Re: [Gimp-developer] JPEG quality factor - some remaining odds and ends
Liam R E Quin wrote: > On Tue, 2010-01-19 at 22:38 +0100, yahvuu wrote: > [...] > >> II. Range of actually useful values for IJG quality value >> >> For GIMP's target users less than half of all possible settings >> are useful: > possibly - I've often used values as low as 35% or sometimes lower. > "The sweet spot" depends hugely on your image and your purpose - > consider providing a "lowsrc" alternate image for low bandwidth > Web users for example, or a thumnail. good point -- that proves wrong my claim that there were only a handful of useful quality settings. Also, further discussion about those 5 values >95 becomes moot, i guess.. regards, yavuu ___ Gimp-developer mailing list Gimp-developer@lists.XCF.Berkeley.EDU https://lists.XCF.Berkeley.EDU/mailman/listinfo/gimp-developer
Re: [Gimp-developer] JPEG quality factor - some remaining odds and ends
> Once a user starts to use jpeg they have to decide what to do with > "quality" setting. Bigger number = better quality is not too hard to get > your head around. A bit of experimenting quickly reveals what works best > for a particular task. > > You quickly realise what ranges don't fit your needs and focus on those > that do. End of story. > > I see no use what so ever in creating some new, grouped setting in its > place. This would essentially involve exactly the same learning process > and reduce control and compromise results. What could be an option is to have named presets. Right now, you can save the JPG settings for later use in exactly one preset. Depending on the task, you may want to have different presets, so naming the presets and being able to recall them later when you do a similar task could perhaps be quite useful. In such a case, Gimp could also ship with 2-3 default JPG presets which are optimized for a specific task, like photo lab image, web image, etc. to make it easier for the first-time (or occasional) Gimp user to get started with the JPG settings. Torsten signature.asc Description: This is a digitally signed message part. ___ Gimp-developer mailing list Gimp-developer@lists.XCF.Berkeley.EDU https://lists.XCF.Berkeley.EDU/mailman/listinfo/gimp-developer
Re: [Gimp-developer] JPEG quality factor - some remaining odds and ends
On 01/19/10 22:51, Liam R E Quin wrote: > On Tue, 2010-01-19 at 22:38 +0100, yahvuu wrote: > [...] > >> II. Range of actually useful values for IJG quality value >> >> For GIMP's target users less than half of all possible settings >> are useful: > possibly - I've often used values as low as 35% or sometimes lower. > "The sweet spot" depends hugely on your image and your purpose - > consider providing a "lowsrc" alternate image for low bandwidth > Web users for example, or a thumnail. > > Most of the preview images on www.fromoldbooks.org are saved at 75% > (usually with "smoothing" to reduce artifacts a little) > > >> 95 >>. no-go: just wastes disk space -- ever heard of XCF? >> 100 > > Actually I use 97% a lot, and 100% too -- because I want jpeg format, > not some application-specific thing that won't work for most users. > Export is about interchange, the end product, you shouldn't ever use > jpg for a file you're going to edit again, and you shouldn't normally > use xcf for interchange unless you know they're using (a compatible > version of) GIMP... Once a user starts to use jpeg they have to decide what to do with "quality" setting. Bigger number = better quality is not too hard to get your head around. A bit of experimenting quickly reveals what works best for a particular task. You quickly realise what ranges don't fit your needs and focus on those that do. End of story. I see no use what so ever in creating some new, grouped setting in its place. This would essentially involve exactly the same learning process and reduce control and compromise results. > > >> III. Parameter Triaging >> >> The "Subsampling" parameter is more important than its current >> position inside the "advanced parameters" section suggests. > Yes - in particular it affects colours, especially reds. > > Liam > > I agree , this setting could come out be more visible. It's annoying having to dig in there just to check what it is set at. /gg ___ Gimp-developer mailing list Gimp-developer@lists.XCF.Berkeley.EDU https://lists.XCF.Berkeley.EDU/mailman/listinfo/gimp-developer
Re: [Gimp-developer] JPEG quality factor - some remaining odds and ends
On Tue, Jan 19, 2010 at 10:38:40PM +0100, yahvuu wrote: > Hi all, > > recent discussion on gimp-user brought up some usuability issues > of the JPEG export dialog [1]. Actually, there's nothing new to say > about it... the big "JPEG quality" thread did cover it all [2]. > However, due to the sheer size of that thread, i think a summary of > open issues is useful. > Can't recall if there was ever a discussion there about the possibility of entering a starting desired file-size in the dialog, whereupon the slider would be adjusted, from where it could be tweaked up or down. Having limited storage space on my server, I am always doing the quality/size compromise. Would be handy if the default starting-point could be adjusted to size rather than quality. Scott. ___ Gimp-developer mailing list Gimp-developer@lists.XCF.Berkeley.EDU https://lists.XCF.Berkeley.EDU/mailman/listinfo/gimp-developer
Re: [Gimp-developer] JPEG quality factor - some remaining odds and ends
On Tue, 19 Jan 2010 22:38:40 +0100 yahvuu wrote: > II. Range of actually useful values for IJG quality value > > For GIMP's target users less than half of all possible settings are > useful: > http://yahvuu.files.wordpress.com/2009/08/ijgqualityrange.png > or, in ASCII: > > 0 > . > . no-go: blocky garbage > . > 60 > . if in dire need > 80 > 90 the sweet spot > 95 > . no-go: just wastes disk space -- ever heard of XCF? > 100 Sorry. I use 100 because the photo labs that I use for online printing only accept image upload in jpg format. 95 may be good enough, but for 80x60cm prints, I don't want good enough. -- Jon Senior ___ Gimp-developer mailing list Gimp-developer@lists.XCF.Berkeley.EDU https://lists.XCF.Berkeley.EDU/mailman/listinfo/gimp-developer
Re: [Gimp-developer] JPEG quality factor - some remaining odds and ends
On Tue, 2010-01-19 at 22:38 +0100, yahvuu wrote: [...] > II. Range of actually useful values for IJG quality value > > For GIMP's target users less than half of all possible settings > are useful: possibly - I've often used values as low as 35% or sometimes lower. "The sweet spot" depends hugely on your image and your purpose - consider providing a "lowsrc" alternate image for low bandwidth Web users for example, or a thumnail. Most of the preview images on www.fromoldbooks.org are saved at 75% (usually with "smoothing" to reduce artifacts a little) > 95 > . no-go: just wastes disk space -- ever heard of XCF? > 100 Actually I use 97% a lot, and 100% too -- because I want jpeg format, not some application-specific thing that won't work for most users. Export is about interchange, the end product, you shouldn't ever use jpg for a file you're going to edit again, and you shouldn't normally use xcf for interchange unless you know they're using (a compatible version of) GIMP... > III. Parameter Triaging > > The "Subsampling" parameter is more important than its current > position inside the "advanced parameters" section suggests. Yes - in particular it affects colours, especially reds. Liam -- Liam Quin - XML Activity Lead, W3C, http://www.w3.org/People/Quin/ Pictures from old books: http://fromoldbooks.org/ Ankh: irc.sorcery.net irc.gnome.org www.advogato.org ___ Gimp-developer mailing list Gimp-developer@lists.XCF.Berkeley.EDU https://lists.XCF.Berkeley.EDU/mailman/listinfo/gimp-developer
[Gimp-developer] JPEG quality factor - some remaining odds and ends
Hi all, recent discussion on gimp-user brought up some usuability issues of the JPEG export dialog [1]. Actually, there's nothing new to say about it... the big "JPEG quality" thread did cover it all [2]. However, due to the sheer size of that thread, i think a summary of open issues is useful. I. Displayed value range for the quality slider It's surprisingly delicate to choose an innocuous range for that slider: - 0..100 is an invitation for confusion: 100 = 100% quality = 100% information?!? - using the same range as another application makes it easy to think those settings were comparable between the applications. (Which they usually are not, for various reasons). Photoshop uses 0.. 12 for "save as JPEG" 0..100 for "save for web" Lightroom: 0..100 - starting at 0 feels odd: zero quality = nothing at all = no image?!? - other don'ts: photoshop allows slider stops between numbers, but doesn't display the corresponding values [3] II. Range of actually useful values for IJG quality value For GIMP's target users less than half of all possible settings are useful: http://yahvuu.files.wordpress.com/2009/08/ijgqualityrange.png or, in ASCII: 0 . . no-go: blocky garbage . 60 . if in dire need 80 90 the sweet spot 95 . no-go: just wastes disk space -- ever heard of XCF? 100 The IJG library takes an integer for the quality value, so there are only about 15-35 distinct useful settings for the quality slider. (don't beat me on the exact numbers, i'm willing to provide suitable image comparisons when/if these numbers become important) III. Parameter Triaging The "Subsampling" parameter is more important than its current position inside the "advanced parameters" section suggests. Photoshop is the reference here -- it automatically switches from 2x2 to 1x1 subsampling for the higher quality settings [4]. regards, yahvuu [1] https://lists.xcf.berkeley.edu/lists/gimp-user/2010-January/016436.html [2] http://marc.info/?l=gimp-developer&m=118377279721823&w=2 [3] https://lists.xcf.berkeley.edu/lists/gimp-user/2010-January/016493.html [4] http://blogs.gnome.org/raphael/2007/10/23/mapping-jpeg-compression-levels-between-adobe-photoshop-and-gimp-24/ ___ Gimp-developer mailing list Gimp-developer@lists.XCF.Berkeley.EDU https://lists.XCF.Berkeley.EDU/mailman/listinfo/gimp-developer
