- Posts: 4
- Joined: Sat Mar 18, 2017 8:21 pm
Sat Jul 14, 2018 12:52 am
I’ve been using the BMPCC on and off just about since it came out, and I’ve shot a variety of projects, in both ProRes and raw. This has usually been dictated by budget and time constraints. However, I’ve been doing some tests for a music video I’m shooting next week and for the first time there’s some flexibility about this, so we’re comparing the formats. So this is actually the first time I’ve shot a direct comparison, and I’ve come across an interesting discrepancy.
I shot a lot of bracketed exposures, including a grey card. For each bracket, I shot 10 seconds of ProRes HQ and 10 seconds of raw, without changing any other settings on the camera. I’ve put them in the same timeline in Resolve, and applied the default REC.709 LUT for the camera (I’ve also tried it in ACES cct with very similar results). The camera was set to ISO 800
For every exposure, the raw shot comes out darker. Adjusting the exposure setting in the raw tab to 0.5 (is this approximately calibrated in stops?) seems to brighten it up so they match. Some very approximate code values from the grey card pre adjustment, taken from the waveform and colour picker:
No LUT ProRes 384
No LUT Raw 330
With LUT ProRes 410
With LUT Raw 330
There is also a very slight shift in colour on the green/magenta axis, only really noticeable in shadow areas.
My understanding was that whilst raw offered a higher bit depth and less compression it should match the ProRes in tonality and colour, so all other things being equal (exposure, etc.) it should cut pretty seamlessly. So I wonder if I have something set up wrong in Resolve. I’ve been through all the settings I can find, but I can’t figure out what it could be.
More concisely, my questions are:
1. Is the 38.4% figure for a greycard on a waveform correct for the pocket camera? I see it crop up online, but I can’t find a proper whitepaper with a transfer function and matrices for BMDFilm.
2. Where should the grey card fall on the waveform after the LUT is applied? 40%? 50%
3. Should the ProRes and raw match exactly if the project is set up correctly? Both in terms of exposure and colour.
4. What could I be missing in Resolve given that they don’t match?
As I say, I get a similar discrepancy in ACES cct. It’s probably worth mentioning that this is in the beta version of Resolve, although I get the same results trying it in Resolve 14.
Thanks
I shot a lot of bracketed exposures, including a grey card. For each bracket, I shot 10 seconds of ProRes HQ and 10 seconds of raw, without changing any other settings on the camera. I’ve put them in the same timeline in Resolve, and applied the default REC.709 LUT for the camera (I’ve also tried it in ACES cct with very similar results). The camera was set to ISO 800
For every exposure, the raw shot comes out darker. Adjusting the exposure setting in the raw tab to 0.5 (is this approximately calibrated in stops?) seems to brighten it up so they match. Some very approximate code values from the grey card pre adjustment, taken from the waveform and colour picker:
No LUT ProRes 384
No LUT Raw 330
With LUT ProRes 410
With LUT Raw 330
There is also a very slight shift in colour on the green/magenta axis, only really noticeable in shadow areas.
My understanding was that whilst raw offered a higher bit depth and less compression it should match the ProRes in tonality and colour, so all other things being equal (exposure, etc.) it should cut pretty seamlessly. So I wonder if I have something set up wrong in Resolve. I’ve been through all the settings I can find, but I can’t figure out what it could be.
More concisely, my questions are:
1. Is the 38.4% figure for a greycard on a waveform correct for the pocket camera? I see it crop up online, but I can’t find a proper whitepaper with a transfer function and matrices for BMDFilm.
2. Where should the grey card fall on the waveform after the LUT is applied? 40%? 50%
3. Should the ProRes and raw match exactly if the project is set up correctly? Both in terms of exposure and colour.
4. What could I be missing in Resolve given that they don’t match?
As I say, I get a similar discrepancy in ACES cct. It’s probably worth mentioning that this is in the beta version of Resolve, although I get the same results trying it in Resolve 14.
Thanks