Batch Processing Voice Over Lines

Question

So you had a stereo track needing to be cut up into 300+ lines of triggerable game audio dialog.

You strip-silence, bake in your fades after adjustment of front and back end handle padding (seems like B's T's and S's can trick strip silence into pre-mature 0 crossings), and rename.

Now you have a stereo track of 300+ lines of solid edit dialog, named correctly, and ready to be implemented.... crap.... you realize that they vary in dynamic, in addition to some swanky noise reduction done by the voice actor him/herself blowing up the T's and S's and Ch's. You need compression, EQ or de-esser, and maybe even some fancy FX to beef it up a bit.

If you run it through an aux, youve just baked it into another stereo track needing to be cut up again. If you highlight everything and Audiosuite it, youve just baked everything into a track needing to be cut up again. If you normalize everything so you can level it all out before fx processing... well you get the picture, it seems like every option and technique to quell such issues seems to end in cutting it all up again.

Any ideas in how to keep the individual files intact, while batch processing?

Answer 1

Batch processing can only be used for certain processes. For Gain/Normalization, Batch Fades, Etc. I'd recommend Sound Forge. I've yet to find an editor since my switch to Mac as a main platform that I'm happy with.

What you need to do if you're using Pro Tools is develop a workflow that saves you as much time and involvement as possible.

One huge thing to realize is that Pro Tools has a function in Audio Suite where you can select to process it as one large file, or to process individually. This only works with certain types of processing though, as some files need to be analyzed per region (like the "Find Level" function in "Gain") and you can't automate that in Pro Tools like you can in Sound Forge.

So you have to sort it into steps based on what you can do as a batch and the most appropriate time to perform certain functions without jeopardizing the effectiveness of what is needed.

Here's an example Workflow in PT: (this might be a bit lengthy in description)

The first thing you want to do is separate all your parts to different tracks based on noise types and then run your noise reduction on them. That way you're avoiding to have to chop them all up yet again and you can run it as one big file per track if they have the same noise. Then you can chop and separate based on sibilance processing types, Sort those to different tracks and repeat the process with the next type of processing. If I remember correctly, you might be able to batch normalize, just make sure it's set for Peak and not RMS. Also make sure that you have the correct processing type set (per region or whatever they call it) and then hit ok/process.

The next thing you're going to want to do after your last round of processing and strip silence is to select all and Batch Fade.

Then sort your regions onto tracks that are named as close as possible to the naming convention you need these files named as. Ex: Dia_Duke or VO_Radio_Announcer or whatever the specs are. You're basically sorting by character (I'm presuming the game dev wants the characters name in the file so they can properly code it into the game).

Then go region by region and Consolidate (Shift+Option+3). What this will do is create a new file that's already mostly named because PT names a newly created file on a track after whatever the name of the track is and adds an underscore and a number to the end.

Then, if you require additional naming then you go file by file (I'd do this just after consolidating while the file is still selected) *Turn of Tab to Transient and Tab and Shift+Tab are your friend to jump region to region and select without having to touch your mouse. Rename (Shift+CMD+R), Then select all and Export Region Definitions (Shift+CMD+Y), Select all then Export Regions as Files (Shift+CMD+K) which will allow you to export all your cleaned up, processed renamed and trimmed files into a specific directory.

*Remember that holding down the Option/Alt Key + Clicking in Pro Tools is a "Do To All" Modifier based on what you have selected (be it track names or regions) and even works in some dialog prompts (like when it keeps asking you if you want to export the region definition for each file. Holding down Option will say "OK" to all.

All this is really is understanding your tools, what they're capable of and developing a workflow to save you time and efforts to make sure you are doing the least amount of work possible without having to repeat any efforts. Every time I have to do something I think it over and wonder how can I get this down into with the least steps possible with the tools I have. I also use a modification of this process to chop up my field recordings into libraries. Then I take them into my search program(s) and add the meta tags as a batch process and then modify them one by one as needed. It's also similar to how I do ADR Editing as well.

Someone else might have some better suggestions or modifications to this process to suggest also.

Hope this helps!

Answer 2

protools won't help you here, you'll need an outside program. Fortunately there are quite a few out there.

Soundforge is kind of the defacto standard for this type of thing, but its PC only so it may not be your cup o tea. The reason it rocks though is because you can set up a plugin chain and run the batch while setting the output to a new folder - resulting in a processed folder with correct file names all done in one fell swoop.

On the mac side you've got Izotope RX2 which has a batch process function and an ability to run VST plugins. This is really only viable if you need one plug or effect though, since you're not auditioning an entire chain.

Soundminer is another entirely viable option here that we've used before. Take your edited and named files and run them through a renamer, adding the suffix "-pre" to them. Chuck all of your pre files into soundminer and open up the vst rack. Add plugins to your heart's content, and jump around the various files to make sure that your results are consistent throughout. Once you're satisfied just hightlight all of them and spot them into a fresh protools session. Export into a new folder and remove the suffix. done.

Answer 3

We use Adobe Audition in the studio for this. You can set up and save a batch and even add plugins into the process, we've got altiverb running on a batch for example. Just open it up, select run batch, select the files you want and the export folder, all the settings and wham leave it running come back an hour later, all done :)

Very easy to use and set up.

Answer 4

Some other additions for the pile, for those of you stumbling across this answer years later:

Tsugi's Audio Bot http://www.tsugi-studio.com/?page_id=632 *Internal functions to crop, pad, normalize, fade in and out, remove quietest parts etc… *VST plug-ins host, so you can use your favourite effects. *Creation of plug-in / processing chains. *Command line version available for easy integration in your audio pipeline. *case-by-case pricing

Audio Ease's Barbara Batch http://www.audioease.com/Pages/BarbaBatch/ *Fades *dynamics / gain *extracts regions *tons of formats

Nugen LMB dialog Normalizer http://www.nugenaudio.com/lmb-loudness-management-automatic-batch-processing-file-application_21

Answer 5

For all game audio dialogue work, there is actually a specialized software called Alto It's from the same company listed above.

It does pretty much all you need + it can even read files directly from your game audio middleware project, check how sentences are concatenated in interactive audio, compare between languages etc...

It also arguably has the best audio file renamer your will find, with some really cool options.