Announcement

Collapse
No announcement yet.

Audio sample rate / frame rate

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

  • Audio sample rate / frame rate

    Hi there,

    I am cutting a feature film on FCP, shot on 35mm, 3-perf, 24fps. The footage will be telecined/mastered on HDCam SR at 23.976.
    The lab will deliver digital dailies in Pro Res 422 HQ for the offline edit. The dailies will be synched at the lab (for preview purposes), but we may end up re-synching manually in FCP from the original audio files to ensure accuracy. We plan to go back to the HDCam SR master for an online and filmout to 35mm.

    I plan to work in a 23.976 sequence in FCP.

    In terms of audio: 1) at what sample rate and 2) with what framerate/TC reference should it be recorded to ensure a smooth workflow with no sync issues down the line?

    I appreciate any advice you can give me!

  • #2
    48Khz/16-bit should be the audio. Personally I'd let the lab do the sync- then you have zero worries... Otherwise it depends what frame rate the original audio recorder was using. Typically it's 29.97 or 23.98.

    Noah
    Last edited by NoahKadner; 04-04-2010, 12:09 PM.
    GH2, GoPro HD, AC160, Canon EOS 5D Mark II and Canon 7D training at www.CallBoxLive.com
    40% discount for 2-poppers! Use code 2pop2012 on these titles.

    Comment


    • #3
      Thank you Noah!

      I'm still confused: what determines the sync in terms of the audio--is it the sample rate or some audio frame rate?

      If sound is recorded at 48KHz (thus matching the 24fps film shot) how will it sync up with the 23.97 pulled down picture in FCP (or on a 23.97 HD tape)?

      I find that the lab's synching is not always accurate--so I would feel more confident if we did it ourselves.

      Thanks again--any advice would be super helpful!

      Comment


      • #4
        Sampling rate and bit-depth have nothing to do with sync- all they determine is the actual quality level of the audio. The same amount of audio will be recorded per second regardless.

        What matters is the timecode used on the recording device- that's where you set 29.97 or 23.98, etc- assuming we're talking about a professional timecode equipped audio recorder. And even that only matters if you're trying to post sync by timecode alone.

        If you're planning to sync manually you can just forget about that stuff anyways because you're just matching the actual audio not the timecode of video and picture. All the more reason why I highly recommend that if the lab is offering to audio sync your dailies for you- just let them.Quality going into FCP would be identical. If the lab can't get it in sync- please do yourself a huge favor and go to another lab. That's basic 101 quality-control stuff. If they can't even get the audio sync right I wouldn't want them touching my neg or color grading my dailies either.

        That and have a serious confab with the audio recordist on set. This stuff can be very simple if you communicate properly, or it can also be a total disaster if you don't.

        Noah
        Last edited by NoahKadner; 04-05-2010, 05:21 PM.
        GH2, GoPro HD, AC160, Canon EOS 5D Mark II and Canon 7D training at www.CallBoxLive.com
        40% discount for 2-poppers! Use code 2pop2012 on these titles.

        Comment


        • #5
          Thank you Noah very much.

          The lab recommends that audio gets recorded at 48.048 at 24fps and that they would do the pull down to 48K. Does that seem reasonable?

          Comment


          • #6
            What does your sound recordist say? Doesn't sound out of line but you should test that out before going too long.

            Noah
            GH2, GoPro HD, AC160, Canon EOS 5D Mark II and Canon 7D training at www.CallBoxLive.com
            40% discount for 2-poppers! Use code 2pop2012 on these titles.

            Comment


            • #7
              Originally posted by NoahKadner View Post
              Sampling rate and bit-depth have nothing to do with sync
              Bit-depth does not. Sample rate does. That's why they record 48048 then shooting 24 fps and pulling down to 23.98 during transfer.

              have a serious confab with the audio recordist on set.
              It is a sound post question. Production won't have any trouble to do either 48000 or 48048. It is post that must decide what's better. Actually I see some problems with your post workflow already. Why do you waste money on HDCAM-SR dailies?
              Michael Aranyshev
              Editor

              Comment


              • #8
                Thank you Mishka and Noah.

                I understand that sound is recorded at 48.048/24fps and pulled down during transfer to HDCam to 48/23.97fps (that's how it will be edited and mixed). So far so good.

                What happens when we do a filmout? Picture will go back to 24fps. How will sound get affected? (is there a way to remove the audio pulldown, while keeping the sound as 48K? If so, why record at 48.048 in the first place?)

                Thank you so much!

                PS. We're using HDCam-SR dailies since that's what we're mastering the project on.

                Comment


                • #9
                  All of that is stuff the lab should be doing for you... Not only should you not be messing with audio pulldown in FCP, it really doesn't even have the proper facility to do accurate audio pulldowns or pullups anyway... Let the lab handle it.

                  Noah
                  GH2, GoPro HD, AC160, Canon EOS 5D Mark II and Canon 7D training at www.CallBoxLive.com
                  40% discount for 2-poppers! Use code 2pop2012 on these titles.

                  Comment


                  • #10
                    Originally posted by souzos View Post
                    why record at 48.048 in the first place?)
                    Because there is a difference between re-stamping an audio file and converting sample rate. Re-stamping is fast and lossless. Converting sample rate is slow and can produce artifacts. It is better to do as the last step when you have full sound team working on your project, not just a telecine intern or an assistant editor.

                    PS. We're using HDCam-SR dailies since that's what we're mastering the project on.
                    They finish film on internegative but they don't strike internegative from every take. Telecine to something cheap. Edit. Make an EDL and get real scans on real scanner, 2k 10-bit log DPX. Grade them.

                    To capture film dynamic range on the tape you'd transfer logarithmic. That will look dull and washed out. Bad for editing. If you transfer with gamma correction you will lose a lot. Bad for grading.
                    Michael Aranyshev
                    Editor

                    Comment

                    Footer

                    Collapse
                    Working...
                    X