Komodo not using jpeg2000 wavelet, but "normal" motionJPEG?

[AndreeMarkefors]

I didn't have time to watch the whole live stream.

It was mentioned that Komodo's r3d aren't exactly the same format as we're used to.

I understood it as Komodo is using legacy DCT motionJPEG for compression and not jpeg2000 wavelet that's been in RED cameras until now. This may be down to internal components being more tuned to doing fast motionJPEG compression?

I've been happy with the motionJPEG in my Canon 1Dc and 1Dx mkII (few issues or surprises).

But was this brought up again for discussion later? Phil quickly mentioned that he might talk about it, and what it would mean.

 

[David Rasberry]

I didn't have time to watch the whole live stream.

It was mentioned that Komodo's r3d aren't exactly the same format as we're used to.

I understood it as Komodo is using legacy DCT motionJPEG for compression and not jpeg2000 wavelet that's been in RED cameras until now. This may be down to internal components being more tuned to doing fast motionJPEG compression?

I've been happy with the motionJPEG in my Canon 1Dc and 1Dx mkII (few issues or surprises).

But was this brought up again for discussion later? Phil quickly mentioned that he might talk about it, and what it would mean.

I doubt Red would move to an inferior DCT compression scheme over wavelet. More likely with the massive improvements in available processing power they would move to something like processing larger blocks of pixels to reduce quantization noise. Redcode is wavelet, but it is not JPEG 2000 or any form of standard JPEG since those formats are either RGB or YUV or YCbCr, not Bayer raw. Not to say they would not use some elements of the JPEG wavelet compression schemes for discretely compressing the Bayer pattern without exporting an encoded file. But they are probably compressing luminance data only for the three filter groups of a Bayer pattern without reference to color since that is defined in the Debayer process. The recorded raw file would contain pixel luminance data only as with uncompressed raw.

 

[AndreeMarkefors]

I doubt Red would move to an inferior DCT compression scheme over wavelet.

I went back and rewatched that segment (around 2h18) and after a little bit of an awkward pause Phil says he'll talk about it but that "no one has talked about it yet". He also didn't seem worried and said that pixel quality was good.

 

[Sylvain Morjane]

I went back and rewatched that segment (around 2h18) and after a little bit of an awkward pause Phil says he'll talk about it but that "no one has talked about it yet". He also didn't seem worried and said that pixel quality was good.

I'm also being intrigued by that segment. Felt like that was something they weren't really allowed to talk about. Anyway, camera seems promising so let's no worry to much.