8

For a project I'm working on, I need to produce approx 5000 voice-only recordings of approx 15 seconds each (typically each file is a couple of short sentences spoken aloud). I've approached this as follows (I'm using Adobe Soundbooth for all the steps below):

  • I make the recordings in as few takes as possible.
  • I then fix up each of these files, normalising them and applying noise reduction (specifically, I use the "Clean up audio" function and "Advanced Dynamics" which seem to give a nice, clear result)
  • I then cut the recordings into the individual files (which takes forever)

At this point I find that there can be quite large variances in the volume between the individual files so I use the "Volume Correction" task in Soundbooth to match them all to a reference file.

The issue is that, even after doing this there can be some files that are way off. So as a further job, I go through every file making manual volume correction. This takes a LONG time, longer still because there's only so much of it I can do in one sitting without losing the plot. So it's this last step that I really want to avoid. I don't want to skimp on the quality - I need to ensure that each file is very closely matched in terms of perceived volume - but I would really like to find a reliable way to automate this.

Can anyone help?

Thanks in advance for any guidance you can give.

P.S. I'm open to considering other software (than Soundbooth) if it will cut out any of the manual labour of producing these files.

Improve this question
6
  • "I then fix up each of these files"? Why don't you edit them all in one track, having made all takes with the same gain settings?
    – leftaroundabout
    Commented Apr 18, 2012 at 20:02
  • True, I could. As it's a fairly short job though (we're not talking about many files at this point), I just haven't bothered in the past. I've also been put off by the fact that the larger the file is, the slower each process is to run, but it's a good point.
    – BinarySolo
    Commented Apr 19, 2012 at 9:22
  • Well that's only a real issue in task-based programs such as Soundbooth. IMO general DAWs are much more practical for pretty much any task, probably also for yours.
    – leftaroundabout
    Commented Apr 19, 2012 at 11:13
  • OK, thanks. Do you have a recommendation on that front? And would it be likely to do a better job of normalising volume across all the other files, to the extent that they wouldn't need to be manually checked as I'm describing above?
    – BinarySolo
    Commented Apr 19, 2012 at 13:03
  • Yup, Reaper. No, it wouldn't do better in normalising volume across all the files, because it favours a workflow in which files as seperate entities have no particular meaning and you focus on tracks instead. Which is really much more comfortable for tasks such as yours, but it won't help you with the files Soundbooth already messed up.
    – leftaroundabout
    Commented Apr 19, 2012 at 13:48

7 Answers 7

Reset to default
7

You could use sox

sox infile outfile gain −n −3

will normalise the level to -3 dB relative to full scale deflection.

Improve this answer
4

Adobe Audition CS5.5 has a Match Volume feature for batch processing multiple files.

  1. Open Audition
  2. Effects > Match Volume
  3. Drag the files you want to match into the window
  4. Set the level you want to match all the clips to
  5. Click Match Volume

Alternatively you can click Batch and you will get options to rename the input & output files and choose the format.

Have a look at the Adobe help article for Automating Common Tasks (under the section: Match Volume across Multiple files). It explains the options and goes over 'Favorites' (like Actions in Photoshop) which can be used to batch process files a certain way (going beyond just matching volume). Worth a read!

Adobe Audition CS5.5 Match Volume

Improve this answer
3
  • Thanks for the answer. That looks very similar to the process I'm running in Soundbooth. Do you have any evidence that it would be better than the same feature in Soundbooth? Or if not, can I reasonably expect that for the type of task I'm carrying out (normalising volume of short voice recordings) that I can get good results from an automated volume matching process? Or will it always be necessary to do some manual adjustment after?
    – BinarySolo
    Commented Apr 24, 2012 at 6:51
  • I havent used Soundbooth before so I cant speak on its quality compared to Audition. Though Id say they are probably very similar considering they are both by Adobe. My thought process was that you would batch them all at the end of your project. But if you were to stick with the same settings each time (and not use a setting that is taking the average) you should be able to batch small quantities at a time and get the same results. It also really depends on how you are recording it. Ideally you'd record each clip the exact same way each time thus minimising the need to process them after.
    – OrangeBox
    Commented Apr 24, 2012 at 7:05
  • I am recording the exact same way, so the original recordings match reasonably well. I just want to boost the volume on them and given my equipment, doing that in post seems to be my best option. If I normalise the volume across the whole uncut file, though, it does it with reference to the peaks and it doesn't get me where I want. So I use the volume matching function in Sb on all the cut files and I get it to match each of them to a reference file, but that's what's giving me problems. I'll give the demo version of Audition a go and hopefully it will give me a better result. Thanks again!
    – BinarySolo
    Commented Apr 28, 2012 at 6:32
2

for normalization in todays crowded market - where we are experiencing "loudness wars" you do not want to be normalizing to peak levels. You normalise to a loudness standard.

There is an open-source project that will achieve this for you - bs1770gain which is the updated version of r128gain.

This will normalize to a loudness level and will thus consider all the compression techniques used in a clip. Once processed through this tracks will sound of similar loudness.

http://bs1770gain.sourceforge.net/

Improve this answer
1

For the normalization part you can also use normalize:

normalize v0.253 (c) 2000-2004 Manuel Kasper <[email protected]>.
All rights reserved.
smartpeak code by Lapo Luchini <[email protected]>.
Visit http://neon1.net/ for updates.

Usage:  normalize [flags] input-file

    -l <ratio>   don't find peaks but multiply each sample by <ratio>
    -a <level>   don't find peaks; amplify by <level> (given in dB)
    -m <percent> normalize to <percent> (default 100)
    -s <percent> smartpeak: count as a peak only a signal that has the
             given percentile (50%-100%)
    -x <level>   abort if gain increase is smaller than <level> (in dB)
    -p           prompt before starting normalization
    -b <size>    specify I/O buffer size (in KB; 16..16384; default 64)
    -o <file>    write output to <file> (instead of overwriting original)
    -q           quiet (no screen output)
    -d           don't abort batch if user skips normalization of one file
    -h           display this help

error levels: 0 = no error, 1 = I/O error, 2 = parameter error,
              3 = no amplification required, 4 = out of memory,
              5 = user abort

    - wildcards are allowed in 'input-file' (e.g. normalize *.wav)
    - 'input-file' needs to be a PCM WAV file.
Improve this answer
0

MP3 files can be bulk corrected to same volume with ' MP3Gain '

Improve this answer
2
  • I didn't think the OP was dealing with MP3 files.
    – JoshP
    Commented Apr 12, 2013 at 12:15
  • While this link may answer the question, it is better to include the essential parts of the answer here and provide the link for reference. Link-only answers can become invalid if the linked page changes.
    – Saaru Lindestøkke
    Commented Apr 23, 2013 at 19:08
0

There is also open source project R128 which supports new standard.

Improve this answer
0

I would personally use SoundForge to do this and use the normalize function.

  • Set to normalize everything to 0 dB.
  • Or if that sounds like crap, to 6dB then.

This should almost guarantee that your files are all the same perceived volume. Soundforge has the ability to do this in a batch function as well.

HOWEVER, I'm not certain if you have a Mac or a PC; if you've a Mac there's really no SoundForge per-se (1.0 doesn't count) so I'd go with what was said up above about Adobe Audition and batch processes.

Improve this answer

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.