Re: cOMp chunk [was: EXIF support in PNG]
Willem van Schaik <[email protected]>
| Newsgroups | gmane.comp.graphics.png.general |
|---|---|
| Message-ID | <[email protected]> |
just to let you all know that I will vote against a COMP chunk for me it is adding complexity for very little gain sorry, Willem On 2017-01-01 12:01, John Bowler wrote: > On Sun, Jan 1, 2017 at 10:49 AM, Glenn Randers-Pehrson > <[email protected] <mailto:[email protected]>> wrote: > > What I'm proposing is a set of chunks which the application > never sees; if I ask the encoder will write an sPLT as cOMP/sPLT > but I never seer that; I give the encoder an sPLT. When an app > reads this using a decoder which supports cOMP the decoder > reports an sPLT, not a cOMP. > > > You cannot count on applications not seeing the chunk. > You can, however, count on the fact that some applications won't > recognize it, and will report cOMP but not the embedded sPLT. > > > I'm still proposing something that doesn't *require* any support in the > application, only the decoder and encoder. > > > I'm talking about a cOMP chunk that contains a chunk that doesn't > have its own compression, either because the chunk > spec doesn't allow it or because the author chose not to compress > it. The cOMP chunk should contain a compression_method byte in order to > future-proof the chunk against improved compression methods that > might come along, just as IHDR, zTXt, etc. do. > > > Well, yes; I was talking about adding keywords. I'm certainly > interested in the possibility of allowing different compression methods > (e.g. RLE) though it would slow down adoption. > > The chunk I'm talking about is just a piece of math that transparently > results in a smaller PNG from the application point of view. It's a > completely reversible lossless transform that *can* be implemented > invisibly within the current specification. It it carries extra > information there will be bug reports from someone, somewhere that the > information is lost by some application and then the chunk ceases to be > transparent. > > Or, to put it another way, when (20 years time?) all PNG *decoders* > support cOMp/cOMP/COMP/COMp I can find quite a few PNGs and, completely, > totally, losslessly make them smaller, and no app will know unless it asks. > > John Bowler > > > ------------------------------------------------------------------------------ > Check out the vibrant tech community on one of the world's most > engaging tech sites, SlashDot.org! http://sdm.link/slashdot > > > > _______________________________________________ > png-mng-misc mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/png-mng-misc > -- Willem van Schaik <[email protected]> http://www.schaik.com/ ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, SlashDot.org! http://sdm.link/slashdot