darbyjohnston / DJV

Professional media review software for VFX, animation, and film production
https://darbyjohnston.github.io/DJV/
BSD 3-Clause "New" or "Revised" License
688 stars 59 forks source link

12 bit & 16 bit DPX support #48

Open darbyjohnston opened 5 years ago

darbyjohnston commented 5 years ago

Original ticket: https://sourceforge.net/p/djv/feature-requests/64/

nawadoln commented 4 years ago

Hi Darby,

I provided you 3 files 10bit, 12bit and 16bit. If you have Nuke or other image sequence player they can open it but DJV can't open the 12 bit file at the moment.

links: https://we.tl/t-uOJJQTYTq3

cheers

darbyjohnston commented 4 years ago

Great, thanks for the sample files, that's very helpful!

darbyjohnston commented 4 years ago

Hi, out of curiosity when you open the 12-bit files in the other applications, are you actually working with 12-bit data or does the application scale it up to 16-bits?

nawadoln commented 4 years ago

Well I'm not sure what the software does in the backend. But I guess No because when zoom in detail, grain still the same no change unless it has a way to convert data but not affect visually.

Thanks

On Tue, Sep 29, 2020, 6:22 AM Darby Johnston notifications@github.com wrote:

Hi, out of curiosity when you open the 12-bit files in the other applications, are you actually working with 12-bit data or does the application scale it up to 16-bits?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/darbyjohnston/DJV/issues/48#issuecomment-700334134, or unsubscribe https://github.com/notifications/unsubscribe-auth/ARDDYB5R3K5P3HRIRNICHKDSIELEJANCNFSM4ITJQKRQ .

darbyjohnston commented 4 years ago

I assume that internally it scales the data up to 16-bits which would take more memory but not affect the quality. Do the applications show the actual pixel values when picking colors (0-4095) or a floating-point number (0.0-1.0)?

I'm curious because converting to 16-bits would be the easiest way to support this, but as I mentioned it does take more memory and the DJV color picker would show the values as either 16-bit (0-65535) or floating-point which might be an issue.

nawadoln commented 4 years ago

Hi Darby, I think you're right, the sample pixel values shown as 0-1, maybe it is converting to either 10 or even 8 bit in the background. [image: image.png]

On Wed, Sep 30, 2020 at 12:59 AM Darby Johnston notifications@github.com wrote:

I assume that internally it scales the data up to 16-bits which would take more memory but not affect the quality. Do the applications show the actual pixel values when picking colors (0-4095) or a floating-point number (0.0-1.0)?

I'm curious because converting to 16-bits would be the easiest way to support this, but as I mentioned it does take more memory and the DJV color picker would show the values as either 16-bit (0-65535) or floating-point which might be an issue.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/darbyjohnston/DJV/issues/48#issuecomment-700882490, or unsubscribe https://github.com/notifications/unsubscribe-auth/ARDDYB2TSSVHJZIB5D3PPT3SIIN6ZANCNFSM4ITJQKRQ .

-- Best Regards Nawadoln Siributr