BRAW in Resolve playback performance

The compressed constant bit rate versions perform very well. The two variable compression, constant quality, versions, not so well. A fair bit of skipping even though the playback frame rate indicator shows a constant 24fps. These were 23.976 4.6k clips in a 24fps timeline and that may be the problem. But it could also be a question of fine tuning.

The compressed constant bit rate versions perform very well. The two variable compression, constant quality, versions, not so well. A fair bit of skipping even though the playback frame rate indicator shows a constant 24fps. These were 23.976 4.6k clips in a 24fps timeline and that may be the problem. But it could also be a question of fine tuning.

Are you reading the file from a designated NVMe SSD or from the ARECA?

Misha, the Areca is way fast enough to playback this stuff. The clips play fine from the media pool page. It's when they're downrezzed into a 4096X1716 timeline that the playback is jerky. That's not the case with 4.6k CDNGs so I assume it's a Resolve tuning problem.

(I also found that the constant compression clips have the same problem)

The good news is that Q:0 keys as brilliantly as one would expect.

Misha, the Areca is way fast enough to playback this stuff. The clips play fine from the media pool page. It's when they're downrezzed into a 4096X1716 timeline that the playback is jerky. That's not the case with 4.6k CDNGs so I assume it's a Resolve tuning problem.

(I also found that the constant compression clips have the same problem)

Bandwidth wise it should be no problem, how about I/O-wise?

Misha, the Areca is way fast enough to playback this stuff. The clips play fine from the media pool page. It's when they're downrezzed into a 4096X1716 timeline that the playback is jerky. That's not the case with 4.6k CDNGs so I assume it's a Resolve tuning problem.

(I also found that the constant compression clips have the same problem)

It makes no sense. Once BM RAW is decoded (and it decodes way easier than DNG) then it's all the same processing pipe. Maybe some bug/beta issue. I actually noticed similar problem- playing 2.40 aspect BM RAW sample was using more resources than higher resolution 16:9. Something odd.

Try disabling 1 card as a test.
Also as a test try 4096x1712 resolution.

Misha, the Areca is way fast enough to playback this stuff. The clips play fine from the media pool page. It's when they're downrezzed into a 4096X1716 timeline that the playback is jerky. That's not the case with 4.6k CDNGs so I assume it's a Resolve tuning problem.

(I also found that the constant compression clips have the same problem)

It makes no sense. Once BM RAW is decoded (and it decodes way easier than DNG) then it's all the same processing pipe. Maybe some bug/beta issue. I actually noticed similar problem- playing 2.40 aspect BM RAW sample was using more resources than higher resolution 16:9. Something odd.

Try disabling 1 card as a test.
Also as a test try 4096x1712 resolution.

Neither suggestion had the least effect, sorry to say.

Misha, the Areca is way fast enough to playback this stuff. The clips play fine from the media pool page. It's when they're downrezzed into a 4096X1716 timeline that the playback is jerky. That's not the case with 4.6k CDNGs so I assume it's a Resolve tuning problem.

(I also found that the constant compression clips have the same problem)

Bandwidth wise it should be no problem, how about I/O-wise?

If the clips play fine from the Media page how would I/O have anything to do with it?

Misha, the Areca is way fast enough to playback this stuff. The clips play fine from the media pool page. It's when they're downrezzed into a 4096X1716 timeline that the playback is jerky. That's not the case with 4.6k CDNGs so I assume it's a Resolve tuning problem.

(I also found that the constant compression clips have the same problem)

It makes no sense. Once BM RAW is decoded (and it decodes way easier than DNG) then it's all the same processing pipe. Maybe some bug/beta issue. I actually noticed similar problem- playing 2.40 aspect BM RAW sample was using more resources than higher resolution 16:9. Something odd.

Try disabling 1 card as a test.
Also as a test try 4096x1712 resolution.

Neither suggestion had the least effect, sorry to say.

Do oyu have BM RAW processing set to GPU or CPU in Resolve preferences? Try also playing with this as test.

Do oyu have BM RAW processing set to GPU or CPU in Resolve preferences? Try also playing with this as test.

It was set to GPU. Changing it to CPU made the issue slightly worse

BTW I've been playing with the green screen material that Nate Porter shot yesterday at https://drive.google.com/drive/folders/1MH2xLzaHsxILQpwlB5jt2gDYI8H-n7PY?usp=sharing if anyone else wants to try to duplicate.