Comments on: Color Spaces https://webkit.org/blog/73/color-spaces/ Open Source Web Browser Engine Thu, 19 Nov 2015 21:42:00 +0000 hourly 1 https://wordpress.org/?v=6.5.2 By: cmacaskill https://webkit.org/blog/73/color-spaces/comment-page-1/#comment-17941 Wed, 14 Feb 2007 23:57:01 +0000 http://webkit.org/blog/?p=73#comment-17941 I’m a huge Mac user and fan who was at NeXT during the let’s do a gamma of 1.8 days, back when monitors were dimmer and CMYK was the thing.

I don’t think the color mismatches are a Safari issue, per se, but I did pour my heart into this blog post that make the front page of Digg:

http://blogs.smugmug.com/onethumb/2007/02/14/this-is-your-mac-on-drugs/

It seems to me it’d take the pressure off the Safari team if OS X shipped in the config that Apple suggests its users switch to — a gamma of 2.2. Tell me where I’m wrong.

Thanks,
Chris

]]>
By: jfriedl https://webkit.org/blog/73/color-spaces/comment-page-1/#comment-14890 Mon, 18 Dec 2006 11:10:37 +0000 http://webkit.org/blog/?p=73#comment-14890 Hi Dave,
Understanding the problem with treating unprofiled images differently than Flash/CSS, let me ask about images for which the color space is tagged in the Exif? An image can be tagged as being sRGB with the ColorSpace tag, or as being AdobeRGB with the InteroperablityIndex tag. This is how most digital cameras communicate the color space of their image. Web-site-adorenment images (ones that might need to match with flash) generally don’t have these tags, so it seems that you could take a big step forward by recognizing these tags. It would require that colorsync have sRGB and AdobeRGB profiles on hand, but don’t they already?

The up side to this is that images straight from the camera now are processed properly. I don’t see any downside with what you mentioned (Flash/CSS).

It’s not a complete resolution of the issue, but it’s a huge win-win step.

I hope you’ll consider it.
Jeffrey
(Author of the article you cited in your original post)

]]>
By: Manuel Silva https://webkit.org/blog/73/color-spaces/comment-page-1/#comment-12431 Sun, 29 Oct 2006 09:31:32 +0000 http://webkit.org/blog/?p=73#comment-12431 Mark and others:

Sorry, excuse.

Is possible suggest a new topic ?

Safari, digital certificates and java policy.

I have some questios about it.

Thanks

Manuel

]]>
By: jacobolus https://webkit.org/blog/73/color-spaces/comment-page-1/#comment-12430 Sun, 29 Oct 2006 07:37:27 +0000 http://webkit.org/blog/?p=73#comment-12430 jfriedl:

> Why do you say this? Is it because some some small percent of web authors might do image-to-background blending? (I’m not saying that’s a bad reason — heck, I do image-to-background blending — I ask because I want to understand your reasons).

Yes, this was indeed the reason claimed in my discussion with hyatt and others on IRC.

> As for CSS, as has already been mentioned, that’s defined to be sRGB, so should be translated from that to the local monitor/printer space.

The impression I got was that they are going to try to move towards this as time goes on.

> So that leaves Flash. It sounds as if you’ve made 99%+ of the images on the web render incorrectly so that Flash authors can have perfect color matching with standards-broken non-sRGB CSS colors? Or did I misunderstand something in your post?

Yep, that’s right. I think the idea was that web designers would rebel if their color-matched flash no longer matched.

]]>
By: Mark Rowe https://webkit.org/blog/73/color-spaces/comment-page-1/#comment-12418 Thu, 26 Oct 2006 10:15:39 +0000 http://webkit.org/blog/?p=73#comment-12418 Manuel, if you have a bug to report in Safari the best way to get it addressed is to report it via the appropriate channels rather than in comments on a completely unrelated blog post. Bugs in Safari should be reported at the Apple bug reporter, and bugs in WebKit should be reported in Bugzilla. Thanks!

]]>
By: Manuel Silva https://webkit.org/blog/73/color-spaces/comment-page-1/#comment-12417 Wed, 25 Oct 2006 22:32:34 +0000 http://webkit.org/blog/?p=73#comment-12417 Spiff:
Thank you for your reply, I really do appreciate it!
Your explanation is absolutely correct.
More, strangely, or not, I have the private key in Keychain !
I think the private key is created when renew the certificate.

The problem in Safari is about download certificates. When I collect the certificate !
And this is a problem in some CA. I tried in some CA internation and d`ont work. The problem is the Safari not open the window -DOWNLOAD, and the content of the certificate (cert. and public key) it is inside of Safari -like this page, but above, mixed with this posts.

If the process is functioned, run, the window -Download, open, download the certificate for desktop. This does not happen. The certificate it is inside of Safari, in the main window, mixed with the content of the page.

If you or other will be interested I I can send a print secreen.

My intention is that somebody tells this problem of Mr. HYATT and other developers for fix this problem, bug ?

Why Safari d`ont download, transferred, the certificates for descktop ?

Best regards

Manuel

]]>
By: jfriedl https://webkit.org/blog/73/color-spaces/comment-page-1/#comment-12415 Wed, 25 Oct 2006 01:39:37 +0000 http://webkit.org/blog/?p=73#comment-12415 Hi. I’m the author of the post to which you responded. I appreciate your response.

> First of all, if you correct unprofiled images to sRGB, you have to correct all drawing to sRGB

Why do you say this? Is it because some some small percent of web authors might do image-to-background blending? (I’m not saying that’s a bad reason — heck, I do image-to-background blending — I ask because I want to understand your reasons).

As for CSS, as has already been mentioned, that’s defined to be sRGB, so should be translated from that to the local monitor/printer space. I can’t imagein a few translations are very CPU intensive, so that’s not really an issue….

So that leaves Flash. It sounds as if you’ve made 99%+ of the images on the web render incorrectly so that Flash authors can have perfect color matching with standards-broken non-sRGB CSS colors? Or did I misunderstand something in your post?

I should be clear that I know little about Flash except that I usually seem to dislike how it’s used, so I have little care about how it renders, and I certainly don’t think that everything should be dumbed down to its level just because it’s poorly designed.

I should also admit that I made up the “99%+” number (it’s probably a low estimate)

Finally, I’m sorry that you considered my criticizims of Safari’s color management to be “pot shots.” I have perhaps harsher words for Apple than for Microsoft because I don’t expcect cluefullness from Microsoft (and they don’t disappoint). If there’s anyone I would expcect to get this right it’d be Safari and Firefox, and it’s disappointing to see that’s not the case.

From what I understand from your post, I don’t agree at all with your reasons, but at least it’s good to know that there are reasons. Why, though, was the ability to set a default color space for unprofiled images removed from ColorSync?

Jeffrey

]]>
By: Spiff https://webkit.org/blog/73/color-spaces/comment-page-1/#comment-12414 Tue, 24 Oct 2006 18:51:34 +0000 http://webkit.org/blog/?p=73#comment-12414 Manuel, in your post I see what I assume to be a user cert (user.crt), the cert from the intermediate CA (ca.crt) that signed/issued the user cert , and the cert from the root CA (root.crt) that signed/issued the intermediate CA’s cert. The one thing that P12 files usually (always?) have, that I don’t see in your post, is the private key that goes with the public key in your user cert. It’s my understanding that the thing that makes a P12 file special is that it contains not only your user certificate (and optionally the trust chain for your certificate) but also the private key that goes along with the public key in your user certificate.

Before you can make progress on this, you probably need to go find your private key — the one you generated (or was generated for you) when you created your certificate signing request.

]]>
By: macmoz https://webkit.org/blog/73/color-spaces/comment-page-1/#comment-12413 Tue, 24 Oct 2006 13:43:29 +0000 http://webkit.org/blog/?p=73#comment-12413 Thank you so much, jacobolus. Because I’m more in the interface-design, gifs and pngs should perfectly blend with css-colors. That wasn’t the case in Safari (I worked without a colorprofile in RGB in Photoshop). So option 3 would be the best for me, convert to (or work in) sRGB, and than strip the profile.

But, how do I strip the profile after generating the GIF or PNG? With Photoshop or imagemagick? And do i have to use this ‘convert +gamma 0’, as Jeff mentioned?

]]>
By: jacobolus https://webkit.org/blog/73/color-spaces/comment-page-1/#comment-12411 Tue, 24 Oct 2006 04:10:30 +0000 http://webkit.org/blog/?p=73#comment-12411 macmoz:

As far as whether to include the profile or strip it: that depends on what your goal is. If you want images which look the same on all platforms, and have vibrant colors, you should convert your image to sRGB and include the profile, and then only Mac gecko browsers will suffer.

If you want images which look the same on all mac browsers but could care less about windows/linux, you should convert the image to the profile of a generic Mac monitor. I’m not positive, but I think that the “Apple RGB” profile is close enough, and include the profile. This way at least clueful windows clients (if any ever show up) will show your image correctly.

If you want images which match your css colors, then you should convert to sRGB, and then strip the profile. Unfortunately they’ll look kind of unsaturated on the Mac, but you can get all the colors on your site consistent that way.

For an example of the kind of “unsaturated on the Mac” I’m talking about, see: http://www.flickr.com/photos/tunaranch/269794437/

But there really is no good solution at the moment. Hopefully if Mac gecko browsers get their act together, then option 1 will work out well enough. Hopefully Safari will also follow the css/svg/etc. specs, and then, as long as everything stays sRGB, we should have consistency across the board.

]]>