Yeah, when zoomed out, Nuke can skip scanlines, so ZIPS means it can do
less work to read say every other scan line, at least when the viewer's
driving the op tree...



On Thu, 5 Jan 2017 at 14:29, Deke Kincaid <[email protected]> wrote:

> Nuke has always been faster with zips.  Zip 16 lines files were always
> considerably slower.
>
> On Wed, Jan 4, 2017 at 3:53 PM, Larry Gritz <[email protected]> wrote:
>
> Out of curiosity, Piotr, do you recall what was the original reason for
> zip vs zips? Under what conditions would one prefer zipping the scanlines
> individually? Was it just for the sake of making it easier to parallelize
> the compression among threads?
>
>
> On Jan 4, 2017, at 3:42 PM, Piotr Stanczyk <[email protected]>
> wrote:
>
> Hi Zachary,
>
> The difference between ZIP and ZIPS is that ZIPS compresses a single line
> at a time whereas ZIP does it in chunks.
>
> For reference, http://openexr.com/TechnicalIntroduction.pdf,  pp16
>
>
> On 4 January 2017 at 15:16, Larry Gritz <[email protected]> wrote:
>
> In EXR speak, "zip" is the usual deflate/zip compression in blocks of 16
> scanlines, and "zips" is the same compression method, but applied to each
> scanline individually. Both are supported by OIIO.
>
> The compression/decompression is handled on the OpenEXR library side, not
> by the OIIO code itself, so in theory anything should be fine with either
> one. Though I do see a comment in my code indicating that only "zips" has
> been found to be reliable for deep files.
>
> Can you give more detail about what's going wrong? You're rendering out an
> exr with "zips"? Writing it with OIIO? Then it's "borken" when you
> "process" them... in what way is it broken, how processed, in what program
> does it appear broken? (Is that when also reading with OIIO, or some other
> app?)
>
> Which versions of OIIO and OpenEXR are you using?
>
> Can you either send me one of these broken files to inspect, or else give
> a recipe for how to create one (ideally using only OIIO components)?
>
>
> On Jan 4, 2017, at 2:28 PM, Zachary Bauer <[email protected]>
> wrote:
>
> Hello,
>
>
>   A question about OIIO compression support:
>
> I am having redshift output Multi-Part EXRs and their headers say
>
> their compression is "zips". Typically what I see is "zip" compression
>
> for my EXRs from other renderers (PRMAN). Processing these redshift
>
> images using Open Image IO is resulting in broken images, but if I
>
> convert them to "zip" compression using oiiotool first, everything seems
>
> to work ok.
>
>
>
>
>
> Is "zips" compression different than "zip"?
> Does OIIO support "zips"?
>
> Thanks,
> Zachary Bauer
> Senior Software Engineer (Tools)
> Blizzard Entertainment
>
> _______________________________________________
> Oiio-dev mailing list
> [email protected]
> http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
>
>
>
>
> --
> Larry Gritz
> [email protected]
>
>
>
>
>
>
>
>
>
> _______________________________________________
>
>
> Oiio-dev mailing list
>
>
> [email protected]
>
>
> http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
>
>
>
>
>
>
> _______________________________________________
> Oiio-dev mailing list
> [email protected]
> http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
>
>
>
>
> --
> Larry Gritz
> [email protected]
>
>
>
>
>
>
>
>
>
> _______________________________________________
>
>
> Oiio-dev mailing list
>
>
> [email protected]
>
>
> http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
>
>
>
>
>
>
> _______________________________________________
>
> Oiio-dev mailing list
>
> [email protected]
>
> http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
>
>
_______________________________________________
Oiio-dev mailing list
[email protected]
http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org

Reply via email to