Welcome to our community

Be a part of something great, join today!

Register Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

ARRI 4K Super 35 Camera Will be Launched Next Month

Attempted some highlight recovery on some Swing clips, but still a work in progress

1-14-2-1-14-2.jpg


1-18-1-1-18-1.jpg
 
Last edited:
The Alexa 35 footage(even still using AWG3-ALC3 INSTEAD OF AWG4 ALC4) does really well at controlling "OUT OF GAMUT COLORS" with the AWG4 LOGC4 Transform Luts. The "REDs" in this image are just a little "OUT OF GAMUT" considering the intensity of the color. I used a slight 'HUE vs SAT" curve to bring them back just a little more.

1.png

Screenshot-5365.png

2.jpg





3.png

Screenshot-5366.png

4.jpg
 
Last edited:
Here's a screen record of the video above next to a vectorscope .You can see how well the flashing red light is being controlled with the AWG4/ALC4 transform Lut applied.


 
rand thompson said:
Out of Gamut Colors
Click to expand...

I just released my actual Color/Gamma Transforms after testing them for a long, long time. Examining ARRI's LogC4 LUTs, the V1 ones, it's pretty interesting that some of the "old ways" are still apart of their color science and pipeline. If you look closely on how that material reads on the waveform and Vectorscope, the red colors are exceeding 100% IRE and smashing into the ceiling of both their SDR and HDR Outputs.

RED doesn't go this route for their transforms as they try to keep it wrangled within scopes and waveform.

If you want a general summary of the differences in the DNA of their Color Science, overall the most interesting difference I'd say and this has been true generally speaking for a long time, RED leans to a very accurate natural image with subtle perceptual choices while ARRI is aiming at a specific aesthetic. Previously they stated this was influenced by film, this time around there's a few notable changes that I think are more their own thing. Taking material from both manufacturers in and out of their color science reveals a lot, shooting tests reveals a lot more. Some fascinating differences between their old color and new color stuff too.
 
Phil Holland said:
I just released my actual Color/Gamma Transforms after testing them for a long, long time. Examining ARRI's LogC4 LUTs, the V1 ones, it's pretty interesting that some of the "old ways" are still apart of their color science and pipeline. If you look closely on how that material reads on the waveform and Vectorscope, the red colors are exceeding 100% IRE and smashing into the ceiling of both their SDR and HDR Outputs.

RED doesn't go this route for their transforms as they try to keep it wrangled within scopes and waveform.

If you want a general summary of the differences in the DNA of their Color Science, overall the most interesting difference I'd say and this has been true generally speaking for a long time, RED leans to a very accurate natural image with subtle perceptual choices while ARRI is aiming at a specific aesthetic. Previously they stated this was influenced by film, this time around there's a few notable changes that I think are more their own thing. Taking material from both manufacturers in and out of their color science reveals a lot, shooting tests reveals a lot more. Some fascinating differences between their old color and new color stuff too.
Click to expand...

Phil,

I'd be really interested in a "Write-up" if you have the time to do It. I've been basically messing around with how the AWG4/ALC4 TRANSFORM LUTs react with the older AWG3/ ALC3 Color Space so I haven't yet seen these Luts true performance with the AWG4/ALC4 Color Space yet. Yeah, I can't wait to see what we all find out over the next few months about the similarities and differences between the RWG/LOG3G10 workflow and the AWG4/ALC4 workflow. But the AWG4/ALC4 workflow is basically just ARRI's version of the RED IPP2 workflow.
 
Last edited:
rand thompson said:
I'd be really interested in a "Write-up" if you have the time to do It. I've been basically messing around with how the the AWG4/ALC4 TRANSFORM LUTs react with the older AWG3/ ALC3 Color Space so I haven't yet seen these Luts true performance with the AWG4/ALC4 Color Space yet. Yeah, I can't wait to see what we all find out over the next few months about the similarities and differences between the RWG/LOG3G10 workflow and the AWG4/ALC4 workflow. But the AWG4/ALC4 workflow is basically just ARRI's version of the RED IPP2 workflow.
Click to expand...

ARRI's new workflow being similar to RED's IPP2 workflow is a big step in the right direction and will minimize some chaos in post. I whipped up my transforms according to the white paper and a lot of math and testing. I'm certain Resolve will add this to their own CST plugin, perhaps even in the next beta. I do specify a different name on my transforms rather than calling them IDTs or CSTs. I use CGT mainly because you are transforming Color Space and Gamma Curve. I also take into account exposure compensation on some key manufacturer's transforms like RED or ARRI to account for 18% Gray being a different target. Actually helps a lot with studios using mine and other's transforms a lot to know who's you're using.
 
Phil Holland said:
ARRI's new workflow being similar to RED's IPP2 workflow is a big step in the right direction and will minimize some chaos in post. I whipped up my transforms according to the white paper and a lot of math and testing. I'm certain Resolve will add this to their own CST plugin, perhaps even in the next beta. I do specify a different name on my transforms rather than calling them IDTs or CSTs. I use CGT mainly because you are transforming Color Space and Gamma Curve. I also take into account exposure compensation on some key manufacturer's transforms like RED or ARRI to account for 18% Gray being a different target. Actually helps a lot with studios using mine and other's transforms a lot to know who's you're using.
Click to expand...

Yeah! I'm using ACES so hopefully they also add it to the 'ACES TRANSFORM" OFX as well since I use both the Project and Node based ACES workflows. I think I remember the new AWG4/ALC4 18% gray point being 28% and the AWG3/ALC3 being around 34 to maybe 38% .
 
What's New in DaVinci Resolve Studio 18 Public Beta 6

DOWNLOAD PAGE
Support Center | Blackmagic Design



Support for Blackmagic RAW SDK 2.6
Support for the ARRI ALEXA 35 camera.
Support for encoding Dolby Vision H.265 Profile 8.4 on Apple silicon.
Support for bidirectional tracking in the advanced panel.
Support for fast review playback mode in the advanced panel.
Improved aliasing behavior when playing camera gyro stabilized clips.
Optimized GPU decode performance for R3D clips.
Addressed issue with reading source files for fully cached clips in some scenarios.
Addressed left channel artifacts in specific stereo nested configurations.
Addressed some fixed bus projects from 17.4.6 loading as FlexBus projects.
Addressed an issue with elastic wave on small audio clips.
Changing camera gyro strength no longer restarts analysis.
Addressed a crash with camera gyro based stabilization.
Addressed a colorspace tagging issue for Blackmagic gen 5 color science ProRes clips.
Addressed a saturation mismatch when color matching BRAW clips with gen 5 color science.
Addressed surface tracker artefacts in some occlusion scenarios.
Addressed a cache regeneration issue when changing Open FX native interface parameters.
Addressed an issue with decoding some Canon AVC clips on Apple silicon systems.
Addressed the collaboration person list not reflecting all active users in some cases.
Addressed Fusion inspector not reflecting changes from the Edit page.
Addressed incorrect behavior for circle and rectangle paint with quick key and mouse inputs.
Addressed an issue with viewing polygons in quad view with vectorscopes and waveforms.
General performance and stability improvements.
 
EVERYTHING ALEXA 35 IS HERE IN RESOLVE 18 BETA 6

Screenshot-5384.png

Screenshot-5376.png

Screenshot-5388.png

Screenshot-5385.png


Screenshot-5383.png

Screenshot-5386.png


Screenshot-5387.png
 
Last edited:
Phil Holland said:
I just released my actual Color/Gamma Transforms after testing them for a long, long time. Examining ARRI's LogC4 LUTs, the V1 ones, it's pretty interesting that some of the "old ways" are still apart of their color science and pipeline. If you look closely on how that material reads on the waveform and Vectorscope, the red colors are exceeding 100% IRE and smashing into the ceiling of both their SDR and HDR Outputs.

RED doesn't go this route for their transforms as they try to keep it wrangled within scopes and waveform.

If you want a general summary of the differences in the DNA of their Color Science, overall the most interesting difference I'd say and this has been true generally speaking for a long time, RED leans to a very accurate natural image with subtle perceptual choices while ARRI is aiming at a specific aesthetic. Previously they stated this was influenced by film, this time around there's a few notable changes that I think are more their own thing. Taking material from both manufacturers in and out of their color science reveals a lot, shooting tests reveals a lot more. Some fascinating differences between their old color and new color stuff too.
Click to expand...

I was listening to Harald Brendel on the Colorist Meetup podcast yesterday and he said LogC3 didn't have enough room for that extra 2 stops over so that's why they needed to switch LogC4.
Alexa S35 produces an 18bits linear signal and Arriraw recording is 13bit ( 12bit Prores LogC4 )
I guess now they need Log-to-Log transforms for LogC4 to LogC3 for Alexa S35?
Is Red working on a new color science as well? Like LogC4 equivalent?
 
Kemalettin Sert said:
I was listening to Harald Brendel on the Colorist Meetup podcast yesterday and he said LogC3 didn't have enough room for that extra 2 stops over so that's why they needed to switch LogC4.
Alexa S35 produces an 18bits linear signal and Arriraw recording is 13bit ( 12bit Prores LogC4 )
I guess now they need Log-to-Log transforms for LogC4 to LogC3 for Alexa S35?
Is Red working on a new color science as well? Like LogC4 equivalent?
Click to expand...

For anybody "going backwards" the key thing to note when going from LogC4 to LogC3 is how that works with ARRI's classic LUTs like the 2021 or K1S1. The two stops in the head get smashed hard and won't exactly look great contained in those output transforms. What I like to do, and I've been testing this all extremely aggressively for the last month, is to create a custom highlight roll-of to tune the highlights if I was to use those output LUTs from ARRI.

As for RED, bare in mind ARRI's new color workflow is alarmingly similar to RED's IPP2 in log execution and strategy. Different curves, but same idea basically. I don't know if RED is working on new color science (I mean they are always working on stuff), but in terms of their current latest wide gamut and expanded log curve RWG/Log3G10 (compared to REDlogFilm/Cineon), they likely don't need to update it for any reason anytime soon as it can hold 19+ stops already for most practical purposes and can hold beyond that into the shadows. Don't actually know how far as that depends on precision and cleanliness of the noise floor.

IPP2 is now 5-6 years old. And though it went largely missed here, on Facebook there was a nice glowing conversation regarding Graeme and Co.'s efforts on IPP2 with compliments actually coming from other manufacturers regarding their efforts towards image processing and color science keeping everybody on their toes. IPP2 is very, very forward thinking.

IPP2 came with lots of new things, like a new debayer, highlight extension, and the color science. Recently, though it went EXTREMELY under the radar with a mysterious note in the change notes, there was a new demosaic at the end of 2021 could likely be considered part of IPP3, but I don't think RED wanted to add to any confusion as it's just a debayer update. But I can't underline this enough, it's made every RED camera stand the test of time with it's improved draw, in regards to both detail as well as color. I posted some A/B comparisons on that front and the difference is staggering in some cases.
 
I've been trying to post a test of the "SWING" clip I uploaded here as a PRORES 4444XQ file but both RESOLVE keeps crashing and the video I posted on both Youtube and Vimeo had "Blacks" that were really lifted. So Here are a few frames from that 33 second clip until I can find out what's going wrong with Resolve 18 beta 6.

1-1-4-1.jpg


grade-2-1-4-3.jpg


grade-3-1-4-5.jpg


grade-3-1-4-6.jpg


grade-4-1-4-7.jpg


grade-5-1-4-8.jpg
 
I think the Alexa 35 ProRes 4444XQ LogC4 files I posted on page 16 and 17 get deleted today and tomorrow , when they do I will upload them all again and this time for another 7 days this time. So let me know if one of the files you may have wanted to download has been deleted and I will upload it again.:thumbsup:
 
Here is the short ''SWING'' Test Video made from the ProRes 4444 XQ file I uploaded that I tried to export out of Resolve 18 beta 6 yesterday, don't know if it looks correct or not on your computer monitor.

 
Last edited:
rand thompson said:
Here is the short ''SWING'' Test Video made from the ProRes 4444 XQ file I uploaded that I tried to export out of Resolve 18 beta 6 yesterday, don't know if it looks correct or not on your computer monitor.

Click to expand...

Blacks are a bit too clamped. To much -- evening for a middle grade.
 
Alex,

Yeah, I uploaded another version. I was using a DTCL in ACES to invert the RRT. So I think that is the problem. It seems that every time I look at it it keeps getting darker and darker. I don't think I will be using that DCTL again, I'll just stick with the RRT in ACES and maybe try again later.
 
I updated all of the ALEXA 35 PRORES 4444XQ files on page 16 transformed from their ARRIRAW files originals in the ARRI REFERENCE TOOL for 7 more days.

https://www.reduser.net/forum/everyt...49#post3794749

PAGE 17 FILES

https://www.reduser.net/forum/everything-else/off-topic/3791062-arri-4k-super-35-camera-will-be-launched-next-month?p=3794914#post3794914
 
Last edited:
Phil Holland said:
Rand, the files are publicly available on ARRI's ftp.
Click to expand...

Phil,

If it's a problem with me posting these Files here just let me know and I will delete them. I was just doing it because the ARRIRAW Files on the Server range from 9GB to as much as 174GB per file. So I wanted to make them more digestible for people who couldn't download as much as 700GB+ of ARRIRAW files. How about when these 7 days are up I don't repost them?
 
You must log in or register to reply here.
Back
Top