After saving for web and device, no color space assigned ??!!??
RovingEyePhoto
Registered Users Posts: 314 Major grins
Seems I asked this question way back when, and came away with something that, in today's light, still makes no sense to me. By "today's light", I mean I've now processed thousands of pics through Photoshop CS3, and still one aspect of converting my TIFs to JPEGs doesn't sit right.
After opening my TIF in CS3 and resizing, I "save for web and devices", I find that a fast and convenient way to fine tune JPEG size by using the slider in the "save for web and devices" pop up box. In that pop up, I check the box that says "Optimized", and in the menu that drops by clicking the button to right of the "Preset" box (which I default to "none"), I check "Convert to sRGB" and "Include XMP". And then I save. However, and this is where the old mystery reaserts itself, when I reopen the resulting JPEG in CS3, I up comes a pop up box that tells me image has no assigned colorspace, and gives me the options to leave it that way or check the colorspace I'd like assigned (sRGB being one of the options). Way back when, I came away with the understanding from here at dGrin that my JPEG is in sRGB colorspace, it just doesn't read that way when I reopen in CS3, and in my rush to get started, I accepted that. But now, thousands of images later, water under the bridge so to speak, it occurs to me that I finally ought to revisit the question so that, even if the past thousands are deficient, the future thousands will have a propoerly assigned sRGB colorspace. I use sRGB, BTW, because all at this point are for screen viewing, not print. I always can go back to my RGB TIFs if I want files later for print, already processed through CS3 (actually through ACR except for sizing and converting) as far as I have the patience to take them, so in that respect, they're ready to go.
I'd be forever indebted if someone can clarify what I've been achieving in terms of colorspace of JPEGs already produced, and how to achieve my desired nirvana going forward. When viewing my JPEG thumbnails in Bridge, BTW, the Metadata "File Properties" section indicates "Color Mode" as RGB, but doesn't even show a side caption for "Colorspace". If I open the JPEG in CS3 and click the choice of assigning sRGB colorspace and save, the resulting saved JPEG in Bridge does show Metadata "Colorspace" side caption, and sure enough indicates sRGB.
Big mystery, box is checked that says "Convert to sRGB", but CS3 and Metadata doesn't read it, pretty clear I haven't got it, eh?
Anyway, I'm accepting that I've probably been fooling myself on this for the past year, but time to face reality, revisit the question, and get educated. Many thanks in advance to dGrinners out there taking the time to set me straight. I don't claim any particular expertise, but have read a lot on CS3 and work hard at using just the upper regions of it (someday will get into the real detail, no time now), and IMO produce pretty good results. The JPEGs in my Flickr and Smugmug galleries (links below) are examples of JPEGs produced exactly as described, so maybe taking a look at them might tell you something I've maybe failed to make clear. The images there look OK to me, but if in reality they are not in sRGB colorspace, am I fooling myself, could they look better, am I lining myself up for some huge re-do job down the road if desiring to do so?
Again, many thanks in advance for taking the time.
After opening my TIF in CS3 and resizing, I "save for web and devices", I find that a fast and convenient way to fine tune JPEG size by using the slider in the "save for web and devices" pop up box. In that pop up, I check the box that says "Optimized", and in the menu that drops by clicking the button to right of the "Preset" box (which I default to "none"), I check "Convert to sRGB" and "Include XMP". And then I save. However, and this is where the old mystery reaserts itself, when I reopen the resulting JPEG in CS3, I up comes a pop up box that tells me image has no assigned colorspace, and gives me the options to leave it that way or check the colorspace I'd like assigned (sRGB being one of the options). Way back when, I came away with the understanding from here at dGrin that my JPEG is in sRGB colorspace, it just doesn't read that way when I reopen in CS3, and in my rush to get started, I accepted that. But now, thousands of images later, water under the bridge so to speak, it occurs to me that I finally ought to revisit the question so that, even if the past thousands are deficient, the future thousands will have a propoerly assigned sRGB colorspace. I use sRGB, BTW, because all at this point are for screen viewing, not print. I always can go back to my RGB TIFs if I want files later for print, already processed through CS3 (actually through ACR except for sizing and converting) as far as I have the patience to take them, so in that respect, they're ready to go.
I'd be forever indebted if someone can clarify what I've been achieving in terms of colorspace of JPEGs already produced, and how to achieve my desired nirvana going forward. When viewing my JPEG thumbnails in Bridge, BTW, the Metadata "File Properties" section indicates "Color Mode" as RGB, but doesn't even show a side caption for "Colorspace". If I open the JPEG in CS3 and click the choice of assigning sRGB colorspace and save, the resulting saved JPEG in Bridge does show Metadata "Colorspace" side caption, and sure enough indicates sRGB.
Big mystery, box is checked that says "Convert to sRGB", but CS3 and Metadata doesn't read it, pretty clear I haven't got it, eh?
Anyway, I'm accepting that I've probably been fooling myself on this for the past year, but time to face reality, revisit the question, and get educated. Many thanks in advance to dGrinners out there taking the time to set me straight. I don't claim any particular expertise, but have read a lot on CS3 and work hard at using just the upper regions of it (someday will get into the real detail, no time now), and IMO produce pretty good results. The JPEGs in my Flickr and Smugmug galleries (links below) are examples of JPEGs produced exactly as described, so maybe taking a look at them might tell you something I've maybe failed to make clear. The images there look OK to me, but if in reality they are not in sRGB colorspace, am I fooling myself, could they look better, am I lining myself up for some huge re-do job down the road if desiring to do so?
Again, many thanks in advance for taking the time.
See my work at http://www.flickr.com/photos/26525400@N04/sets/. Policy is to initially upload 10-20 images from each shoot, then a few from various of the in-process shoots each time I log on, until a shoot is completely uploaded.
0
Comments
I played around with this, and the only way I could get results similar to yours (jpeg document saved without color management) is not to have the ICC box checked in the main menu for "Save for Web and Devices". Do you have this checked?
I'll be! That seems to do it. You've saved a grown man a lot of heartache. Now, how come I never knew that, or why isn't the ICC check a default.
Many thanks. I'll make sure to have ICC checked from here on out.
I take it all this means my all earlier JPEGs have no colorspace assigned. If not, what are they in, what gamut? I understand basics, but obviously not enough here. Any suggestions for all those past JPEGs to show them best as possible on screen?
Thanks again. Huge help.
That is correct - they have no colorspace assigned. From your initial post, it seems they are all sRGB. You could open them in PP3, Edit->Assign Profile to change them to sRGB, and save them. You could even do this with an action.
Depending on your Color Settings, you might just be able to do this by opening and saving the files. You should set your RGB Color Management to Convert to Working Profile, and set Missing Profile to Don't Ask. You might run a test to make sure this works.
Hey John, you've hit the nail on the head with this response. My main concern is that the pics converted into JPEGs over the past year are in fact in sRGB colorspace, as driven by my option selection on the "Save for web and devices" pop up. So long as they're actually in sRBG for best viewing on Flickr, Smugmug and such, my main concern is met. I'm only secondarily concerned that CS3 can't recognize there in sRBG colorspace. The CS3 pop up asks if I want to assign sRGB, which I took to mean they were not already in sRGB, but you're saying they're just missing the tag to tell cS3 that they are. Which I think I can live with. I don't do home printing, zero interest in that, and any outside printer I might employ surely would want either my original RGB TIF, or something I easily can convert to from that. This seems also to confirm what I think I'm seeing on Flickr and Smugmug, which is anything but washed-out pics, which is my understanding of what RGB colorspace would give, the more extremely vibrant colors not in the RGB gamut. Think I have that logic right. If you don't mind giving another minute or two to this, are there holes you see in my understanding and thinkng?
Again, many thanks. Your responses have been a huge help.
I never use "Save for Web" that is a command strictly for folks producing images for web designers, graphic artists for web pages, web sites, etc.
I use Convert to Profile -> sRGB. That way I KNOW the profile has been attached properly to any file I plan to upload to the web.
Moderator of the Technique Forum and Finishing School on Dgrin
However, I've done some testing, and I think I was mistaken.
I took an image in sRGB working space and generated two jpegs using Save for Web and Devices - one with ICC on and one with ICC off (both had Convert to sRGB and Generate XMP set). Then I read both of them in, assigning sRGB to the one without a tagged color space. Putting one on top of the other (in layers) with difference mode, there were definitely differences between the two.
So I guess I don't know what PS does when you have Convert to sRGB on ICC unchecked. It doesn't seem to be the case that it generates the same image bits without tagging it.
That is interesting. I'll try the same experiment, but do a third, not check "convert to sRGB", which means result will be in the original RGB, and see how that compares to the others. Would you mind taking a few minutes and doing the same. I've a feeling you're much more discerning in you you see these things than me, so your observation undoubtredly would be more valid. Frankly, I'm not even sure I'll see a difference, hopefull will, but without the background you nave in finishing, mybe not.
I have no idea of what's considered a lot of hits in Smugmug, and presumably if finish was way sub-standard I wouldn't be getting many, but last time I looked, galleries have had over 16,000 hits over the 1 1/2 months I've posted there. So does that tell us something?
Look forward to hearing back, and many, many thanks for helping out.
I'm not sure what this means. The differences aren't great (when I do a difference of the two, I can't see them - it takes an AutoLevels on the difference to see them faintly). The odd thing is that when you check or uncheck ICC, it regenerates the optimized JPEG (I have a fairly large file, and I see the progress bar for optimizing across the bottom).
So I was guessing that the differences are not due to different color spaces (there isn't another color space handy, since everything is set up for sRGB), but rather some small differences in the JPEGs.
Then I did the same experiments with the Granger rainbow, hoping to verify this. This time, the differences were much more noticable, and in ways that suggested a difference in color space.
So the bottom line is I don't know what color space your earlier images are in. You say they seem to look OK on SmugMug, and there's nothing obvious to do to "fix" them, color-space-wise anyway. So I'd just be happy with the older images, and follow Pathfinder's advice for the future.
Hey, John, don't know how to thank you. I tried same progression as you, and sure enough, my eyes aren't trained well enough to tell the difference. I will follow your advice, should serve me well, although checking the ICC and sRGB boxes in "Save for web and devices" seems to do the same thing, and allows me fine tuning on size I find helpful. Again, dGrin saves the day.
I've got a whopper coming up. Was shooting with tripod indoors at ISO 100, had to increase to 200 and 400 for a small number of shots, but when dialed back to 100 I inadvertently dialed one click too far, and ended up at ISO 3200, where I shot for the next two hours before noticing it. Obviously way beyond anything my Oly E-3 can legitimately handle. I know, stupid rookie mistake, shoot that long without checking settings, but I'm no rookie, so that leaves just plain stupid. Anyway, shoot involved rare access to a Maserati/Ferrari dealership and expanded showroom, lousy backgrounds, but beautiful autos! In processing results, I'm finding remaining noise not that objectionable viewed on screen, subjects are all hard metal shapes and forms, milled and chromed and painted, focus is sharp, so maybe something redeemable. As soon as I get a few more processed, I'll start a thread here attaching a few, see if what I'm seeing is wishful thinking, or actually an alternative look to go for in the right situations. Stay tuned ...
My normal working space is Adobe RGB, so I had to change Color Settings to do the experiments, and then change them back. At times I wasn't careful enough in this. I also forgot one crucial step when enhancing the difference between two layers, and I did the Auto Levels on the wrong thing. Talk about rookie mistakes.
Anyway, for both images I tested (including the Granger Rainbow), there is, indeed, no difference whatsoever between the images with ICC on and ICC off. So my original guess was correct, and your untagged images are indeed sRGB.
Sorry for the confusion.
You know, that's something I find here a dGrin I don't see everywhere. You guys are really interested! If I could reach some level of expertise, be patient enough long enough to really gain traction in selected portions of PS, I'd love to do the same thing.
Thanks for the clarification. As hard as I tried to discern a difference, I couldn't, so your re-check confirms what I'm seeing. Many thanks.
I'll put up something soon demonstrating my ISO 3200 experience. For a guy who's never shot over 800 in his life, and at that only rarely, this is like earth-to-mars, never seen such clatter. But it's indeed a "look", something all it's own, not as pretty as grain, but actually kind of interesting in all this metal I shot. Will be interesting to hear comments. Sure they'll range from "lots of noise" to "interesting effect", really not much succor, but any port in a storm ...
Just an OCD-oriented
Since the nature of Web photos has changed, Save for Web has also changed. Apparently enough photogs complained "hey, my web images are for online galleries and need profiles and metadata" that Save for Web now has enough options for us Smugmug users as a one-stop export. In CS4 Save for Web there is a Convert to sRGB option, there is an Embed Color Profile option, and there is a popup with different metadata inclusion/exclusion levels in it. That's getting pretty far from "it's just for web designers."
Now to all, going to back to post #1 it is very important to understand that those three options are distinctly different things:
Convert to sRGB only performs a conversion - nothing to do with embedding.
Embed Color Profile only embeds the profile - nothing to do with converting (if it's in Adobe RGB, embedding embeds the Adobe RGB profile with no conversion unless you also checked the Convert to sRGB box).
Image metadata (in SFW) has nothing at all to do with conversion or profile - just EXIF, copyright, keywords, captions, etc. which are also very important in a Smugmug upload workflow.
In Save for Web, for Smugmug you want Convert to sRGB on, Embed Color Profile on (if that's what you want), and you set the metadata pop-up depending on how much image info you want to make publicly visible. But for color the most important option is the Convert to sRGB checkbox.
All that said, I'm still using an action that does Convert to Profile (sRGB) separately and does a Save As JPEG, nothing wrong with that workflow either unless you want to restrict the range of embedded metadata.
Its sticky, its default is to embed.
Of course you could use the Convert to Profile command and simply not save the file. No need to warn folks of that right?
As for embedding the profile, IF they are destined for the Web, there are only two web browsers that will even recognize the embedded profile. Since one would assume most people using Save for Web are actually uploading the files to the web, its questionable if the default for the check box should be on or off, but again, like the Save dialog, its sticky. Once set, its not going to change without user intervention.
Author "Color Management for Photographers"
http://www.digitaldog.net/