Cheers!
nah.. Too much text…
Still would use JPG!
This is one of the most informative articles and blogs I have ever read on this subject. I was pretty well convinced that I will use Tiff as my output but there are a couple of you who suggested that if I have VueScan Pro, I should use dng + jpeg. No one really commented further about that so I am a bit confused. I do have that program and I use Lightroom. I have all my Nikon digital pictures in Lightroom as NEF (Nikon raw)files. LR does allow you to import your files as dng but I did not do that. My intention is to import all of my scanned files into Lightroom as Tiff’s since I have not yet used the dng capability but the contributors to this blog aroused my curiosity. Can anyone comment further, either positively or negatively, regarding the dng + jpeg method?
Hi,Great Article. Thank you so much for sharing!
Cheers!
My choice for archival would be PNG because the format is widely supported and decompression is fast. For work in progress I would use either TIFF or BMP without compression, or PSD.
One reason why PNG is not a good choice for direct output from scanning software is that compression to this format is rather slow. This becomes significant when working with large files. So it makes no sense to spend time compressing a scan that still needs editing (inluding simple edits such as cropping or rotation), especially if compression only reduces the size to around 50% of the original. If the scanner does encoding on the fly without a temporary uncompressed file, and the CPU can’t keep up, then the scanning head would be requird to stop, and possibly create a tear in the image.
If you expect data coruption, then an uncompressed format has an advantage. As long as the file size doesn’t change and unreadable sectors are replaced with other data instead of omitted, then corruption will only affect the pixels that could not be read. A compressed file will be corrupted by a 1 byte error until the end. Recovery might be possible for some formats, but it is not trivial.
JPEG-LS is an interesting format with faster compression and better ratio than PNG, comparable decompression speed, but unfortunately it doesn’t support metadata.
Interesting article. Thank you for publishing it. I am confused, however, about the compression vs. non-compression issue. To get right to it, I’m not sure why using compression makes a person a basic need person, or why that’s relevant.
From my perspective, using more disk space than one needs to doesn’t make sense. If you can store the exact same file for less, than it makes sense to do so, regardless of whether you’re a basic person, advanced person, or an upside-down pineapple cake. Just using more space because you have “advanced” needs, without a benefit for that increased space usage, all to retain a mental image of oneself as an advanced user, doesn’t make sense. This is like saying “basic people pay $1.99 for a taco, but advanced people should pay $3.99 for the exact same taco”.
Well, why?
I guess I’m just not getting it.
Interesting article, though.
Hi Rex,
I believe why Curtis refers to some people as ‘basic needs’ people is that they need only to have an acceptable quality of digital image. They’re generally not concerned about preserving the maximum possible file data and quality that other people, who may want to do so in case they want to edit their images for particular needs in the future. So far as compression goes, if it’s LOSSLESS compression, then no data or quality is lost to the compression operation. On the other hand, LOSSY compression does discard some data and quality. The greater the amount of compression, the greater the loss of data and quality. While the TIF/TIFF format is generally LOSSLESS, it can be set to use JPG/JPEG LOSSY compression, at least in some programs. If the same image is scanned as a TIF, then again as a JPG, the two files will NOT be identical since the JPG format always applies at least some compression. Thus, you don’t have the “exact same file for less” with the JPG instead of the TIF version. For people who just want a digital copy of an original photo, particularly if all they want to do with the digital image is email it or post it on a blog or web site, the JPG version, which will be smaller than the TIF version, will most likely meet their needs, regardless of the amount of compression applied. Becaused the JPG format is LOSSY, each time a JPG file is opened, edited in any way, then re-saved, it is compressed again so repeated saves continue to reduce the amount of data in the file and the file’s quality. TIF and PNG formats, both using LOSSLESS compression, if any, normally give larger file sizes but don’t lose any image quality. For those who want to retain the maximum possible quality and/or do extensive editing, possibly intending to eventually print their images, TIF and PNG are better formats to use. Such files can always be converted to JPG after editing and, if necessary, resizing, for use in email or posting online. Hopefully this will clarify things for you.
Ah, okay, so he’s talking about lossy compression then? He doesn’t have an objection to flipping on LZW or ZIP compression when saving the TIFF?
If so, that makes complete sense.
For some reason, I thought he was saying “Don’t use LZW or ZIP compression on TIFF files if you’re an advanced user”, and that’s what I didn’t understand. I thought there was some philosophical objection to any compression for “advanced” users.
It sounds like my interpretation was incorrect.
Thanks for the help!
What about tags in the files? Is anybody using this?
In windows explorer there is only one field for PNG and that is ‘date taken’. This cannot contain ‘x’.
For TIFF there are a whole lot of fields to use. Any suggestions on which fields to use for extended filenames?
[date] – [event] – [location] – [people] – [scanner settings] – [numbering system].tiff
The field ‘tags’ seems useful for searchable tags for example all of the persons in the photo, location etc. This is searchable in windows explorer. Can these fields be read by a photo managing software?
I’ve bought a Canon CanoScan 9000F Mark II for my retired mother to scan our photo collection. She hasn’t started yet as I’m still doing research.
I’ve really learned a lot from this site. Thank you!
Glad I found your website. I am so uninformed on this I only understood about a 10th of the discussion. But, I feel as if I have found a source that with study will help me to become more informed before I begin to scan my 30 years of photos of Switzerland. I am eager to start the scanning, but first, the learning. Thanks!
No worries at all Barbara — knowing 10% of the conversation is a really good starting point! This stuff is hard at first. And you certainly don’t need to understand a 100% to get started on your collection.
But, I’m certainly with you if you decide you don’t want to start scanning and organizing until you’ve got a pretty firm grasp on a lot of it. It’s not a good feeling worrying that you may have to redo everything you’ve already done.
So, here’s to the learning… and we’re (me and everyone who visits this website regularly) behind you when you’re ready to finally jump in and get started.
Cheers!
Hello there, You’ve done an incredible job. I’ll certainly digg it and personally recommend to my friends. I am confident they’ll be benefited from this website.
Nice article..photocopiers are very important for office promotion..thanks for sharing
Nice answers in return of this query with real arguments and describing everything regarding that.
What? Lossless compression is lossless. It is mathematically identical to the original, and just happens to waste less space.
Hi Curtis,
Thank you so much for all your work on this blog. It’s been incredibly helpful; I am just beginning to archive my family’s photo collections, and have found myself referring back to your site multiple times a day for weeks!
I finally decided to invest in the Epson V550, which is now set up. I’ve decided to use the .tif format (though I am curious about the .dng format and would love to learn more about the pros and cons). From what I’ve read, it seems you use the Epson Scan software, which is what I’m planning on using as well. Right now I’m in the File Save Settings within the program, but I see there are two ‘.tif’ options:
1. TIFF (*.tif)
2. PRINT Image Matching II (TIFF) (*.tif)
I looked up the difference in the user manual, and (if I am understanding it correctly) it says the latter will provide a higher quality scan for printing. First and foremost, my goal is to archive, but I’m sure at some point down the line I will want to print hard copies of some of the photos. That being the case, should I use the latter? Is there any reason to choose the the first option over the second? Any advice and information you can provide is appreciated!
And again – thank you so much for sharing your experience; it’s been invaluable!
I have started scanning my photos using tiff and the ‘extended’ filenames and I’m happy with how it’s going. I do have a question regarding my digital photos though – it seems they are automatically jpg files of about 1-2 mb, obviously I can’t scan them as tiffs because they are digital, not hard copies (unless I have the last 15 years of photos printed!) so I guess they are already compressed and there is not much I can do with them? I have mostly used a point and shoot camera.
Hi Debbie. There’s certainly nothing wrong with .JPG masters taken with your point and shoot camera. Most people shoot and save their masters in this format with this type of camera, so you are not alone. If there’s a setting deep within your camera settings to adjust the quality that it saves them as, then this would well be worth taking advantage of. Many cameras ship with the setting set to the middle — like 5 quality out of 10 — in order to preserver space on the internal memory card. But, bumping that up to or near 10 will improve the quality often by a lot!
So, this being said, if you keep your masters saved in the .JPG format, just be careful when using an image manager with them. Opening/viewing them will not hurt them, but opening and then re-saving them as .JPG files over time will degrade the image quality (even more) each time it re-comresses the .JPG file. So, since quality is important to you, if you aren’t using a non-destructive image manager, and you re-save your edits as a new .JPG file (or to itself, replacing your original image — which I wouldn’t recommend if you want to preserve your original as it was shot), then make sure your image manager is set to save at the highest quality .JPG that you can to reduce the amount of quality loss.
Curtis, the Typing Gremlin got you in the second line of your last paragraph in response to Debbie. It Should read: “Opening/viewing them will NOT hurt them, but opening …” The Gremlin left out the word “not” for you. The rest of your reply is correct. An option would be to save the original, master JPGs as PNGs. That won’t increase any quality, but edits made to the PNGs won’t lose quality with each subsequent Save like they would to edited JPGs. If a particular shot needs extensive edits, adjust exposure, adjust contrast, adjust color balance, cloning or other retouching of several spots, with a Save or Save As after each edit, (a good practice in case the computer crashes and all existing edits since the last save are lost), it’s possible to have 6 or more saves for one file in just one editing session. That many save operations are likely to become noticeable with reduced image quality. That is not a problem with PNGs, which can be saved as JPG if desired when all edits are done.
Art, thank you for the catch! The Typing Gremlin most definitely took advantage of my tired state last night, after working for a long time on my taxes! (I just corrected my comment above)
If someone were to convert their .JPG’s to .PNG’s, .PNG files still don’t have a clean workflow with IPTC metadata though correct? Last I had read, they still don’t support this and several other types of metadata to be embedded.
Yes, you are correct about PNG and IPTC metadata, although I believe they can have XMP sidecar files created by programs like Lightroom, PS, PS Elements, ACDSee, or other programs. However, it should be possible to copy the IPTC, EXIF, or other metadata from the original, master JPG into the JPG saved/converted from the edited PNG, so long as the metadata has been preserved in the original JPG. Some of the less expensive or free image editing programs available may not read/write or even preserve such metadata anyway, except possibly within a proprietary database instead of within the individual, actual file. That’s one reason to keep a copy of any and all metadata in a separate RTF or TXT file with a matching file name for each and every image file. RTF (Rich Text Format) and TXT (plain text) files are likely to be readable 1 month, 10 years, 100 years from now in whatever software is prevalent then. JPG, PNG, TIF image formats will also likely be readable with future software, although just as now, not all software will necessarily recognize any existing metadata embedded in the image files, let alone any metadata buried in today’s proprietary databases for programs. To be even more sure of preserving important images and their metadata, as well as archiving such files to various media as media change with technology (e.g. floppy disk, ZIP/JAZ disk, IDE to ATA to USB hard disk, CD, DVD, BD,???) and migrating to each new media generation, use archival paper and inks to print hard copies of both images and relevant metadata and preserve these hard copies in archivally-safe storage containers. That way, so long as humans can see to read, they’ll be able to see any particular image with its metadata, regardless of how much technology changes.
I just want to say a few quick things about PNG…
It actually does support 16 bit per channel, and from my experience any PNG viewer can handle 16bpc PNGs even if it can only display 8bpc.
Epson is a Japanese company, and from my knowledge Japanese businesses are typically quite conservative when it comes to adapting new technologies (especially in computer software*). Therefore, it’s very possible that PNG isn’t supported as a scanned “save as” format simply because it is a newer format (“only” 15 years) compared to JPEG, TIFF, BMP, etc.
Consider that Windows 7’s built-in TWAIN scanning software lets you save as PNG even if you’re doing so with an Epson scanner! (which is what I’ve been doing)
*Example: the 2011 PC version of the Japanese visual novel “Fate/stay night: Realta Nua” uses MPEG-1 (seriously?!) videos at 640×480 with “burned-in” letter-boxing (actual video content only uses 640×360 area).
I have recently purchased the Epson V370 scanner with the intention of scanning my paper photos (again, as I want to scan at a higher (600) resolution). I haven’t started yet, aside from a couple of practice runs to check out the scan software (Epson Scan), mainly because I am not sure which format to scan in. I was going to use Tiff, for the reasons Curtis mentions but after reading this article and all of the comments I am confused, I didn’t even know about PNG, only JPG.
I would consider myself to have advanced goals rather than basic because archiving is important to me but I am relying on the scanning software and the expertise of other user’s knowledge and experience.
I do not have any numbers after my name and a lot of the comments went way over my head so to speak, so back to basics – Curtis do you still prefer to use Tiff 4 years from originally writing this article? I don’t mind spending time scanning and I have a separate hard drive to back up in.
Over the last two days I have read quite a few of your articles and really appreciate the effort you put into answering questions and comments as well as inspiring others to ‘get started’.
I should mention that another reason for deciding on tiff is that I am scanning multiple photos and then need to crop them into single photos. I am assuming that the scan I make is scanned as one photo therefore when I crop them I am changing the dimensions of each photo which is why I think tiff will work better for me as the original sizes will remain the same. Or is there a better way of separating each photo when doing a multiple scan? As I mentioned in my previous post, I have the Epson Perfection V370 Photo scanner.
I also want to add the file names suggested in another article and add comments to each photo. Yep, long days ahead!
Debbie, with Epson Scan you likely have a choice of saving as TIF, JPG, or PDF. You can choose any one format per scan. Please use the PDF option only for text documents, NOT photos. It compresses your files and makes it very difficult to edit photos. JPG files are always compressed so you will get better quality using TIF. If you invest in VueScan Pro (about $90.00), you can save each scan as a DNG (absolutely all the data your hardware can capture),TIF (loses some data through software processing, even without compression), JPG (always compressed), and as PDF. You have a choice of any single format or any desired combination of formats by checking the options on the OUTPUT tab in Professional mode. I recommend DNG for maximum quality plus JPG for immediate use without further editing other than resizing for web or email.
Thank you Art, you are right about the choices I have to scan, PNG is not an option, and I have started using Tiff. I can crop each photo while it is in preview and then the photos are scanned separately so it works well and I am happy with it. It took about 2 minutes to multi scan 6 photos. I will scan all my photos as Tiff even though I don’t intend to crop or edit a lot of them – YET. I’m sure if I scan certain photos as JPG, they will be the ones that I want to crop in the future, Murphy’s law and all that.
I have taken the wise advice in naming each photo with year/month/day and using xx when required. I am adding name/event/location then settings then giving them a number using the 00001 system, also using _ and -.
At the moment I am just adding these when I save the photo, I had a look at Picasa but it is being retired in favour of Google photos, see here http://googlephotos.blogspot.com.au/
I am about to check Google photos out as I would like to be able to leave a comment on each photo rather than in the name.
Debbie, I agree with everything Art told you in his reply. I just wanted to jump in as well to answer you specifically in my case, yes, I still do prefer to use .TIFF (uncompressed version) over any format such as .JPG or .PNG.
The only exception would be, as Art pointed out, .DNG … if you have the luxury and desire to work with them.
If you choose to use JPEG compression, be sure to disable Chroma Subsampling. Some software will not allow it to be configured, and will either always use it, which is very bad, or tie it to an overall quality slider. Chroma subsampling will halve color resolution, usually in both dimensions, which might not be noticeable on most images, but will be visible on red, saturated parts, such as fire or flowers. You might discover that those sections are blocky when you zoom in to them while editing, when it is already too late.
A scanner is almost equally capable of resolving any color, and delivering a ‘sparkling’ detail, if it is present in the copy.
My software has a dumb TIFF/JPEG writer which would aways use subsampling and produce significantly poorer quality at any setting, compared to plain JPG. Photoshop would open such files with nasty 2×2 blocks.
Quality 98 or Photoshop’s 12 is ‘nearly’ lossless, and an acceptable choice.
Just a note about the file formats you mentioned. This is very good information but I want to add just a little more to it.
I am a retired mechanical engineer and now I do freelance work at home. I use AutoCad 2010 and often go to a customer site to photograph a part of machinery or electronic equipment. I have learned that Autocad allows importing a photo onto a layer so I can then draw over it on another layer. But the Tiff and jpg images can’t be manipulated as much as the png image. Apparently AutoCad has accepted this file format to handle the best in a drawing.
I realize this will apply to very few of your readers if any at all. I just wanted to mention it to you.
Best Regards, Franklin Reid
basic or advance type of person? Curtis…. if we are reading your “sort” articles and posts… or those from Art… (no offense to any of you) we all have to be advance users
Good site!!!
Keep in mind that TIFF turns 30 in 2016. You should be able to easily open a 1986 TIFF file in todays graphics programs, and possibly even open one of todays images in a 1986 graphics program (that predates Photoshop by about 4 years). So for longevity, its hard to beat that.
From the point of a computer scientist I have to say you not really have a clue what you are talking about, if you are talking about compression. The term “loseless” in computer science (and so in this case too) means that you are able to reconstruct your original file (or image in this case) bit by bit without any variance.
You don’t get any information (which is what you call quality) more with an uncompressed tiff than anybody else with the same tiff compressed loseless.
The only difference will be, you have an file maybe 10 times as big as the compressed one.
The argument of proprietary algorithms who are not readable in 20 years are gibberish.
The most common algorithms for compression within TIFFs are LZW and ZIP.
The first one is around for more than 20 years and the second one is used for image-compression for more than 10 years. Each of them has plenty of libraries and programs who can decompress them and if there would be none of this tools left, they are documented this good nearly every programmer could recreate such tools within no time.
There are a much higher chance your HDD or filesystem will not be readable in 20 years than getting problem with zip-compression.
Great point… I’m sorry you’d decided the troll way to put it.
So does this mean I have to re-scan all my paper photos and slides that I’ve foolishly saved at .jpg?
Hi Wendy. I don’t think it means you have to rescan all of your photos that you scanned and saved as a .JPG file. As a purest who is teaching people how to secure as much quality as possible from our photos as each of our abilities and hardware allows us, I suggest people to save as an uncompressed .TIFF file when given the choice because it rules out any quality being tossed out from the process of compression.
And please, don’t think of yourself as a fool for saving as a .JPG. This wasn’t a foolish move by any means. Not knowing the difference or even choosing to scan as a .JPG doesn’t make you a fool. They are both very viable formats to save your photos. They just compliment two types of separate goals — one more for those with basic goals, and one for people with more advanced goals.
If it makes you feel any better, know that most scanning services out there where you mail in your photos, they scan them and then mail them back to you, offer .JPG files with all of their basic packages. For most of their customers, .JPG files offer the easiest file format for them to do whatever they want with it without any other knowledge of their computer or software on it.
For example, if all scanning services scanned in 48-bit uncompressed .TIFF file, and sent their customers 80 MB images for all of their analog photos, it’s very possible their customer support numbers might start to light up with people wondering why they can’t post these massive files onto Facebook!
But, for those inclined to learn a couple more steps in the archival workflow, the uncompresseed .TIFF offers a whole lot more room to store and edit your master images with the highest amount of quality and detail as possible.
If you still feel like you missed out on your chance to save as .TIFF files, my personal suggestion and opinion would be for you to scan the rest of your collection from here on out and save as uncompressed .TIFF files. (I keep saying uncompressed because there are versions of .TIFF that save as compressed — so if you have the option, you want to make sure you have it set as uncompressed). Then, when you are finished scanning the rest of your collection and you’re done, THEN if you still want to rescan the first photos you saved as .JPG, then rescan them then.
The most important thing is to get digital copies archived as soon as possible and backed up preferably in more than one location. So, by finishing the scanning first, you’re putting the longevity of your photo collection first, before your personal goals of having all of them saved at the highest quality.
“But, for those inclined to learn a couple more steps in the archival workflow, the uncompresseed .TIFF offers a whole lot more room to store and edit your master images with the highest amount of quality and detail as possible.”
Except that a PNG does the job just as well, and doesn’t require leaving the file uncompressed to keep, which is an awful way to store image archives unless you enjoy wasting money on storage space. The only reason you’d use a TIFF file is if you want to save an image’s layers for editing purposes, which isn’t necessary when you’re scanning in photographs unless you’re actually doing extensive editing on those, and if you’re still in the editing process the only advantage to TIFF is that you can view more easily in photo viewers more easily than native editor formats like PSD or XCF. And I think TIFF supports both lossy and lossless compression, so again, if you use lossless compression, there’s literally no reason for your average person to store it uncompressed.
Lossless is *lossless*. Period.
Also, you can store EXIF metadata with PNG.
Hello everyone, I have another question for you experts: when I scan booklets of my cd collection and save to PNG and open them in photo managers, the scans appear not in the position you usually read them but in 90° angle to the right. Is there in loss when I rotate the scan 90° to the left with for example microsot office 2010? I’ve noticed the size of the file increasis a bit after a rotate and save and the increase varies depending on what photo manger I use for the task. Should I use any particular software for rotating them to the original “reading” position?
Thanks
Elias, since no-one else has replied, i’ll offer my own not-too expert opinion:
No, rotating any exact multiple of 90 degrees will result in no loss to image quality because images are rectangular arrays of square pixels. (Rotating by any other angle, on the other hand, involves “re-sampling” and thus potential degradation of quality.)
Any slight changes in file size, if it occurs, is likely a result of other factors discussed at length above.
“I consider compression, even the lossless kind, to be an act of compromise to the overall integrity of your collection.”
that’s pretty bad claim belief in an structurally contradicting statement (what is the compromise of a lossless storage ? a lossless coding is a coding like any other, just more efficient), but to base and write a whole article on this single nugget of wisdom is plain fraud.
Your argument against compression seems completely illogical to me:
“You’re putting more trust in that mathematical formula to “destruct” your photo now and “reconstruct” it later than you may want to give it. Some formats have so many variant types of compression, who is to say that “shareware” graphics program you found one day off the internet to compress your images will be around 20 years from now?”
You’re putting trust in many “mathematical formulas” every time you use your computer for anything at all. What if the scanner driver accidentally sets every third pixel to a few shades lighter? What if the program you use to save the TIFF image is saving the TIFF incorrectly? You’re already putting your trust into far more places than you can possibly imagine by scanning and saving an image, so why sacrifice storage space to prevent an extremely obscure case of improper data compression?
If you use a well-known, well-tested image editor you’re not going to lose data by saving with a format that supports lossless compression. It’s called “lossless” for that very reason, and these algorithms are very well standardized. Most of these applications are even using the exact same library of code to perform the compression, not randomly making up their own.
If you’re worried about data loss, you’re looking in the wrong place. How many times have ACTUALLY heard of or witnessed an image editor compressing a common image format incorrectly and in an irreversible way? Now ask yourself, how many times have you heard of or witnessed a data storage device failing? I guarantee you’re infinitely more likely to lose data due to a hard drive failure and insufficient backups than a broken compression algorithm.
A couple of people with advanced goals suggest PNG is great.
Library of Congress accepts both uncompressed TIFF and PNG as for archival.
http://www.digitalpreservation.gov/formats/content/still_preferences.shtml
National Archives of Australia has chosen PNG as the “preservation format” for bit-mapped images and converts images in other bit-map formats to PNG during its ingest process.
http://www.naa.gov.au/Images/Digital-Preservation-Software-Platform-v1_tcm16-47139.pdf
Library and Archives Canada has adopted PNG as a recommended format for still images.
http://www.collectionscanada.gc.ca/obj/012018/f2/012018-2200-e.pdf
Mariano, thanks for posting these links for us! It’s very interesting to see how new technology and in this case formats are being used and adopted.
I originally wrote this post on “PNG vs. Tiff” some time ago. And the PNG format, as well as others, has grown. So, there is always a need to go back and re-evaluate whether a set of conclusions needs to be modified. Maybe the PNG format has evolved: is now accepted by more scanning software companies as a “save as” option, allows you to save IPTC metadata inside, and you should therefore really consider it as an option in your own personal photo collection. Or maybe it hasn’t.
One thing that jumped out to me after looking briefly at the pages these links took me to, was the question whether .PNG was primarily added mainly as a format “option” because it solves the problem of transport. What I mean is, how would people get their digital photos TO these organizations. Asking people to burn a DVD and mail it through the postal service is asking a lot of people now. But, having someone email or even upload a photo to a website is less of a chore and may generate a greater response. In this case, PNG becomes a great option because it reduces file size making it easier to send from people’s home broadband lines, but still retains the “lossless” quality.
While I agree that asking individuals to snail-mail digital image files to any archive, particularly national ones such as those referred to above, is asking a lot, don’t forget that people likely wishing to donate files to such an archive are probably going to be planning to contribute collections with numerous images, probably numbering in the hundreds at least. Burning an extra copy of such images to sufficient DVDs/BDs (Blu-ray Disks) for the archives is not going to be a major time factor when personal archive disks are being created. Considering the amount of time involved in emailing such a number of files, even as relatively compact .png format, it’s not a project I’d relish undertaking. I’d much rather package a batch of DVDs and take them to the post office to mail them to the recipient. The other factor to consider is the time required at the archive to download emailed files and then transfer them to archival media, be it DVDs/BDs or external hard disk drives. For budgetary reasons, most archives operate with minimal paid staff and hopefully, lots of volunteers, but they’d likely be reluctant to use staff time to sit at a computer downloading a large batch of emailed image files.
Until such time as the .png format is revised to include EXIF information with the image information in a single file, each image will require both its .png image file AND its appropriate side-car file with the EXIF information. It seems to be all too easy to accidentally miss including all the side-car files in a batch of image files if they’re to be emailed, either at the initiating end or the receiving end. If they’re all kept in the same folder on the originator’s computer, they’re much more likely to be missed when burned to DVD/BD.
Certainly, most archives will place a higher value on files with EXIF data, either contained in the image file as with .tif or in a side-car file as with .png, than on files without that data.
These are just a few recent thoughts about the .tif/.png discussion.
“Many may disagree with me, but I consider compression, even the lossless kind, to be an act of compromise to the overall integrity of your collection.”
There are no sides to this, nothing to disagree on. Lossless format is what the word says. I suggest getting more knowledge before _offering advice_ to people.
Some of the files could not be read at all. This happened a few years ago. Threw the disk away. The cd drive would read and read. And then say disk could not be read. Happened on any drive. Some images I got off perfectly. And then others had damage. Where you couldn’t see some of the image. Or the file wouldn’t load at all.
Hi Graig Smith,
Can you send Curtis screen shots of any of the files from your dvd disks that have gone bad to show what the files look like when you try to view them? I’m not questioning the truth of your comments, just interested in seeing what disc rot actually looks like. Possibly other readers would also be interested in seeing what it looks like so Curtis could post any shots you might be able to provide him as a learning tool for all.
As Malcolm and I both mentioned late last month in this forum, hard disks can become unreadable, even if they’re in mint condition. Over Christmas, I visited a friend who had received a brand new Dell computer but he’s unable to connect the hard disk from his old computer to retrieve his data. The new computer has a totally different, although now standard, interface to control the hard disk(s) and it’s incompatible with the older standard. Fortunately, he believes he has most, if not all, of his images and other data transferred to CD since he’s unable to access the old hard disk. A few years ago, adapters were available to connect old disks to new computer mother boards so they could be used but when I checked online this week, such adapters were not to be found. Again, as Malcolm asks about SCSI drives, will today’s hard disks be accessible in a few years? This is a point worth considering. As with optical media or tape media, serious archival and backup strategies should include plans to migrate all data to new media every few years to retain accessibility.
Art Taylor
“Optical media is preferable as long as the dyes in the media is archival grade and they are store properly…”
i disagree, as i said before, i had dvd disks go bad after only a couple of years. disc rot does happen. even if you have it stored in a dust free scratch free environment which is how i stored my disks. honestly. i have had better luck with hard drives. the only thing is hard drives are prone to catastrophical failures, where you can loose a whole disk at once. which is why i am using online backup. online backup. sends stuff out for backup hourly or so every time you add some images, it sends them off and they are already backed up. time machine isn’t the only backup i use. it’s just an extra. my main backup is crashplan’s service.
“The best media for backup is tape (still).”
i agree. all sorts of large data companies use tape backup. it’s reliable. i bet crashplan uses tape backup on their servers.
You imply that with lossless compression there are “lots of filters and computations altering the outcome of the master image during the scan”. It seems odd to me that most scanning software would do this, but not implausible. Have you actually tested this and/or do you have any evidence you can point to? And if you did test this, was the final image quality affected to any discernible degree?
An excellent point Malcolm makes is the need for TWO tape drives. This is an issue seldom likely to occur to most of us until it’s too late to do anything about it. I also appreciate his comments about TIFF vs PNG formats. Some valid points are raised.
Thanks, Malcolm for some excellent information about RAID. Any information about a RAID controller failure rather than a RAID drive failure? I seem to recall reading somewhere that a controller failure is at least as likely, if not more so, than a RAID drive failure. Do you know if one could leave the RAID drives powered on, although disconnected from the controller, while a faulty controller is replaced?
Malcolm touches on the issue of hard disk drive controllers when he mentions SCSI (Small Computer System Interface), which is used to be (but no longer are) standard on Macs and was available with an optional SCSI controller add-in card for PCs. The contemporary controller on PCs was IDE (Integrated Drive Electronics), later referred to as ATA, in turn superceded by EIDE (Enhanced Integrated Drive Electronics), then by SATA/PATA, and eSATA (for connecting external drives to the computer’s motherboard. See http://en.wikipedia.org/wiki/Hard_disk_drive#Access_and_interfaces for more information. Today’s PCs often don’t have IDE controllers so older hard disks may pose problems being connected to modern computers, let alone any internal mechanical problems such as Malcolm mentions, such as the lubrication drying out. Today, SSD (Solid State Device/Drives) are becoming common since they use no moving parts. A friend of mine has several older IDE/EIDE drives which he can no longer connect to his current computer since the computer lacks a suitable drive controller. To access any data on these older drives, he has to hook them up to an older computer in his computer room. Since he likes to tinker with computer hardware, he often has such a unit available but not always. For those who have only one computer available, this should be an item to seriously consider before deciding to archive exclusively to hard disks.
While I’ve had no personal experience with an SSD, one acquaintance of mine, who uses a very recent model Mac, has reported no problems with his SSD, although I believe he uses it for data storage rather than as the system disk with the Operating System on it. A friend, who built his own PC from parts a little over a year ago and who uses an SSD for his system disk, reports that while initially it was extremely fast in comparison with SATA drives he’d used previously, over time he has noticed its performance has slowed noticeably so that it seems to be no faster now than a SATA drive, installed at the same time, would have been. This MIGHT be just because he’s added more programs over time and no longer has as much free space available as he had at first. Maybe it’s a (common) feature of SSDs. Since I’ve heard of only these two specific drives, used differently by different individuals, I don’t want to draw any conclusions about the pros and cons of SSDs, just to mention there may be unforeseen differences which might cause one to consider if an SSD is worth the extra money for a comparable size disk instead of a SATA/USB/Firewire drive at present. Perhaps the SSD technology should be allowed to mature a little more before we all rush out and replace all of our present drives with SSDs. At best, we should be prepared to migrate our archived data to newer HDDs over time, as technology progresses, just as we needed to migrate it from Zip/Jaz to CD to DVD, and now likely to Blu-Ray, if we choose to use optical disks as part of a back-up and archiving strategy.
Art
Wow, this article has got a lot of traction, and (unfortunately) a little misinformation (not a lot, but…)
First, on backup: a hard drive (in “time machine” or anything else) is NOT a backup. It is a copy, an offline copy (probably), and that is A Good Thing, but if one is concerned about optical media as a backup format, one should run screaming from any hard disk. The reason for this is that for backup you have be concerned about all sorts of failure modes, including the often overlooked “how will you read the thing in 10 years?” one (suppose you’d used SCSI disks as a backup media a decade ago… now what?). For an HDD (hard disk drive), you need to worry not only about media faults (not being able to read a particular chunk of media) but also motor/bearing/interface failures, which render the whole thing about as much use as a brick. But even worse, because HDDs have critical moving parts, they also need lubrication in order to work…. and that lubrication will evaporate over time. So eventually a disk sitting on a shelf will become a brick, and the only way you’ll get your data back is by spending a lot of money with a disk recovery business, and crossing your fingers.
Optical media is preferable as long as the dyes in the media is archival grade and they are store properly… but the main reason to prefer it is that the media is cheaper (although the time to burn it isn’t, so that must be considered). Because the media is cheaper, the thinking goes, you’re more likely to have multiple copies, and because it is light and relatively robust, you’re more likely to have remote copies, too.
The best media for backup is tape (still). Unfortunately, a sound backup strategy needs *2* tape drives (one to write, the other to verify), otherwise you risk creating backups that only one drive in the world can read… and that drive will be the one that just went up in smoke. The good news is that “cloud backup” services usually include tape-based backup of the storage, so this is (yet another) reason to go that route.
By the way, if you do have a RAID drive fail, the best thing to do is to keep the whole thing running and use it as little as possible; don’t power it, or the attached computer, off until the new drive is installed (or ready to be installed if you don’t have hot plug), because a huge area of complexity for RAID controllers is figuring out which drives to use if, for instance, the formerly dead disk “gets better”. They should get it right, but it’s your data, not theirs, that will be corrupted if they don’t…
Even more by the way: RAID1 is always a better choice than RAID5, except for the pesky issue of cost and connectivity. So six disks arranged as 3 RAID1 pairs is better than four disks as one RAID5 set, given that we’re all made of money! (Oh, wait…).
Second, others have pointed out the fallacy of worrying about lossless compression. Certainly, if you save anything to a cloud, your data is going to be (losslessly) compressed anyway, so this is a non-issue.
Third, the comments about vector formats being supported by TIFF are (at least partially) valid, but entirely irrelevant to a discussion about scanning. Sure, if you want a single file format for multiple types of image, then this is a very reasonable concern, but for scanning, you’re not going to be getting vector formats.
Fourth, Adobe’s DNG is equally irrelevant to scanning, because DNG is a format designed to be a superset of camera sensor raw formats, and a lot of value of raw sensor data is that things like white balance and exposure can be tweaked. A good scanner will control those (the lamp produces the same color light, for example), so DNG doesn’t really have a role to play here.
Fifth, as has been suggested above, the Achilles Heal of TIFF is that it is not fully standardized (the spec is actually copyrighted by Adobe), and (to make it worse) it is a “Container” type format (like AVI or MOV for video). So a TIFF file can have the image data in one of many formats, including lossy compression (JPEG, no less)! Of course, at first this is only a problem if you make it one (for yourself), but as a data interchange format, TIFF is a long way from being ideal. PNG is also a container-type format, but the structure of mandatory-vs-optional elements is much more rigorously defined.
Sixth: PNG’s big strength is in the “P”: portable. the same file can be displayed on a phone, an iPad, any notebook or desktop machine regardless of whether it is a Mac, PC, or Linux (generally), because the rendering logic is built into all modern browsers (in fact, there are three little PNGs just below the text box in which I’m writing this).
Seventh: the fact that any particular scanner does or does not support a given format is not intrinsically evidence of that format’s suitability for the task. Few scanner vendors care much about their software, and with higher-end applications like Silverfast, other considerations may apply (e.g. PNG requires compression, which requires CPU time, so saving files as uncompressed TIFFs is generally faster during the scan phase, although you may lose that time copying around larger files, and so on). Oddly enough, no-one’s mentioned JPEG2000, which is probably A Good Thing as it would just confuse the issue more (it’s another lossless format, but less well supported than PNG, so why bother….). Anyway, the bottom line is that whatever the source (camera, scanner, software) produces may well not be the best format for your personal workflow, and that format may not be the best for archival.
Conclusion: (sorry it’s gone on this long…) there are plenty of subjective reasons to pick a file format, but when it comes to scanning there are no _objective_ reasons to pick TIFF over PNG or vice versa. Personally, I like to use PNG for scans, mainly because it’s also my final output (what gets sent to the printers), with all intermediate files (including the final “master”) are in PSD. If someone batch scans stuff for me, it all gets converted to PNGs on arrival — it only takes CPU time, and I’ve got plenty of that (as long as I’m not rendering video). This works for me, but probably wouldn’t for Curtis, but for my money the portability of PNG wins the day!
(Why, yes, I have been bitten by unreadable file formats and incompatible TIFFs! How _did_ you guess?)
Looks like I might have beat Curtis to this response, Nancy. As you’ve discovered, most popular print places and photo sharing sites will convert your uploaded photos to jpg if you haven’t already done so before uploading. No, you’re not really gaining anything in print quality by scanning to .tiff IF you get prints done by an on-line company. You could likely get the benefit of your .tiff files if you put them on CD/DVD or USB stick and go to a place like Costco or other bricks-and-mortar store that offers digital printing.
While your .tiff files will take up more space for archiving, you retain more detail and quality for your archives with .tiffs. EVERY time you make a change to a .jpg and save it, you discard more detail and quality. You’d be better to scan routinely to .tiff, do any desired editing to the .tiff, then save an edited copy to .jpg for on-line printing or photo sharing.
One reason for the on-line companies converting everything to .jpg is simply that .jpg files are much smaller than .tiff files so they upload and download much more quickly and require less server space on their computers.
If your scanning software (EpsonScan and VueScan both do) and image-editing software (Photoshop Elements, Photoshop CSx, and some other programs do) offer the option to scan at 48-bit color instead of 24-bit color, using this option will give better color quality in your scans. While your monitor and any printer won’t be able to reproduce such a large range of colors, they do give better results when you do any editing or color-correction to your scanned images. To benefit from this color range, you must scan as .tiff since the .jpg format is limited to 24-bit color. When you’re done editing, save a COPY of your edited image as .jpg for viewing, sharing, and printing but keep your original .tiff with all the colors in case you later decide to re-edit for any reason. Feel free to contact me via Curtis’s blog if you have any questions about this.
Art Taylor
so…here’s my question…
I’ve scanned both to .tiff and .jpg.
using .tiff on the prints that I want the best scan.
BUT when I upload them to popular places to print…they all end up converting them to .jpg anyway…so I’m not really gaining anything on the print…am I ??
If I scan to .jpg …I archive that file…and only make a duplicate for editing. am I still losing detail?
this is part of the “8 years between scans question for me!”
thanks.
Hey Nancy! Yes, Art did beat me to a response for you (see below) but that’s okay! You’re in good hands with him and his response. I couldn’t agree more with what he just told you.
If the places you would like to upload your photos to for printing only allow .JPG files, then it says they think their customers are okay with compromising some image quality and detail for the sake of price and convenience.
Expounding on what Art said, not only is it more convenient for you as a consumer to upload smaller .JPG files than larger uncompressed .TIFF files, it’s also cheaper for the printing company on their end because they aren’t saving and processing these larger file sizes. Seconds of processing time, Gigabytes of hard drive space, all add up to more money.
And they save on tech support staff as well who would have to be trained to help their customers understand how to even save as .TIFF files and manage their larger sizes on their possibly outdated computer hardware. Overall, for these companies we are speaking of, .JPG is a good level of compromise for them and their customers.
Especially considering even if you showed many of them a printout from both a .JPG or a .TIFF source, many of them may not even care if some of the detail is “soft” looking or “pixelated” in the .JPG version. I mean, It’s a picture of their granddaughter! “She’s so beautiful!” They’re happy.
And to make sure Art’s point is made even clearer (if that’s possible), let me add that if you do a nice scan on one of your photos, and then save it as a .JPG file, know that at that moment, you are already “throwing out” some quality and detail. Then if you later re-open that .JPG file and make some edits to it and re-save it again as a .JPG, now you have just thrown out even more quality and detail. So it’s not just when you make edits to a .JPG where there is a loss, it’s from the second you settle on .JPG as your file type to begin with.
So, I would say if the highest quality is really important to you, and you would still like to work in the .JPG format for some of your photographs, try and use editing software that will let you save at the “highest quality” possible at each save. This way you will lose the least amount of detail as possible each time.
Often times, this menu item as you are saving is like a slide bar from 1-10 — 10 being highest quality. If it just says “Save as .JPG” for example, how do we know if it’s saving it as a quality “10” or a quality “5”.
Cheers!
Curtis
Art Taylor, I highly highly recommend against using optical media (DVD/Bluray) as a backup medium. DVDs and Blurays are very susceptible to “disc rot,” which is when the disc appears to be rotting from the inside, which results in data degradation. I believe it is due to rusting or some other chemical change happening within the disc. This is an extremely common phenomenon that can affect discs in as little as 2 years (I have seen it myself happening on large numbers of DVD-R/DVD+Rs, etc). Entire boxes of DVD-Rs can rot at the same time.
Please avoid using optical media as a backup medium.
Hi John,
You are the first, and only, person I’ve learned of who has actually experienced the problem of ‘disc rot’. Although I’ve heard and read about ‘disc rot’, I’ve not recognized it IF I’ve had it happen to me. I’ve checked with others and they have also not recognized the problem, if they have it with their discs.
It would be beneficial to other readers of this blog, and to me, if you would send Curtis some photos showing any images damaged by ‘disc rot’ and any photos showing visual indications on either the label side or recorded side of discs exhibiting the problem. If we could see examples of the symptoms of ‘disc rot’, it might help us to learn if we have media with the problem. It would also be helpful to know specific brand names of discs you’ve noticed particularly susceptible to ‘disc rot’. If you’ve noticed any particular brand(s) showing problem after a specific period of time, it would be helpful to know the names involved. Are they name-brand media or generic, house-brands? I’m certain Curtis would be willing to post example photos, especially if you included a description of what they show, so please send him some examples to share with his readers.
I don’t doubt you’ve experienced ‘disc rot’ but without an accurate description of the symptoms, specific brand names, and photos showing examples of the problem, your warning is of limited value to anyone reading this blog.
Art
I said in an earlier post what happened to me. i had burned cd disks and DVD disks. that even though they didn’t have a scratch on them, no dust on them. stored in a protective case for a couple of years they STILL went bad.. i lost some of the data on just a couple of my dvd disks. i permanently lost scans that i couldn’t get back. it was a bummer. and I also do not recommend storing anything in burned disk format. this comes from personal experience. it’s not a safe format to store things in.
“Maybe someday, the ISO folks will provide a revised version that overcomes these shortcomings of the PNG format.”
I doubt it. the intended purpose of PNG was as a general purpose web graphic. though you never know.
raid is a good way to store images but not as your only storage medium. Drives fail. And if only one drive fails like it did, then at least it did it’s job in that situation. But that’s not the only thing that could go wrong with a raid setup. say you get hit by lightning and all your drives hardware gets fried and none of it works. It can happen. or maybe you just get a brownout and power surges, and multiple drives fail. don’t let raid be your only backup, i know you said you had the images on other drives too, which is definitely a good thing. one way to improve your backup even more is to have offsite backup.
One thing i use is a service called Crashplan plus. And it’s not really that expensive when you consider how much it costs to buy a hard drive. crash plan automatically backs up anything on your computer you tell it to backup, and saves it on their remote servers. so even if your house burns to the ground and all your computers and backups are destroyed. later when you get a new computer you can still restore your data with your crash plan account. and they offer unlimited data for your backups. you can buy it for a year, or more the more you buy the cheaper it gets. i bought a 5 year family plan, and even used it to back my mom’s computer up. recently her computer hard drive died and i used crash plan to restore all her data. crash plan is the best of those online backup companies, i can’t recommend it enough. crash plan works on windows mac and linux too.
What Greg says about TIFF vs PNG is correct. As others have already mentioned, the other disadvantages of PNG, aside from needing an XMP ‘sidecar’ file for EXIF and IPTC data, are its lack of support for layers (useful for work-in-progress) copies of Photoshop/Photoshop Elements PSD files; and its inability to preserve color profiles. Possibly these limitations are why Epson and other scanner manufacturers don’t include an option to save scans as PNG files. Maybe someday, the ISO folks will provide a revised version that overcomes these shortcomings of the PNG format.
I didn’t realize the chances of multiple RAID5 drives failing were similar to or equal to those of a single drive failing. Like Curtis, I wouldn’t rely exclusively on RAID5 or any other version of RAID for backups and archives, although I see no problem using RAID5 as Part of a Backup and Archive Strategy.
‘Cloud’ storage may be relatively safe and secure, if not too expensive for large file volumes, but I’m concerned about the provider being there for months or years, then suddenly gone overnight, taking your files with it. As an option for one part of a backup and archive strategy, it’s probably worth considering but I’d hate to make it my only choice. Probably every ‘cloud’ storage provider uses multiple RAID drives, be they RAID5 or other, but what happens to your files if they happen to all be on the single RAID box that fails? What happens to your data if the provider suffers a catastrophic fire, flood, tornado, hurricane, or other natural disaster? Does your choice of storage provider offer geographically diverse storage facilities with redundant storage space in several locations for every customer?
I know both Apple and Microsoft offer ‘cloud’ storage space for free or very low prices and neither of these companies seems likely to vanish overnight. However, for many decades the Minolta camera company produced quality cameras, lenses made using glass ground in-house to their own specs, and introduced many innovative features in their cameras. They also produced several highly-rated scanners, both film and flatbed, over the years. Then one day in about 2005, they sold the camera production facilities to Sony, discontinued the production and support of their scanners, and completely exited the photographic camera, lens, and scanner markets. Fortunately for Minolta camera and lens owners/users, Sony has continued to provide at least some support for the Minolta cameras and the Sony Alpha D-SLRs use a lens mount compatible with the original Minolta auto-focus mount of 1985 so the older lenses can be used on today’s camera bodies. What do you do if you’ve made ‘cloud’ storage your only backup and archive medium and your provider suddenly vanishes, as Minolta did from the camera field? I’d strongly advise against using only ‘cloud’ storage with any provider.
For what it’s worth, here are my comments on file formats for archival images. As Curtis mentions, TIFF files are compatible across a wide range of scanners and computer platforms, (Mac, Windows, Linux) and virtually all professional, and many consumer-level, image editing software programs can read and write both compressed and uncompressed TIFFs. Having been developed before the advent of computers and desktop publishing, they were designed to include both bitmap graphics, like scanned images, and vector graphics created using mathematical formulae in programs like Adobe Illustrator and CorelDRAW. Today, Photoshop, Photoshop Elements, and probably other programs can create vector elements in bitmap files, using the Pen Tool and the Text Tool. TIFFs can include these vector lines and retain their editability, but PNGs need to have any vector lines rasterized or converted to potentially jaggy-edged bitmap elements. PNGs are also limited to 24-bit color depth, unlike TIFFs. While PNG might be OK for archiving original scans, if you are scanning for maximum quality at 48-bit (or 64-bit with the Infra-red dust and scratch channel included), TIFF is the better format. Also, once you have saved a COPY of your original scan and want to perform any editing on the copy, TIFFs will let you save and re-open any adjustment layers, text layers suitable for future editing of the text, any alpha channels or layer masks, and layers in general. While PNGs can preserve transparency, they do not retain layers as TIFFs do.
Back in the mid-’90s, the company that had developed the GIF format started to try to charge licencing fees to any company that wished to include GIF read-write capability in their software. Not surprisingly, this idea met with some opposition. The PNG format was developed to be an open-source format, that any company could freely use. It was intended to become a replacement for GIF, retaining the ability to include transparency as GIF can, while permitting the full 24-bit color depth of JPG. As Curtis has mentioned, JPG cannot include transparency. While it took several years for PNGs to become popular, their lossless compression, higher color depth, and transparency capability are now being recognized and more widely used. Another limitation of PNGs compared to TIFFs is that PNGs cannot incorporate EXIF or IPTC (the What, Where, When, Why, How) data in the image file. Such information needs to go into a separate side-car file which must be kept in the same folder as the PNG or it becomes separated and lost. EXIF and IPTC data can be included in TIFFs. Also, Adobe’s Lightroom, will not readily work with PNGs but has no problems with TIFFs.
Another format I don’t recall being mentioned yet is Adobe’s DNG (Digital NeGative). Since each digital camera manufacturer uses a proprietary RAW file format which is incompatible with any other manufacturer’s format (i.e. Canon/Nikon/Sony, etc. are incompatible and not interchangeable), and various versions within a particular brand may not be compatible with earlier versions from the same company (i.e. version 2 of brand X is not compatible with version 1 of brand X and cannot be processed using the different version’s conversion software), Adobe developed and released the open-source DNG format and encourages all the camera manufacturers to include DNG as a possible format in their respective cameras. The intent is to attain maximum future file compatibility for archives and libraries, much as the Adobe PDF format has achieved and retained near-universal compatibility for documents. Recent versions of Adobe software (Photoshop, Photoshop Elements, Lightroom) and other programs, such as Zoner Photo Studio Pro, VueScan Pro, can open DNG format files. VueScan Pro can also create DNG files as well as TIFF and JPG files as it scans.
Since DNG files, like RAW files, include all of the data captured by the scanner (or camera for RAW files) and do not do any image processing as for JPGs, maximum scanned data and image detail are available for processing at any time in the future. Color balance, color cast correction, exposure adjustment, etc. can be changed at any time using appropriate software. If one is shooting with a digital camera that provides a RAW option, the RAW file can be embedded in the DNG file so the original camera’s captured information is still available and can be edited in the appropriate camera RAW converter software.
My suggestion would be, and my practice is, scan using VueScan Pro, save as DNG and JPG, for ‘quick-and-dirty’ review and editing. Final editing is done using the DNG file. Note that the lower priced VueScan will not work with DNG format, the Pro version is required.
For backing up my photo archives (10s of 1,000s of slides plus hundreds of negatives and prints from over 40 years of personal photography), I burn at least two copies to DVD-R, using the slowest available burning speed. After burning the first copy, I visually check it in an image viewer/editor, usually Zoner Photo Studio Pro or ACDSee Photo Pro, to be sure it burned correctly, before burning the second copy. The second copy is also visually checked after burning. I do NOT use a back-up program to create archive backups. I drag-and-drop or copy-and-paste the original files to my burning software. This avoids any possible compression with backup software and any possible future incompatibilities with future versions of such software.
As newer media, such as Blu-Ray, come to dominate the market, my archives on DVD will be migrated to such media. When Blu-Ray starts to be replaced with something newer, the archives will again be migrated. Anybody have archived images or other files on floppy disks, ZIP or JAZ disks, or Syquest drives or tape drives that is no longer retrievable because the hardware is no longer available?
Art Taylor
yeah, i don’t use either. i mostly take digital photographs, so really i am pretty done with scanning in general. i just use apple aperture, and throw all my raw images into that. it’s a really nice program for the price you pay. And hey, pentax uses lossless compression on their raw files
i can see your point about an archive though, if you want a museum like level of preservation. but even a lot of backup software will compress the data too.
“So how much risk are we talking here? Honestly, I am not qualified to answer that. But, also honestly, I’ve never had a .PNG file not open correctly — at least not that I was aware of. But why take any chances at all, I mean not even a .00005% chance if you aren’t gaining anything from it.”
i really don’t think theres any risk to using compression. The biggest risks to archival longevity are using incorrect backup strategies. I have lost data before. here’s how it went down. i had a DVD or a CD backup of my data. and my hard drive crashed.. problem 1 i only had one backup. 2 the DVD or cd with some of my images went bad, some of my images were gone. and some others were damaged. in this case maybe uncompressed tiffs would have made a difference to some of my damaged images, but then maybe not. The storage medium was damaged. hard drives crash… and the cd or dvd i used was kept in a dust free case. but it STILL failed.
So now i don’t trust any CD or DVR to backup… i use a hard drive backup, using time machine. and i also use online backup called crashplan. this way if my hard drive crashes, and my time machine drive burns out at the same time, i still have my online backup.
“Many may disagree with me, but I consider compression, even the lossless kind, to be an act of compromise to the overall integrity of your collection.”
i totally disagree. lossless compression means exactly that…. it’s won’t loose any quality. That means you could open and save it over and over and over again and it will always look the same. not a single pixel will change… Lossless compression should not be avoided. it’s a good thing.
besides, TIFF files can use LZW compression too, or even lossy jpeg compression. Avoiding lossless compression because it’s a form of compression is pointless.. PNG just has better lossless compression than TIFF.
the bad thing about tiff is that it has way more options. way way more. it’s had features added and supported and changed all over the place. i have had tiff files open with pieces removed. layers missing. layers doing something they shouldn’t be doing. because various programs don’t always save TIFF files in exactly the same way. Tiffs are overly complicated. besides if you accidentally have jpeg compression on when your saving your tiffs then your really messing up your images.
png being a standard hasn’t been changed very much. Png is ALWAYS a lossless format, and unlike TIFF there are no settings you can accidentally pick that would loose image quality. PNG offers lower file sizes, and you get the exact same image quality that you saved it as, just like tiff with no or LZW compression. And every program that saves a png can open it in exactly the same way… because it’s made to be viewable on the web. So in the future you know your image will be supported and openable.
i highly recommend PNG for image scans.
Hey Graig! Thanks for giving your opinion on this. I knew I would have some that would disagree with me on this point, so I’m glad you stood up for “this side” and put it down in writing.
However, just for the sake of defending what I said just a little further and maybe explaining where I left it a little vague, the quote from me that you pulled read:
I wasn’t debating whether lossless compression (.PNG files are a type of this) was a compromise to the image quality of your photograph. I can only believe when they say it’s lossy, they did all of the test to guarantee that you truly don’t lose any quality. So let’s just give them the benefit of the doubt on that one. What I said was that using compression of any kind, in my opinion, was an act of compromise with the integrity of your photo “collection.” To me there is a difference.
The reason I said many would probably disagree with me, is because for many I assume, the “risk” associated with making something lossless isn’t that much to them. But to me, when you are archiving your photo collection by scanning, you want to work with the theory that you have 1 shot at this. You may be borrowing a paper photograph from a relative and you may never get that chance to scan it again. So, if you have the most “advanced goals” for your collection, I believe you want to make the most conservative choices that you possibly can to generate this scan and protect it in the best possible way.
File size:
When I chose to save all of my scans for archiving as uncompressed .TIFF files over lossless .PNG files, it’s because I saw no benefit with a .PNG file. My research showed that the .PNG format was created to reduce the file size of an image so that it could be transferred over the internet faster. The TIFF file format on the other hand was created to standardized the saved format of images when they are scanned across different models of scanners made by all different companies. It seemed like the safer bet to preserve my master images using the file format that was invented for it. If this was television, I would inject a somewhat silly analogy like, “You can cut a loaf of bread with a paring knife, but why do that when you got a bread knife stored right next to it.” Reality tv shows LOVE quotes like that!
Now, since you don’t lose image quality with “lossless” compression like .PNG uses, maybe I would have considered choosing it if I saw a real benefit. The only benefit I see with .PNG files over compressed .Tiff files is to save a little bit of hard disk space. If my archived collection needed to be portable (PNG stands for “Portable Network Graphics” for anyone that doesn’t know), file size would probably be more of a concern to me. I would rather put my external hard disk full of my master images in a safety deposit box at a bank than to carry them around in a backpack or purse. And hard drive space is so cheap now, even an average sized complete family photo collection of 3000 slides scanned at 3200 dpi/48-bit and saved as uncompressed .TIFF files will fit on a 250 GB hard drive. That’s a tiny drive for 2012 — and dirt cheap on Amazon!
For me, archiving is about making choices for the purpose of preservation and not accepting any more risk than absolutely necessary. What I asked myself over and over was why would I want to put my irreplaceable master image files through a complex mathematical formula to “deflate” some empty space out of them and then turn around and do it again to “inflate” them every time I wanted to open them? (And don’t think your photo manager won’t be opening and closing them a lot.) Especially why do this just to save some MB’s of hard disk space? I did a test and my slides were around 80 MB’s (with previously mentioned settings) as an uncompressed .TIFF file and only slightly smaller as .PNG’s — 68 MB’s. Not that big of a difference.
So how much risk are we talking here? Honestly, I am not qualified to answer that. But, also honestly, I’ve never had a .PNG file not open correctly — at least not that I was aware of. But why take any chances at all, I mean not even a .00005% chance if you aren’t gaining anything from it.
Again, that’s just how I saw it when I was choosing between PNG’s and uncompressed Tiff files for archiving master image files.
IPTC/EXIF Metadata (Captions):
Then, I also did some testing with .PNG files and found there was another compromise to me. PNG files don’t work as well as .Tiff files for storing the text information inside of the images like, captions, location where the photo was shot, who took the photo etc. (Through the years, the best place to store this information has been in the IPTC/EXIF metadata fields used by graphics professionals) This information is very important to me and my photo collection.
When I typed this kind of text information into a photo using Apple’s Aperture, then exported this image out to the desktop as a .PNG file, then RE-imported this same image back into Aperture, ALL of my captions and location information was no longer saved in the image. It was all missing. I did not have this problem at all with the uncompressed .Tiff files.
My research told me that .PNG’s don’t handle this type of metadata well without the use of the newer method called XMP which stands for “Extensible Metadata Platform.” Fantastic! There’s a workaround. But, it seems like XMP isn’t as standardized in as many third party application as IPTC and EXIF already is. So, your XMP captions will save and reload well in the later versions of Adobe products like (the expensive) Photoshop, but maybe not (at least not yet) in all of the non-destructive image managers and image editors that average people will have on their computers and want to use.
Scanning Software Capability:
And lastly, in my experience, some image scanning software won’t even let you SAVE your scan as a .PNG file. Im looking here in my drop-down list in the professional mode of Epson’s EpsonScan and I don’t even see it as an option. So, it would be hard for me to recommend this format to everyone reading this website if I wasn’t sure they could choose the format in the majority of their scanner hardware/software combinations.
So Graig, I’m with you with everything you said. Especially with all of the problems associated with compressed .TIFF files — what a nightmare that is! You know, I don’t think I have even saved an image EVER before as a compressed TIFF file. I always choose uncompressed and then there aren’t any additional choices to go wrong with. The software I deal with, I set my file type to uncompressed Tiff and it remembers this choice for every scan after that. So, I don’t have to worry about accidentally saving it the wrong way.
I am just saying that for me, and what I have decided to recommend on this site, is that when given a choice for archiving master image files, I want to pick the one that is always considering “quality” and “protection.” I believe that uncompressed TIFF files have an edge over PNG files for those who are most serious about their archived photo collections.
For people that are thinking about using a .PNG file, only for the reason to save some hard disk space, I say, if you are really doing this to archive your photos and not just to send some copies through email or post on Facebook, then just please, just go out and buy a slightly larger external hard drive so you don’t have to worry about file size. Problem Solved.
Lossless compression means you an exactly reconstruct, with bit-for-bit accuracy, the original uncompressed image. In other words, you can exactly reconstruct a TIFF file using a PNG file. PNG is ultimately equivalent to TIFF as 5+5 is ultimately equivalent to 10.
And to be mathematically strict, the smaller a file is, the smaller its chances of residing on corrupted sectors of a hard drive. So TIFFs, while exactly the same as PNGs in output, have a higher chance of corruption. Given the same conditions, more PNGs will survive the same disk failure.
I agree with graig that the likely largest threat to the integrity of digital archives is the storage medium. People should absolutely be using a geographically diverse, high-redundancy cloud backup for archiving data, as it is far and away more robust than a few hard drives managed by an end-user. PNG gives tangible advantages in this area as well. Given the same upload speed and a disaster interrupting the upload midway after an equal amount of elapsed time, more of the smaller-but-equivalent PNG images will have made it onto the offsite backup.
Source: M.S. in computer engineering
Appreciate your comment. It’s good to hear from someone with a computer engineering background.
Corrupted sectors on a hard drive is truly something I never really considered to be an issue for people. Thanks for bringing that to my attention. It’s something I really should be talking about more on this website. I have been bookmarking cloud-based backup companies for some time now, so I can do some reviews for this very concern Graig and you discussed.
I store all of my master images on a small drive tower made up of 4 “server grade” hard drives running together in RAID-5. Which as you know, but others might not, this means if one drive goes out, I can replace the drive and the other 3 drives will rebuild this data back to the new drive.
Am I correct to assume this will help me in a case like you mentioned, where a drive experiences a “corrupt sector”?
Yes. If one drive is corrupted, the other drives will be able to restore the data.
Also notice that he mentioned storing them in geographically diverse locations. That is the ideal backup solution for critical data, because a single geological disaster won’t be able to destroy all your backups at once.
Note however, Manhattan and Miami aren’t geologically diverse. Being coastal cities, they can both be susceptible to many of the same geological events. For example, a single tidal wave could take out both sites at once.
If decide this is the backup plan for you, the typical way to implement this at home is to store your data online (in the cloud) through one or more of the many online storage providers. Be sure to find out where they store your data, and whether or not it is encrypted securely. Some of the popular providers maintain a master key by default and are able to decrypt your files, so be sure to check.
All excellent advice Mark! Appreciate the comment.
While we are on the topic, I would like to point out that RAID is *not* a backup solution. Using RAID can in fact introduce more points of failure, and risks of losing data. In your hard drive tower with hardware RAID, there is a chance that your RAID card can fail. If this happens many years down the road, what are the chances that you’ll be able to find an identical replacement? And no, a similar product from another company is more likely than not to be unable to read it.
Furthermore, while your drives are connected, there is a risk of some bug or malicious software that deletes, corrupts, or otherwise destroys your files. This may be unnoticeable if it only happens to a portion of your archives.
RAID is only for improving availability – ie. a drive fails and your deadline is in the morning – you can keep working, or just pop in another drive and soon your performance (if RAID 5) will be up to normal.
Use offline backups; copy files to either other drives or storage media, and send them across the country; in bank vaults, etc.
I agree with John on the compression issue – there is none. In fact, PNGs are probably more likely to work decades down the road – it is an open standard, and anyone can look at the specs and write their own decoder easily. Not so much for the TIFF container; the binary data inside isn’t as well documented.
Be cautious with cloud backups – who’s to say they will stay around, or that they don’t mess up your files? They can be one of several backup locations, though.
Since we are throwing degrees around: BS/MS CE.
Great advice you have for those who believe RAID is a backup solution. People really need to understand what you said if they are relying on it as their only “copy” of their master images.
I personally rely on RAID 5 to help secure the reliability of my first copy of my master images. From there, I have additional backups on other drives. However, I still really need to diversify and get them into different geographical regions as you and Mark are pointing out so kindly.
Thanks for your thoughts Puey! Appreciate it.
Oh, I forgot to mention: there is one great advantage to TIFFs: embedded color profiles. I could save the profile as a file and load that after loading the PNG, though it is likely too much trouble.
Another would be that it’s more likely to be one of the output formats of image capture programs. For example, the version of Nikon Scan which I am using now only outputs that, or some other more esoteric formats.
Yes, I think MANY people will have this issue with their scanning software. Most people use the software that comes with their scanner — such as Epson Scan that comes with Epson scanners. Unless I am completely mistaken, the latest version still doesn’t allow PNG as an option even in the Expert mode.
RAID5 is hardly as safe as you think. Do you even have a hot spare? Did you know the odds of multiple drives failing are almost as high as a single drive failure?
More points for long term archival with PNG: It’s a standard format (as in ISO). TIFF has never been standardized and requires implementers to reverse engineer others work (which is never 100% reverse engineered). The PNG documentation can be gotten from ISO 50 years in the future (assuming there is no apocalyptic situations that occur before then). You can’t get a complete document on TIFF from anyone today, let alone in 50 years.
PNG is a safer archive format. Your metadata argument for TIFF holds up, but as you said, XMP solves that problem. I’ll pick PNG.
FYI, I’m a software engineer.
Hey Greg.
Actually, true story, about 3 weeks ago, I had a drive fail in my 4-drive Raid 5 tower. Beeper was going off and light was flashing for one of the drives. And, I actually did have another drive ready to go — it had some stuff on it so I had to take a day or so to figure out how to move media around before I erased it and could then use it for the tower. (It was the same model number as my others). But, had I not had another drive ready to go, I would have ordered it off Amazon and would have had it in 2 days with Prime. (For those who don’t know, Raid 5 allows you to continue working with a failed drive. It’s not best to though. You want to replace it as soon as possible)
But, what I think is cool about Raid 5, is that it IS safer than not using and redundancy at all on your primary storage drive. What if I was the average person that keeps their photos on their laptop hard drive, and that drive had gone out on me a few weeks ago. Then I would have lost everything at worst, or at the middle, I would have lost whatever was new since my last backup to a second drive, and at best, I would have to download what I was now missing from a cloud service (if I was using one).
Maybe you disagree, but I think Raid 5 for a primary storage for photos is awesome because of the exact situation that I was just in. Raid 5 bought me time to replace the bad drive.
You state that more than 1 drive failing “is almost as high as a single drive failure”; I don’t have stats on that but, I can certainly say I am glad the troubles in my recent incident was just sticking in another drive. It was pretty painless.
So my point is not that people should use Raid 5 alone and think they are completely safe. My point is that if you have the money and want to deal with it, Raid 5 is a great place as the first or primary place to store your photo — in my opinion, it’s safer than a single hard drive. Then look into additional drives and cloud storage for your backups.
PNG for Archival
With all of the software engineers speaking up for PNG, I will admit it’s made me want to look into it more. So I really do appreciate you (all) speaking up in it’s behalf.
As someone who has to stand behind a format to recommend to the masses on this website, and who has so far recommended .TIF, these questions would still need to be answered for me though:
1) If PNG is superior, why doesn’t EpsonScan scanning software allow me to choose PNG as a file format when I use their scanning software? (Epson is still arguably the best scanning hardware company for pro-sumers. And this is just Epson. What other scanners and software don’t offer .PNG as a choice to save in?
2) Why aren’t any (or the majority) of the major scanning services offering .PNG as a format to have your scans sent to you in? Of all the ones I have investigated so far, I am not familiar with one that offers it. It’s always .JPG or .TIFF (for an extra charge).
3) If .PNG really IS the way to go for archival, then why isn’t XMP universally accepted in all of the major Image Managers on all platforms.
So, as an example, consider people with basic goals and basic computer experience who scan with EpsonScan software and who decide to use .PNG to archive their photos. They would have to become more advanced users in order to learn how to batch “re-save” all their photos from .TIFF to .PNG after scanning. And then they may be required to purchase and learn all new advanced software that handles XMP until it’s more universal.
But, again, Greg, your input and others who defend .PNG has definitely piqued my curiosity. Anyone who really knows me knows I am not one to hold onto old technology because I refuse to adapt. And if .PNG is the new future, some day, I hope all companies get on the wagon and support it as well as XMP.
Thanks for your thoughts Greg.
I am some years late but since you said “TIFF has never been standardized and requires implementers to reverse engineer others work” and “You can’t get a complete document on TIFF from anyone today” here it is:
https://partners.adobe.com/public/developer/en/tiff/TIFF6.pdf
Also ISO have standarised TIFF/EP and TIFF/IT, there are also RFC for TIFF-F and TIFF-FX
Thanks “Floppy” for taking the time to leave a link to that for us — appreciate it!
Floppy, ISO may well have set some standards for recent versions of TIF files, but that doesn’t mean that everybody applies that standard. One of the major concerns about ANY file format used for archival purposes is that it can be opened and read at any time in the future. At one time, TIFs created on Macs used a different bit-order than those created on PCs and the two variations were not always 100% interchangeable. Macs could usually open PC TIFs but PCs could not necessarily open Mac TIFs. Fortunately, with current operating systems and software, that’s no longer a concern. However, Corel PaintShop Pro X6 gives several options when I choose Save As TIF. I can choose FAX-CCITT3, Huffman Encoding, LZW Compression, Packbits, or Uncompressed for each file. Sometimes, JPG-compression is offered as an option, even for TIFs. While TIF is generally considered to be a “lossless” format, if the JPG-compression option is chosen, TIF becomes a “lossy” formatI also need to choose to save RGB or CMYK color channels and whether or not to apply an ICC Profile to my TIF.. The TIF format can use layers and can hold 48-bit color information but not all programs can recognize more than one channel or read 48-bit colors of data. Not all software, even today, will necessarily be able to open and read every TIF file, especially if it’s an older file from a decade or so ago. These are some of the concerns about TIF being a ‘Standard’ format. There are simply too many possible variations within the so-called standard..
Actually a RAID 5 is not in any means good for backups. Its there for high availability. And nothing else.
Problem is, if one drive fails, there is a really huge chance that another one will fail during the rebuild of the RAID. Because they are from the same batch..
RAID5, like every RAID, is only there to keep things working in case of a problem. Not for Data security.
RAID5 is btw outdated. RAID6 should be used.
I’m commenting late as well. I stumbled upon this article when researching compression algorithms. I’m an engineering Ph.D. and I research medical imaging techniques. Part of this research includes dealing with a lot of large datasets/images that must be stored and/or transferred across networks. It is very important in any field of science that data is archived with absolutely NO loss in quality/accuracy. It is very common for scientists to deal with massive amounts of data. With all that said, compression is very important to us and we put a great deal of effort/planning into archiving. I would say I have “advanced goals” and I hope this paragraph validates my 2 cents about the subject…
I use both PNG and TIFF. However, I rarely use “uncompressed” TIFF. As the CS guys have mentioned, lossless compression is truly lossless. Digital images are simply a 2D or 3D array of numbers. I do image processing often and can tell you first hand that if you open a lossless compressed image and subtract it from the raw image, the result is an array of zeros. This is just a variation of the Additive Identity property you should have learned in grade school. It all means that the two images are EXACTLY the same.
Now, there is the other issue which is, “which format is less susceptible to corruption?”. The short answer is either. Re-opening and saving a file is not inherently likely to corrupt the file and if a chance at all exists it is very likely a system problem or software problem, not due to the standard itself (in my experience this happens more when saving TIFF than PNG because TIFF is not so standard and is open to being implemented incorrectly; again unlikely overall). A previous comment mentioned how file size affects corruption. An uncompressed file is large and therefore more likely to exist on a corrupt sector. However, compressed (PNG, or LZW TIFF) is smaller and more files would exist on the corrupt sector. Either way there is no reason to consider that trade off especially if you backup and duplicate your archive anyway.
Which brings me to archiving. It is somewhat standard for important research projects to backup data offsite. This means that somewhere other than the research facility, the data is stored on tape. This makes sense for very important data. It doesn’t say that local backup is unreliable, it just says that the chances of your two local drives both failing at the same time is about the same chance that your building will burn down. For photographers, the best solution really is the cloud, while using a local RAID server and DVDs for very important photos. You have a duplicate local copy and an “offsite” copy (which is typically duplicated itself — disclaimer, not all cloud services are equal), and a hard copy that lasts decades. You could do more, but there is a point where your wallet and time can’t satisfy your insecurity about storage reliability. I think the best way to not lose your photographs is to be active and know when you need to manually duplicate or if ever change the format. You can’t predict what changes are to come so it is just best to adapt. With that said, PNG in my opinion is least likely to change and will certainly not be forgotten (i.e. even if we move on from PNG, there will exist converters to whatever format is currently superior). TIFF could be more difficult to read in the future if there was some non-standard way it was saved (though still unlikely).
I prefer TIFF for scientific images because of the options and ability to save text with it. We typically save notes on how data is stored with the actual data so it is not a problem to read in the future. However, for a photographer who doesn’t know how to manipulate/recover/read non-standard data, PNG is probably best. Either format is OK really, but as a photographer there is no reason to use TIFF unless you really like the way it does EXIF, and even then I view this as favoring short term preferences about usability. And I reiterate, the best way anything can survive time is not just through how it’s made, but by its upkeep. The golden gate bridge is not still in use and does not look the way it does because of the steel it is made of, but because it is being continuously painted.
Hello everyone! I’m not a photographer but I scan a lot booklets from my cds collection. Befor starting to use PNG for the reasons given by John I used TIFF. My question now is: What is the best way or software to save my old TIFF in PNG?
Thanks Curtis now I’m using Picassa as you suggested. Thanks.
Elías, that’s a great use for PNG — especially since you probably aren’t trying to save a lot of image metadata with them like captions etc. But, as many have already noticed, it’s not easy for most people to scan something and immediately save it as a .PNG file because most consumer scanning software doesn’t list .PNG as a “save as” option. So, like in your situation Elías, you have to either convert each one, one at a time by loading it in a file viewer/editor and doing “save as PNG”, or you have to batch convert them using an application.
High-end photo managing programs like Lightroom and Aperture will let you select a group of photos and export out as different file types. This would be the optimal solution. I just loaded up my version of Picasa and it appears it sadly doesn’t offer a file type conversion in the export process. (Google is keeping it really simple so basic users make as few as mistakes as possible)
On my Mac, I bought a nice small lightweight program called “Photo Batch” from the Mac App store that does a lot of batch conversions. This does the trick nicely for file type conversions. Mac users could probably also use Hazel or Automator to set up batch conversions. (Someone correct me if I am wrong)
I looked on Windows 8’s App Store and didn’t find anything. Unless I navigated it incorrectly, their selection is very limited right now with applications. I did find a program through searching around on Google that should work for Windows users — AVS Image Converter 3.0. I couldn’t find a price, so it’s either free, or is a trialware application.
That’s great Elías that you are trying out Picasa. Hope you continue to love it!
Thanks for your reply, Curtis. I’m going to try your option and let you know. Also, I’ve found this other on the internet
http://www.rasteredge.com/software/tiff-png-converter/
but haven’t got the time to try it yet. Thanks again
You’re welcome! And thanks for the link to this other application as well. I will add it to my resource page I just started and try this one out as well sometime.
I thought I would chime in on the conversion tools. A free tool that I use quite often and that is still active (since late ’90s) is “Irfanview” (http://www.irfanview.com/) It covers a number of different file types “out of the box” and more when using available plugins. This includes a variety of TIFF variations. The thumbnail viewer that is installed has a batch function available when selecting several images. You can also utilize the command line features of Irfanvew to make the conversion process faster.
Got the answer I wanted on whether to scan my old family photos in TIFF or PNG format. I think I fall into the advanced side and am willing to spend more time and money (hard drive space) to preserve these photos. TIFF it will be. Now it’s just a matter of what DPI to set it at.
Hey Yolanda. I’m pleased to see that you believe you fall into the “advanced side.” I don’t think you will regret it at all. It may take you a little longer, but in the end your photo collection will be so much better off. I’m scanning about 30 photos every morning now. It’s just part of my daily routine and I love it!
As far as the DPI, you will just need to find the sweet spot you feel comfortable with. For slides, for those with “advanced goals” like us, I feel between 3000-4000 dpi is the best. I am currently scanning mine at 3200 (a pulldown setting in my EpsonScan software) which in 48-bit mode (which I chose) makes them about 80 MB’s a piece. I wrote a post about it called “Q&A: What’s the Best DPI or Resolution to Scan Your Film Negatives?” that might help you with more information.
For prints, for those with “advanced goals,” I found that it really depends on the size of the print as to what the best dpi is. You want the resolution to come out big enough that you can do enlargements later if you want — or extreme crops. I also wrote a post about this that goes into a lot more detail that you might get a lot from. At the end, I present a chart with my DPI suggestions for each sized print. The article is: “The DPI You Should Be Scanning Your Paper Photographs.”
I hope this helps you out! Please let me know how your scanning/organizing goes. Feel free to contact me through my “contact” me page.
Best ~ Curtis