Travis has an automatic method for ducking based on compression related to total volume (I think). It's on a very old thread that I can't find right now, but he's here quite regularly and may pick this up and describe it again.

I couldn't get the method he uses to work great with my material but you may have more success.

I experimented with ducking using trial versions of Ultimate S & Excalibur(?) a few months ago. Neither quite did what I wanted. I think I ended up using Excalibur to create the the envelope points, but I had to manually move them along the envelope (<ALT>+drag) as the duck was starting too early and finishing to late around the speaking.

Nick

Ultimate S 2.0 has variable attacks on the ducking, Nick. You might try setting the envelope to be slower on the incoming, and faster on the outgoing sections of the envelopes.

You can also use a side chainer for compressors, or feed the compressor to a bus and use that as the trigger.

Yes, I remember that Spot, and in fact I manually drag my envelopes to match what you're saying.

For me the poblem with both products' method (I seem to remember) was the the edge of the event was triggering the ducking, not the volume. Sadly not all my events have the same length heads and tails before and after the speaking.

It would be wonderful if someone came up with a ducker that a) was triggered by a user-definable decibel threshold AFTER the start of the event (automatically putting the start of the bottom of the duck there), and b) detected the beginning of a user-definable period of silence (below the threshold) and put the start of the "recovery" at that point.

With an aching wrist having literally spent the whole day dragging envelope points to create ducks I'm considering having a go at scripting myself to try and achieve this :-)

Oh, and while we're at it, my other wish for auto-duckers is user-definable envelope types for both the beginning and end of the duck (e.g. smooth at the start, quick at the end or whatever).

Nick

I think I ended up using Excalibur to create the the envelope points, but I had to manually move them along the envelope (<ALT>+drag) as the duck was starting too early and finishing to late around the speaking.

Excalibur will let you control the amount of time it takes to reduce and restore the volume levels. It will also let you complet the reduction before the beginning of the voice event or center the reduction on the border of the voice event.

Wow, thanks guys... glad I asked! I guess I'm not seeing the utility of a ducker based on event splits rather than volume levels. In this particular case, the audio is recorded outdoors and there is a nice ambience created with some real external ambient noises. I think it would be quite noticeable if that "white noise" shut off and turned on around vocal points.

Even if I wanted that, I'd probably rather use a noise gate to bring the volume to zero just so I didnt' have a bunch of extra vocal track tidbit events to keep track of... rather have one big event tied to its corresponding video.

Jetdv... how about Excalibur? Same idea generally as S2 on this point?

Thanks,
Jeff

Yes. Scripting cannot "see" the audio so you have to "cut out" the areas where you wish the music volume to be increased.

> I guess I'm not seeing the utility of a ducker based on event splits rather than volume levels.

There is no one size fits all. If you have an audio voice over that is broken into events with a music bed under it, then Ultimate S2 and Excalibur can duck the music when it encounters a VO event. If your VO is one long event and you don’t want to split it up, then you’ll have use some other way. Ultimate S and Excalibur don’t have access to the audio data so they cannot analyze it. All they see are events on the timeline.

> I think it would be quite noticeable if that "white noise" shut off and turned on around vocal points.

Since the vocals are not being ducked, the music is, you could duplicate the vocal track and slice it into events to use with Ultimate S or Excalibur for ducking the music. Then just mute that duplicate track and use your original track in the mix.

> Jetdv... how about Excalibur? Same idea generally as S2 on this point?

Excalibur and Ultimate S2 both work at the event level. When they encounter an event on the VO track they duck the music on the MUSIC track. If the VO track is one long event, both Ultimate S and Excalibur will just duck the music for the entire time.

~jr

Howdy...

Here's a solution to your problem, although it costs $40...

You can get an easy-to-sidechain compressor here:

http://www.db-audioware.com/dB-D-dynamics-processor-more.htm

There's a demo if you want to try it out to see how it works...

Now... Add this as a track effect to your vocal track, and click 'send' until you see an 'A' in the box. This means that your are sending the signal from the vocal track to the side-chain 'A'...

Now go to your music track. Add this effect again. Click on the 'receive' button until it says 'A'. This means it will compress the music track, but the input level will be controlled by side-chain 'A', which as we previously set up, is the vocal track...

Now adjust the compressor on the music track. Raise the compress very high (all the way to 'limit' if you like) and adjust the threshold. The upshot should be that the music is loud until the vocal track speaks. This then 'trigges' the compressor, which compresses (ie: reduces in volume) the music track. The louder the vocals, the greater the compression of the music.

In this way, the music will still be nice and loud when the speaker is silent, but as soon as the speaker talks, the music backs aways, or 'ducks' under the vocal track... Very handy.

I use this alot for fashion-show type shoots. Need to have loud, pumping music, but also need to hear the description of the outfits by the M.C. This gives you the best of both worlds...

Hope that helps..

Best Regards,

Ben Freedman
www.rooproductions.com
www.gettingstartedvideo.com

You can also achieve a similar result without the dbaudioware tool, although that is a good app.

Put voice on tk 1
Put music on tk 2
Drop the volume of the music level to -6dB (assuming it's a fairly hot music track)
Apply compressor to master bus
Set threshold level to at least -8, maybe more like -10
Set the ratio (amount in Sony comp) to around 10:1
Be sure you've got Smooth Saturation enabled, be sure that Auto Gain is turned on/enabled.

Depending on levels of the V/O recording, and levels of the music, this should be a good place to start. Keep the attack fairly low, same with the release. Experiment a bit with other settings.

The dbaudioware tool will be a bit better overall, because it's using actual inputs from each signal.
It would be sweet if Vegas could support side-chaining in the app itself.

Scripting cannot "see" the audio so you have to "cut out" the areas where you wish the music volume to be increased.

A couple more thoughts:

Can scripting see a volume envelope? Could the envelope output by plugging in Blue Cat's Digital Peak Meter (as mentioned by Spot) be used not itself (because it's probably too "up and down") but to trigger the start and end of a flatter ducking envelope based on volume thresholds?

Secondly is there a way to automatically split the long voiceover event into shorter events based on a volume threshold, in preparation for applying a ducking script? If scripting can't see a decibel level then I think I know the answer to this already.

Jeff, please let us know how you get on and which method you settle with.

Nick