Announcement

Collapse
No announcement yet.

Workflow to blend BGM with dialog/ambient sound?

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

    Workflow to blend BGM with dialog/ambient sound?

    Have a couple of issues to address:

    1. When in Audition CC, I use capture noise print (for the background noise), followed by noise reduction (35%) and compressor (to raise vocals). Given this process, it almost ALWAYS sounds like the talent is "drowning" or sorts. How best can I process the audio so that it doesn't sound like it's drowning?
    2. How do I normalise the dialog with the underscore? Should it be done in PP CC or in audition? If it is the latter, what's the best way to go about doing it?

    Thanks!

    #2
    THe "drowning" is probably artifacting from the noise reduction. THe solution is to get better recordings and not use noise reduction. Noise reduction is a last ditch salvage unless you are applying it very lightly to lower a bit of room noise. Basically if you apply it heavily then it always sounds bad. More expensive NR tools allow you to be a bit more aggressive with out it sounding bad, but that is still a pretty light pass.

    THe reason is that noise reduction is a subtractive process. Some of the noise is at the same frequency as what you want to keep so when you get rid of the noise you get rid of something you want also. A good analogy I've heard is that noise reduction is like trying to get the black out of grey. You can only bleach it so much before it falls apart.

    Using the compressor just amplifies the issues the noise reduction created.

    So...

    The best way is to get better recordings and then have someone who knows sound to do your mixing.

    I'm guessing the first is problematic and the later is not an option. But do think about making both happen if you plan on doing this seriously.

    So probably for you the answers are get better sound. Get the mic closer, turn off stuff that is making noise etc.

    If you HAVE to use noise reduction do so with light passes and, this is very important, make sure that your noise print is ONLY of actual noise.

    People talking in the background, and planes overhead are NOT noise. THey are things you should have redone the shot because of. You can not remove them.

    Noise is steady, and consistent (or you really won't be able to remove it), like tape hiss or air conditioning that stays on.

    Don't use a compressor for volume adjustments, it's not what it was designed for, use the level adjustments in the track or clip.

    As a quick and dirty "how to", you need to get the dialog at the level you want and then adjust the level of everything else to fit.

    That level is really a personal choice thing, unless you are submitting to broadcast, but the dialog sound bounce around in the -20 to -12 range with peaks not much over -6. That is EXTREMELY general but should get you into the reasonable ballpark.

    It is important to set your levels in the tracks and keep the knob on your amp or speakers or computer the same. Otherwise your tracks will vary in volume in bad ways.

    Based on your question I would probably stick with doing your mixing in PP. You need to mix with picture and bouncing back and forth between different programs is probably not the best way to go. If you know PP then really learn it and do it all in PP. When you want to up the game look to getting someone who will focus on sound and have them do the mixing. THey will use what they know and you get the best of both worlds.

    BTW really heavy NR has been used at times to make things sound "under water". Work that into your next script and wow the audience with your awesome under water recordings ;~)
    Cheers
    SK


    Scott Koue
    Web Page
    Noiz on Noise


    ďIt ainít ignorance that causes all the troubles in this world, itís the things that people know that ainít soĒ

    Edwin Howard Armstrong
    creator of modern radio

    Comment


      #3
      Originally posted by Noiz2 View Post
      THe "drowning" is probably artifacting from the noise reduction. THe solution is to get better recordings and not use noise reduction. Noise reduction is a last ditch salvage unless you are applying it very lightly to lower a bit of room noise. Basically if you apply it heavily then it always sounds bad. More expensive NR tools allow you to be a bit more aggressive with out it sounding bad, but that is still a pretty light pass.

      THe reason is that noise reduction is a subtractive process. Some of the noise is at the same frequency as what you want to keep so when you get rid of the noise you get rid of something you want also. A good analogy I've heard is that noise reduction is like trying to get the black out of grey. You can only bleach it so much before it falls apart.

      Using the compressor just amplifies the issues the noise reduction created.

      So...

      The best way is to get better recordings and then have someone who knows sound to do your mixing.

      I'm guessing the first is problematic and the later is not an option. But do think about making both happen if you plan on doing this seriously.

      So probably for you the answers are get better sound. Get the mic closer, turn off stuff that is making noise etc.

      If you HAVE to use noise reduction do so with light passes and, this is very important, make sure that your noise print is ONLY of actual noise.

      People talking in the background, and planes overhead are NOT noise. THey are things you should have redone the shot because of. You can not remove them.

      Noise is steady, and consistent (or you really won't be able to remove it), like tape hiss or air conditioning that stays on.

      Don't use a compressor for volume adjustments, it's not what it was designed for, use the level adjustments in the track or clip.

      As a quick and dirty "how to", you need to get the dialog at the level you want and then adjust the level of everything else to fit.

      That level is really a personal choice thing, unless you are submitting to broadcast, but the dialog sound bounce around in the -20 to -12 range with peaks not much over -6. That is EXTREMELY general but should get you into the reasonable ballpark.

      It is important to set your levels in the tracks and keep the knob on your amp or speakers or computer the same. Otherwise your tracks will vary in volume in bad ways.

      Based on your question I would probably stick with doing your mixing in PP. You need to mix with picture and bouncing back and forth between different programs is probably not the best way to go. If you know PP then really learn it and do it all in PP. When you want to up the game look to getting someone who will focus on sound and have them do the mixing. THey will use what they know and you get the best of both worlds.

      BTW really heavy NR has been used at times to make things sound "under water". Work that into your next script and wow the audience with your awesome under water recordings ;~)
      Hi there,

      Thanks for the detailed reply.

      My questions arise from run-and-gun situations where I don't have the luxury of time to do proper setups / or when I just want to video stuff spontaneously. Most importantly, I shoot for personal consumption and not on a commercial basis, so having another soundman is definitely out of the question.

      Noticed that you mentioned 'more expensive NR tools' that can do slightly better - may I know what that is? Also, it'd be good if you could share the workflow for such situations (to make the best out of it.. although we all know the rule - garbage in, garbage out).

      Thanks!

      Comment


        #4
        The iZotope RX family/suite is pretty much the norm for noise reduction. While it can't completely 'heal' bad recordings, it's just a good tool to have around.
        RJF
        Sound Editor/Sound Designer/Mixer
        San Francisco, CA
        My IMDb

        Comment


          #5
          I see. What would be the right workflow if I'm working in PP CC then? As round-tripping is probably not allowed with 3rd party apps, is it okay for me to finish all my editing first, export to my preferred format, then start working on the sound and replace the bad sound clips with the good ones?

          Comment


            #6
            The "traditional" workflow would be to finish your edits in Premiere. After "locking" your picture, you would then export an AAF or OMF from Premiere. This AAF or OMF would then be opened in Audition/Pro Tools/Logic/etc. In the mentioned audio programs, noise reduction/mixing/dialogue editing/sound effects editing/Foley would be performed along with iZotope'ing the files. After you mix the dialog,music,FX together, you would then export a .wav/.aif file which represents all the audio work done from the mentioned audio programs. That .wav/.aif would be sent to the picture editor (you) and be placed in your master timeline replacing all the existing audio. That is the "round trip" that we're all pretty much involved with from an audio stand point.

            With that said, if you don't want to use Audion/Pro Tools/Logic/etc to do the audio related tasks, you can just use iZotope RX as a stand-alone program. You can bring in individual clips into it, process them, then replace the bad clip in Premiere.
            RJF
            Sound Editor/Sound Designer/Mixer
            San Francisco, CA
            My IMDb

            Comment


              #7
              I'm not sure Rx is the "norm", but it is what I use. Others would be the WAVES NR suite and if you really want to go for it the Cedar NR boxes. In terms of price Rx is the most affordable, I think you can get into it for $300 - $400, WAVES is more like $1,500 and Cedar is like $4,000 for the plugin and like $15,000 for the box.

              All of them work as plugins and most also have stand alone abilities. So you can do a turn around with them.

              Ryan pretty much summed up the normal workflow. IF you wanted to do it all in one application then I would still do picture then audio, or reverse, and not work on both at the same time.

              Run and gun only really relates to production. You can take forever in post (not recommended).

              There are plugins (maybe not for CC?) that are intended to automate things like dialog levels. I have tried a couple of them since for me personally dialog is just not a lot of fun. But I wouldn't pay for any of them. They have had very mixed results and you will end up spending as much time tweaking as you would have in mixing in the first place. They are also generally a pain to set up and get working. I thought they might save a bunch of time doing an initial pass but in the end I don't think I have ever used the levels they generated. There are cheats that I have heard "recommended" like "leveling" software and plugins, but these are really not good. Maybe if your gig is super fast turn around promos or something?

              So the short answer is do mix by hand and ear. Meters are good to make sure you are mixing to about where you should be and they are really important if you have delivery specs, but the "art" is with your ears and you fingers. Something that can speed the proses, if you can get one that works with your software, is a touch control fader. You can get them pretty cheap. Something like this

              I have used and own bigger more fader units but for me I am 90% of the time only working one fader at a time so I probably would get a single fader next time around.
              Cheers
              SK


              Scott Koue
              Web Page
              Noiz on Noise


              ďIt ainít ignorance that causes all the troubles in this world, itís the things that people know that ainít soĒ

              Edwin Howard Armstrong
              creator of modern radio

              Comment

              Working...
              X