Announcement

Collapse
No announcement yet.

Bitrate versus Quality Question

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Bitrate versus Quality Question

    OK for those of you who've been using BRaw for a while now. Which way did you go? I just installed the update to the BMPCC4K and am deciding what will be my starting point. Would appreciate any discussion on this. Thanks…

  • #2
    Just trying things out… I put an old SanDisk Extreme Pro 95MB/S SD card in and recorded a couple of 30 second UHC BRaw 3:1 clips with no issues. I know it's not a deep test but I'm amazed that it would record anything at all.

    Comment


    • #3
      Never understood people asking those questiosn like: Which iso, which codec, and so on. You are a cinematographer. Take the camera, use different lenses and just do some tests. This is what all professional cinematographers are doing for years. braw 12:1 is just fine. With Q5 or whatever setting it also goes up to 20:1 compression. And even this is fine. There are some tech-videos about the REDraw codec but I do not find thelinks yet. Where they spoke technically about raw compression thing. My english is not good enough to write it down.

      But all I want to say is: Just do some tests. You got the camera. You got braw 3:1, 5:1, 8:1, 12:1. What do you want to know? Just film different things with lots of details. Konfetti, hair, stuff with ver yfine details. And just look at the image in fullscreen. Or take some stills and look at it in photoshop. Or maybe underexpose the image 3 stops and bring it up in post again. Or do some greenscreen test. Or whatever YOU want to test.

      (sorry for this little rant)

      Comment


      • #4
        S_Berger, just starting a conversation to begin working with these files. I expect to be the judge of my footage just like I have my entire career.

        I've always found this group to be free of all the petty arguing, nurturing questions like this. A place where conversations leads to learning.

        Frankly I've learned so much from so many around here, you included, that I'm surprised by your response. Did not think putting this question into a defensive posture was really necessary. But that's just my take.

        My question still stands if anyone has any experience with the two file types. Bitrate versus Quality.

        Comment


        • #5
          I'm lazy and not all that technically critical, focus on aesthetics. I don't have a camera yet so am happy to follow others findings and see if consensus forms. When the first braw tests were released they were mostly 12:1 and I couldn't find any flaws. If Koli and Frank say you can't break it I will believe them. Now if later it comes to pass that you need 3:1 or q0 for effects work then so be it. There will always be clients that think you need to shoot whatever the best setting is, and if they are paying that's what you do, but for my own stuff it will be 12:1 on sd cards until I find a situation that breaks it.

          Comment


          • #6
            Originally posted by S_Berger View Post
            Never understood people asking those questiosn like: Which iso, which codec, and so on. You are a cinematographer. Take the camera, use different lenses and just do some tests. This is what all professional cinematographers are doing for years. braw 12:1 is just fine. With Q5 or whatever setting it also goes up to 20:1 compression. And even this is fine. There are some tech-videos about the REDraw codec but I do not find thelinks yet. Where they spoke technically about raw compression thing. My english is not good enough to write it down.

            But all I want to say is: Just do some tests. You got the camera. You got braw 3:1, 5:1, 8:1, 12:1. What do you want to know? Just film different things with lots of details. Konfetti, hair, stuff with ver yfine details. And just look at the image in fullscreen. Or take some stills and look at it in photoshop. Or maybe underexpose the image 3 stops and bring it up in post again. Or do some greenscreen test. Or whatever YOU want to test.

            (sorry for this little rant)

            And then some of us don't have the camera yet and are just curious.

            Comment


            • #7
              For the record: I always use Q0 - the data rate is already so low (compared to DNG), I can see no reason, to use anything less, but the highest quality BRAW I can get.
              Blog: http://frankglencairn.wordpress.com/

              Comment


              • #8
                Hi Marshall,

                I have not done in-depth testing, and I probably won’t, so I will not speak to the quality difference. If I’m doing anything that’s on the steadicam, or I’m at all concerned with maximizing quality, I go for Q0. HDD’s are cheap, reshoots are not. If I’m doing a long interview, where the background is less relevant or static, I use 3:1 5:1 or whatever setting gives me the least amount of compression for the length of time I need. It’s the same way I shoot on RED. Whatever generates the highest quality in the amount of space I have available. With that said, 9/10 it’s Q0 for narrative, including B-roll, and 3:1 for doc style interviews. I like to keep rolling during interviews, because people do often say “can we stop?” followed by a perfect performance of what they wanted to say and exactly how they wanted to say it.

                Comment


                • #9
                  Originally posted by S_Berger View Post
                  Never understood people asking those questiosn like: Which iso, which codec, and so on. You are a cinematographer. Take the camera, use different lenses and just do some tests. This is what all professional cinematographers are doing for years.
                  I'm not a professional cinematographer. I own and use the Pocket 4K, that's ok isn't it?
                  Advice from people who are more critical, people with a lot more experience than me is very valuable. I'm a professional musician of 40 years. I don't hesitate to offer help and advice to others less experienced, or part timers.\
                  I don't see a problem in having a discussion between experienced professionals about the settings they use and the approach they take, that others like me can learn from.

                  Comment


                  • #10
                    Originally posted by Frank Glencairn View Post
                    For the record: I always use Q0 - the data rate is already so low (compared to DNG), I can see no reason, to use anything less, but the highest quality BRAW I can get.
                    I thought the Q0 bitrate was typically close to cDNG 3:1, is it really much smaller than that?
                    My YouTube channel

                    Comment


                    • #11
                      I just installed the BRAW update, and then immediately filmed a torture test - complex patterned fabric blowing wildly with a fan.

                      I cycled through the all the RAW settings from Q0 down to 12:1. Just on sheer motion blur / complex motion, I was struggling to see any difference between any of them. Certainly didn't see any artifacts like you would see on an mp4 encode or anything like that.

                      I'd say motion blur / complex stuff moving quickly is pretty even throughout. There has to be penalties of some kind moving into 12:1, and I imagine it's at the color level.
                      director reel

                      11 Blocks, debut feature on Netflix

                      Comment


                      • #12
                        You can skip the fan, Braw is an I-frame codec.

                        Comment


                        • #13
                          Yeah, movement shouldn't break an I - frame codec, torture tests would need to be colour, grading based, possibly how well fine detail is preserved, but motion blur and that kind of stuff shouldn't really be a variable.

                          Comment


                          • #14
                            Motion blur might even be easier on the codec – loss of details.

                            Comment


                            • #15
                              Moving the camera, or at least shooting stuff that moves (to make sure to introduce some motion blur) is a good way to keep the test grounded (assuming you want to shoot a movie/film/video). I'm guessing even the slightest amount of movement would be a great equaliser in terms of visible detail in codecs.

                              One exception is if you want to shoot video as a "high frame rate" photo camera application, just to grab the best frames as stills afterwards. That would require different tests.

                              Comment

                              Working...
                              X