Audacity Forum

Hi,

I have a situation where I have probably 15,000 (ouch) short (1-15 second) .wav files that I need to batch process to be the same volume.

Is there a way to use Audacity to batch process a folder containing wavs to overwrite each existing wav with a volume corrected and normalised version of itself?

Thanks.

Making files the same "volume" is not the same as making them have the same "peak amplitude". One file can have a higher peak level than another but still sound quieter. "Volume" depends on frequency content and other factors as well as on "peak level", thus setting "the same volume" is not as simple as it may at first appear.

If you are on Windows there is a tool specifically designed for setting WAV files to the same volume: WaveGain.

This is the Audacity forum, so we don't directly support the WaveGain application, but as a starting point I find it easiest to use WaveGain with "Just calculate, don't apply", first, and then use "Calculate and apply". The first step will show how much gain will be applied to each track.
See also: ReplayGain which is used by WaveGain to calculate the amount of gain to apply.

WARNING: WaveGain acts directly on the input files so I would highly recommend that you apply it to copies of your files and NOT to the original files.

I am posting this here as there is no forum that I can find for wavegain

This is the results I am getting after running the radio gain

Gain | Peak | Scale | New Peak | Track
---------------------------------------------
-2.39 dB | 16327 | 0.76 | 12399 | 3225.wav
+1.08 dB | 9558 | 1.13 | 10823 | 3441.wav
+12.89 dB | 5642 | 4.41 | 24884 | 3731.wav

Obviously 3731 was the quietest but now it is the loudest. I thought they would all go to a peak that was the same, or very close. 3225 and 3441 are close and they sound good compared to each other whereas 3731 is definitely louder, as far as I can tell anyway.

Additionally, I guess I can use audacity after (or before) I have made volumes equal to remove background hiss/noise.

Many thanks again

nocdacity wrote:3225 and 3441 are close and they sound good compared to each other whereas 3731 is definitely louder, as far as I can tell anyway.

As I wrote previously, "loudness" is a complex issue, so whatever automatic method is used there will always be some audio samples that defeat the system.
It is also important that you are listening with speakers/headphones that have a "flat" response across the audio frequency range - for example, if I listen to bass sounds (even really "loud" bass) on my laptop's speakers, then there it will be extremely quiet or may even be inaudible because the tiny laptop speakers do not reproduce low bass frequencies. You can read more about loudness here: http://en.wikipedia.org/wiki/Loudness

nocdacity wrote: Additionally, I guess I can use audacity after (or before) I have made volumes equal to remove background hiss/noise.

Any other processing (such as noise removal) should be done before the equal loudness processing.

Cheers.

Are there any posts on batch noise profile removal with Audacity?

The effectiveness of Noise removal is highly dependent on how well the "noise profile" matches the actual noise that you want to remove. Unless the noise profile is exactly the same for all of the files in the batch then the effectiveness of Noise Removal will be significantly compromised. Having said that, you can still use the Noise Removal effect in a "Chain" http://manual.audacityteam.org/o/man/ba ... ssing.html

Thanks, will have a look

OK, I am having an issue where I cant capture a noise profile when creating a chain.

I go to create a chain, choose noise removal and then click edit properties and click capture noise profile but it is not capturing it.

Any pointers?

Thnaks

I think I figured it out. If I use the front end to capture the sound profile then simply apply the chain it will use the latest sound profile captured.

Actually, I was wrong, Audacity then asks me to save and close current file before I can batch process files, it then seems to forget the current captured noise profile.