Did anyone do tests on the average BRAW compression?

The place for questions about shooting with Blackmagic Cameras.
  • Author
  • Message
Offline

Sean van Berlo

  • Posts: 221
  • Joined: Wed Jul 01, 2015 6:33 am
  • Location: The Netherlands

Did anyone do tests on the average BRAW compression?

PostThu Feb 13, 2020 8:52 am

Hey good folks!

Out of curiosity before I myself end up filling up a lot of cards - did anyone do any tests on the average compression rate of Q5 and Q0 in different scenarios?
Offline

Sami Sanpakkila

  • Posts: 48
  • Joined: Fri Sep 13, 2013 3:48 pm

Re: Did anyone do tests on the average BRAW compression?

PostThu Feb 13, 2020 9:40 am

I tested all of the BRAW compression ratios and only saw any differences with 12:1. Then I shot a few projects with 5:1 and then moved on to Q5 and have been using that with no regrets. The quality difference between the compression ratios is so small compared to the amount of space you save it makes no sense to me to shoot anything else. Depends on your situation though!

One thing I've noticed is that when you start recording in Q5 the available recording time is adjusted after a while and it is actually a little more than the estimate the camera gives you. Q0 takes up twice the space and if you're not going to do green screen of VFX work I highly doubt you'll find it an added benefit.

I've shot in low light, smoke, strobes, fast movement, foliage... Haven't broken the Q5 yet. It is a very good option.
Offline
User avatar

VINOVINKKELI

  • Posts: 20
  • Joined: Sat Feb 09, 2019 8:46 pm
  • Location: Finland
  • Real Name: Teemu Saarinen

Re: Did anyone do tests on the average BRAW compression?

PostThu Feb 13, 2020 10:56 am

Sami Sanpakkila wrote:I tested all of the BRAW compression ratios and only saw any differences with 12:1. Then I shot a few projects with 5:1 and then moved on to Q5 and have been using that with no regrets. The quality difference between the compression ratios is so small compared to the amount of space you save it makes no sense to me to shoot anything else. Depends on your situation though!

One thing I've noticed is that when you start recording in Q5 the available recording time is adjusted after a while and it is actually a little more than the estimate the camera gives you. Q0 takes up twice the space and if you're not going to do green screen of VFX work I highly doubt you'll find it an added benefit.

I've shot in low light, smoke, strobes, fast movement, foliage... Haven't broken the Q5 yet. It is a very good option.


Nice to hear real-world experiences with this Q5 setting. I haven't tested Q0 or Q5 at all. I have mostly shot with 5:1 and 8:1. Some specific cases 3:1.

I always wanted to know that I am shooting enough quality with the constant settings as I have read and heard that Q5 setting might at times go a bit too low when it shouldn't. Of course, a green screen or VFX stuff is better with constant bitrate (or Q0) and as high quality as possible.

But as you Sami mentioned those cases "low light, smoke, etc" and not seeing problems, started to think that I really should try out Q0 setting myself aswell. The used space is valuable but so is the quality. Always looking for the best of both worlds.
Intel Skylake i7-6700K @ 4.0GHz, 32GB RAM, Nvidia GTX 960 4GB, Win 10 Pro, System disk: Samsung NVMe M.2 960 EVO 250GB, Workdisk#1(MATERIAL): Samsung EVO850 1TB, Workdisk#2(CACHE): Samsung EVO850 500GB, DaVinci Resolve 15.2.3, Dell 34" Ultrasharp
Offline

Tim Lota

  • Posts: 160
  • Joined: Fri Dec 13, 2013 12:48 pm
  • Location: Essen, GER

Re: Did anyone do tests on the average BRAW compression?

PostThu Feb 13, 2020 12:08 pm

VINOVINKKELI wrote:..as I have read and heard that Q5 setting might at times go a bit too low when it shouldn't.


I've read that too, I even think it was on this board. But since it was when BRAW first came out I always thought they might have fixed it, because A LOT of people here are using Q5 with no regrets.
I remember someone saying that Q5 produced way more noise in the shadows than any other setting, even more than 12:1. But nobody else confirmed it..

Good to know that it works in different, stressful scenarios. I'm planning on using BRAW in the future, in a scenario where I'll possibly only have one chance for a lot of shots. Since I need something that's as high quality as possible, while saving as much space as possible, I've always thought about going 8:1 or 5:1. If Q5 doesn't have a noise problem after all, I'll happily take it over any other setting. I'd be awesome if some other folks could chip in on this, too. Thanks.
Offline

Michael_Andreas

  • Posts: 1481
  • Joined: Sat Jan 05, 2019 9:40 pm
  • Real Name: Michael Andreas

Re: Did anyone do tests on the average BRAW compression?

PostThu Feb 13, 2020 1:06 pm

I'm wondering about the possibility of a high data rate under certain conditions, like a noisy exposure. On the BMPCC product page is this statement: "The new constant quality options preserve image quality by completely removing the upper data limit. " While the chart implies that Q0 has a range of compression ratios from 2:1 to 5:1 and that Q5 has a range of compression ratios from 7:1 to 20:1, the statement "completely removing the upper data limit" makes me wonder if it could exceed this range, maybe even recording with no compression at all.

Does anyone have anything to go on to indicate that this is or isn't the case?
_________________________________________________
DR Studio 16.1.2.026, Win10Pro 1909/18363.535 - i7-6700K@4GHz, 32GB RAM
RTX 2070 8GB, "Studio" driver 442.19 2/3, prev: 441.66, 431.86
OS,Library: 1TB SSD - Project: 1TB SSD - Cache: 1TB NVMe
Offline
User avatar

rick.lang

  • Posts: 12608
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: Did anyone do tests on the average BRAW compression?

PostThu Feb 13, 2020 2:51 pm

I believe I posted quite awhile ago that Q0 appears to be able to record at higher data rates than 2:1 if the scene has sufficient detail and contrast. When I did my tests, I examined the actual file sizes for the duration of time as well as the information available in DaVinci Resolve Inspector. But for typical scenes the data rate is close to 3:1.

I don’t have as much experience with Q5, but it also seemed to exceed its published range with good results in my tests.

If you want predictable results you use constant quantity such as 3:1, but if you know your media can cope with higher data rates the constant quality Q0 is excellent. I can manage for my typical shoots with Q0. But not when I have a 12 hour shoot.

If media and storage are issues, I’d shoot Q5 rather than 12:1. Because often that will use even less storage than you expect.
Rick Lang
Offline

Tim Lota

  • Posts: 160
  • Joined: Fri Dec 13, 2013 12:48 pm
  • Location: Essen, GER

Re: Did anyone do tests on the average BRAW compression?

PostThu Feb 13, 2020 3:24 pm

I'm kinda undecided between Q5 and 8:1 (with 5:1 as an option in the back of my head).
Offline
User avatar

rick.lang

  • Posts: 12608
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Re: Did anyone do tests on the average BRAW compression?

PostThu Feb 13, 2020 4:02 pm

That’s a good choice either way. I’d suggest you try both and see if you can see a difference even when examining a frame versus motion pictures.
Rick Lang
Offline

pnguyen720

  • Posts: 249
  • Joined: Tue Mar 19, 2019 9:17 pm
  • Real Name: Phong Nguyen

Re: Did anyone do tests on the average BRAW compression?

PostThu Feb 13, 2020 5:01 pm

Tim Lota wrote:
VINOVINKKELI wrote:..as I have read and heard that Q5 setting might at times go a bit too low when it shouldn't.


I've read that too, I even think it was on this board. But since it was when BRAW first came out I always thought they might have fixed it, because A LOT of people here are using Q5 with no regrets.
I remember someone saying that Q5 produced way more noise in the shadows than any other setting, even more than 12:1. But nobody else confirmed it..


I'd like to know this too. Can BM confirm this?
Offline
User avatar

carlomacchiavello

  • Posts: 887
  • Joined: Tue Aug 28, 2012 6:04 pm
  • Location: italy

Did anyone do tests on the average BRAW compression?

PostThu Feb 13, 2020 8:22 pm

Tim Lota wrote:I remember someone saying that Q5 produced way more noise in the shadows than any other setting, even more than 12:1. But nobody else confirmed it..

If Q5 doesn't have a noise problem after all, I'll happily take it over any other setting. I'd be awesome if some other folks could chip in on this, too. Thanks.


Noise is present if you shoot in the worst way.
Q5 is more compress that mean less dectails that mean less noise not more noise, no one codec raw or compressed add noise in compression.
Braw add a moderate denoise to optimize compression, more compressed more denoise


Inviato dal mio iPhone utilizzando Tapatalk
Offline
User avatar

carlomacchiavello

  • Posts: 887
  • Joined: Tue Aug 28, 2012 6:04 pm
  • Location: italy

Did anyone do tests on the average BRAW compression?

PostThu Feb 13, 2020 8:28 pm

Search on web frank glecairn website he did a lots of test about braw very complete and excellent comparison

https://frankglencairn.wordpress.com/20 ... er-at-all/

Inviato dal mio iPhone utilizzando Tapatalk
Offline

Justin Jackson

  • Posts: 651
  • Joined: Thu Apr 28, 2016 3:50 am

Re: Did anyone do tests on the average BRAW compression?

PostThu Feb 13, 2020 10:23 pm

Well shoot.. I have been using 12:1 all the time.. I think I tried the Q5 and it resulted in larger files. For a 90 minute soccer match, I was seeing around 300GB's of file size using 4KDCI.

Is Q5 and 12:1 similar in quality, and file size? Or is Q5 much larger file sizes? Trying to understand the trade off. I assume 12:1 and Q5 are MUCH better quality than the ProRes options, given that it's 16bit RAW?

Personally, I can't really tell much difference between all these different options. Ideally I want to maximize space while reducing any quality loss. I have a 500GB SSD (T5) that when using 12:1 can fit 2 games on it. If Q5 pushes me in the 300gb range, I am now not able to fit 2 games (and I typically record two games back to back) on it.
Custom DIY AMD1950x 16-core/32-thread, liquid cooled, 64GB 3600Mhz RAM, 950Pro-512GB NVMe os/apps, 2x500GB 850 Evo RAID 0 SATA3, Zotac 1070 8GB video, USB 3.1Gen2 RAID0 2x4TB, 2x2TB Crucial MX500 SSD SATA3.
Offline
User avatar

rick.lang

  • Posts: 12608
  • Joined: Wed Aug 22, 2012 5:41 pm
  • Location: Victoria BC Canada

Did anyone do tests on the average BRAW compression?

PostFri Feb 14, 2020 4:31 am

Justin, this is an example where you need predictable data rates; you’ve tested Q5 and for your requirements, it’s generating more data than you want and more data than 12:1. If 12:1 quality was fine for your needs, then go with that.
Rick Lang
Offline

Tim Lota

  • Posts: 160
  • Joined: Fri Dec 13, 2013 12:48 pm
  • Location: Essen, GER

Re: Did anyone do tests on the average BRAW compression?

PostFri Feb 14, 2020 7:11 am

carlomacchiavello wrote:l
Noise is present if you shoot in the worst way.
Q5 is more compress that mean less dectails that mean less noise not more noise, no one codec raw or compressed add noise in compression.
Braw add a moderate denoise to optimize compression, more compressed more denoise


Yes, I know. This is why I was confused why it would produce more noise.

Thanks for the link!
Offline
User avatar

carlomacchiavello

  • Posts: 887
  • Joined: Tue Aug 28, 2012 6:04 pm
  • Location: italy

Re: Did anyone do tests on the average BRAW compression?

PostFri Feb 14, 2020 7:12 pm

There is a balance between compression kind / quality and iso setup that give you better result.

Fixed compression (3:1-5:1-8:1) give a more smoothing result of higher iso be cause you force more to reduce dectails with higher compression

Variable compression (q0,q5) give you more noise presence be cause q5 try to lower data rate but noise force it to keep higher.

Q5 give you better result with lower iso (less noise) better balance between size and noise

Fixed lower compression like8:1 give you better balance on higher iso be cause to compress more there is a moderate higher denoise of clip.

Anyway every situation should be verified, I suggest to check with waveform the noise level to decide the kind of compression.


Inviato dal mio iPhone utilizzando Tapatalk
Offline

Jim Giberti

  • Posts: 270
  • Joined: Tue Mar 03, 2015 2:03 am

Re: Did anyone do tests on the average BRAW compression?

PostFri Feb 14, 2020 10:55 pm

I've said it before - Q5 all day long.

Amazing results in every way.
Offline

Justin Jackson

  • Posts: 651
  • Joined: Thu Apr 28, 2016 3:50 am

Re: Did anyone do tests on the average BRAW compression?

PostFri Feb 14, 2020 11:00 pm

Interesting.. 12:1 more ISO. Good to know. I am fine with 12:1. I don't do much if any grading, maybe a little depending on time of day.
Custom DIY AMD1950x 16-core/32-thread, liquid cooled, 64GB 3600Mhz RAM, 950Pro-512GB NVMe os/apps, 2x500GB 850 Evo RAID 0 SATA3, Zotac 1070 8GB video, USB 3.1Gen2 RAID0 2x4TB, 2x2TB Crucial MX500 SSD SATA3.

Return to Cinematography

Who is online

Users browsing this forum: Dmitry Shijan, Neil-G, Ola Harström and 14 guests