Bulk Audio Processing Of Video Clips

kooki wrote on 5/19/2008, 9:07 PM
I use Vegas 8 to process MPEG2 Video clips and SF9 to normalize the sound levels and add some compression with wave hammer.

Currently, I open about 50 files in Vegas. Then, one clip at a time, select the audio to open in SF, Normalize and Wave Hammer it into shape, save the new audio file, open back in Vegas, and repeat the process for all 50 MPEG2s.

As I am processing hundreds of individual clips, I'd like to automate the audio processing so that I can check the processed audio before doing the final MPEG2 rendering.

Can anyone suggest how best I can achieve this?

Thanks for any help.

Comments

farss wrote on 5/19/2008, 9:40 PM
You can use batching and scripts in SF however why not just add Wave Hammer to the Buss Master or where ever it works best in your signal chain. I use SF for audio microsurgery and NR but 95% of my audio needs I meet within Vegas.

Bob.
johnmeyer wrote on 5/20/2008, 5:48 AM
Doesn't Sound Forge come with Batch Manager? I have it, but can't remember whether I got it separately or whether it was bundled.
Chienworks wrote on 5/20/2008, 6:12 AM
Prior to Sound Forge 6 there was a batch processing program. It sorta fizzled out. Now with Sound Forge 8 and 9's scripting the same functionality is provided right in Sound Forge with scripts.
farss wrote on 5/20/2008, 6:16 AM
As someone who has SF I'm still curious as to why go to all that trouble when the same thing can be done in Vegas and non destructively too.

Bob.
johnmeyer wrote on 5/20/2008, 6:48 AM
As someone who has SF I'm still curious as to why go to all that trouble when the same thing can be done in Vegas and non destructively too.

I just did a VO yesterday. Needed to replace small clicks in my breathing with surrounding sound. I have no idea how to do that in Vegas. The SF Replace made easy work of it.

As to destructive vs. non-destructive, I think the answer lies in the fact that a sound file is exponentially smaller than a video file and therefore can be altered in a relatively short time. By contrast, changing gigabytes of video would take ages. At some point, however, if you give enough instructions to the non-destructive Vegas engine, it bogs down so much that you can't see what is going on. At that point, you have to "pre-render" which is basically "destructive" in the sense that a new file is created rather than just creating the effect on the fly.

the other reason why I think sound editors are destructive and video are not is that you can still effectively get work done if you only see every third frame of video, or if you only see 320x240 instead of HD. By contrast, you can't do a whole lot, besides just cuts, if you try to listen to an approximation of the final sound. I'm not sure I can explain why that difference exists, but I think that's at the heart of it.
kooki wrote on 5/20/2008, 6:52 AM
The problem with batch processing in SF is that in order to save the result, you have to render the file. If you render to MPEG2, it then undergoes a second render in Vegas. If you render to AVI ... it is very slow, even with a quad. If you render the audio to wav, you then have to delete the original audio file in Vegas and sync in the newly processed file. I don't really like any of those options.

If the audio processing is done in Vegas, you don't have the same ability to monitor and check the audio as thoroughly since the processing is applied during the render.

Will I lose much by rendering to MPEG2 in SF first ... and then importing the completed files into Vegas?
farss wrote on 5/20/2008, 1:15 PM
In Vegas you can just R Click the audio track and select Open Copy in SF. In SF you save the audio back as a file and when you exit SF it appears in Vegas as a Take perfectly in Sync. No need to delete anything or worry about sync.

Aside from that though what you hear in Vegas is as much what you get in SF.

Encoding back to mpeg-2 is indeed lossy for both vision and sound.

Perhaps if we knew where these files have come from in the first place we could offer better advice. There's further uncertainty as there's mpeg-2 which is video that can contain a variety of encoding schemes for audio and there's mpeg-1 layer 2 audio which is also called mpeg2. This is what the audio in HDV is encoded as as.

Bob.
Laurence wrote on 5/20/2008, 2:40 PM
You can do this quite simply with no visual quality loss in Vegas using three free scripts: the "normalize all" script, the "add regions to events" script and the "batch render" script.

First load all your clips on a single track of a fresh Vegas timeline. Run the "normalizie all" script. This will normalize the audio on all the clips on the timeline.

Second run the "add regions to events" script. This will make a separate region out of each event.

Third, run the "batch render" script using the "render regions" option and selecting the HDV mpeg2 template that matches your source clips.

These three steps will give you clips where the video matches the clips you started with (without generation loss because of the smart-rendering) and audio that is normalized.

You can also do many variations while you are at it. For instance you can use just the left side audio if you have a shotgun mic plugged into just one channel, or just normalize tracks with dialog on them.

There are links to these (and other) free scripts http://www.vasst.com/training/ScriptLinks.htmhere[/link] and http://www.jetdv.com/vegas/forum/index.phphere[/link].