VRIMG compression?

[simmsimaging]

I'm finding vrimg files are still 3X larger than the same file converted to exr, but I seem to remember that they should have compression now. I use the vrimg2exr converter and I think the default conversion creates a 16bit linear exr - would that be enough to account for the size difference?

/b

[Shimakaze]

Yes, the difference between 16 bit and 32 bit is pretty huge when it comes to filesize, though not so big when it comes to quality. I recommend putting the vrimg files somewhere safe and use the converted EXR images in normal production. You can of course render directly to EXR but I've had too many renders get corrupted from that to advice it.

[simmsimaging]

Thanks! I know what you mean about the corruption - that's why I switched to using vrimg but I would much prefer exr if only to save the time spent always converting to layered exrs for Photoshop as well as taking advantage of that size difference. 16 bit float has always been sufficient for me.Thanks again also for the vrimg2exr tool - a very handy time saver!b

[3DMK]

Yes, the difference between 16 bit and 32 bit is pretty huge when it comes to filesize, though not so big when it comes to quality. I recommend putting the vrimg files somewhere safe and use the converted EXR images in normal production. You can of course render directly to EXR but I've had too many renders get corrupted from that to advice it.

I usually save to exr
what corruption issue is there??

[simmsimaging]

There are a lot of problems with rendering and saving .exr to network locations (i.e when rendering via backburner) with large files. If your renders are small enough you may not run into it.

[joconnell]

Odd - I usually get the opposite. I render to default vrimg, not sure if it's 16 or 32 bit, and then when converting use the command line tool with "-compression zips" and it makes smaller files.

[simmsimaging]

Unless I misunderstood you, sounds like you have the same thing: on coversion to exr the files are much smaller than the original vrimgs. It could be partly the compression, but I think it's probably also that it goes from 32 to 16bit float.