How to batch process the audio track?

bsdcatharine · Post by **bsdcatharine** » Thu Dec 28, 2006 12:29 pm

I have edited over 100 of short didactic movies. In each of them there is a speech in one audio track that has been edited from few takes made during the same session with a headphone microphone for each movie (but not one audio session for all movies), so the sources are in different WAV files, but its background noise reminds the same. No music added.

Now I would like to batch create videos from the projects and have the audio normalized in each project. It is not going to be burned on CD/DVD, it is a web project. By normalizing I don't think keeping eye on the VU (true mastering... even though it would be fine to have this done for me), just reaching 0 dBFS.

My question is: I want to normalize the whole audio track of every movie, but
- Batch creating Audio track for each movie, manually running Audio editor, normalizing, replacing the original audio with the processed one and then creating the resulting video would be a real PITA. Isn't there some automation? If so, it would be wonderful as I could also add some audio compander to make it sound loud and clean...
- I can't do it by applying audio filter Normalize to the whole track - I am not allowed by Ulead... or am I?
- ...any other idea?

Thank you in advance!

Devil · Post by **Devil** » Thu Dec 28, 2006 1:03 pm

Normalise a single audio clip and copy it. Select all the other audio clips you wish to normalise and right-click and Paste Attributes.

bsdcatharine · Post by **bsdcatharine** » Thu Dec 28, 2006 11:33 pm

Hi again, thx for reply, but I am afraid this is not a good solution:

Suppose that speaker was talking loud first when recording take 1, but in time he was getting tired and started to speak silently. If you normalize each part separately, the background of part two will be louder, because "Normalize" will try to make speech louder.

In other words: I need Ulead to look at all selected chunks of sound as if it was one countinous audio, not a few individual chunks.

In even more other words: In this particular case I need Ulead to find the global peak maximum value of all pieces and then to change the volume of all pieces based on that global maximum value (not local maximal values of each snippet).

Devil · Post by **Devil** » Fri Dec 29, 2006 11:38 am

OK, adjust your first clip using Amplify to the level you want it, then do the copy/paste attributes as I said before.

sjj1805 · Post by **sjj1805** » Sat Dec 30, 2006 2:57 am

An alternative suggestion is to complete your project (cuts, transitions, titles etc.) and then export the audio track as a WAV file.
Open that WAV file with Audacity - a FREE audio editor.

You can now use the powerful editing abilities of Audacity to tweak the soundtrack, adjusting levels where necessary - it also includes a normalise filter.

Now save the modified WAV file and use it to replace the one in your project.

bsdcatharine · Post by **bsdcatharine** » Sun Dec 31, 2006 11:33 am

Ok, I understand that there is no "automated" solution within Ulead.

Thank you anyway for your answers.

As an inspiration for other future readers see http://forum.videohelp.com/viewtopic.php?t=314909. The point is to omit Ulead. Instead, batchprocess it using MS DOS scripting + 3rd party command line tools and then to replace old files with the new ones. Especially the capabilities of "Normalize" 0.7.7 are awesome as it counts with both program peaks and the RMS, therefore not only changing the gain, but also operating with some kind of limiter (although it is not possible to change its attack/release time, hard/soft-knee etc.) What's even better, it can calculate the optimum settings from all the files at the same time and then "truly master" the whole set of recordings as mastering engineers would do... but it's really an alchemy... but the manual is truly worth reading!

Usage: normalize.exe [OPTION]... [FILE]...
Normalize volume of multiple audio files

-a, --amplitude=AMP normalize the volume to the target amplitude AMP [default -12dBFS]
-b, --batch batch mode: get average of all levels, and use one adjustment, based on the average level, for all files
--clipping turn off limiter; do clipping instead
--fractions display levels as fractions of maximum amplitude instead of decibels
-g, --gain=ADJ don't compute levels, just apply adjustment ADJ to the files. Use the suffix "dB" to indicate a gain in decibels.
-l, --limiter=LEV limit all samples above LEV [default -6dBFS]
-m, --mix mix mode: get average of all levels, and normalize volume of each file to the average
-n, --no-adjust compute and display the volume adjustment, but don't apply it to any of the files
--peak adjust by peak level instead of using loudness analysis
-q, --quiet quiet (decrease verbosity to zero)
-t, --average-threshold=T when computing average level, ignore any levels more than T decibels from average
-T, --adjust-threshold=T don't bother applying any adjustment smaller than T decibels
-v, --verbose increase verbosity
-w, --output-bitwidth=W force adjusted files to have W-bit samples