I am not a high-tech person. I recorded a Skype conversation and forgot to put on my headphones. I have my channel (Bill’s channel) and the other person’s channel (Mark’s channel). Mark’s channel is fine. But Bill’s channel also contains Mark’s voice when Mark speaks, because Bill’s computer is hearing not only Bill’s voice but also Mark’s voice as it comes through Bill’s computer’s speaker and into Bill’s computer’s microphone. So every time Mark speaks Bill’s channel records it along with the recording in Mark’s channel. Bill’s channel is markedly distorted when Mark speaks.
I tried taking Mark’s channel and inverting the audio, and then combining that with Bill’s channel with the hopes that Mark’s voice would be canceled out in Bill’s channel. As far as I can tell, it has made no improvement.
I am assuming that there is a way of getting Mark’s voice out of Bill’s channel, but so far I have been unsuccessful. Can anyone help? Thanks!
Cancellation is not going to help because the two voices went through two different sound pathways. Only one, for example, went through your microphone. They’re probably delayed in addition to quality differences.
We can’t split a mixed performance into individual voices, instruments or sounds for editing.
That is, we can’t do it automatically. You can do it manually. Drag-Select each “wrong” voice or words and silence it with Edit > Remove Special > Silence Audio. My tool has a hot-key Command+L. That will put a hole in the dialog where the selection was. See at half-second.
Do you like to interrupt each other? That’s a problem. You can’t both talk at the same time for this solution.
If you have a clean guest, Mark, you have the Hollywood solution of putting Bills voice in the trash and record it fresh and clean in post production from a printed script. How desperate are you? You can totally do that, I’m not making it up, but it does take some editing skills.
There may be no show.
How are you recording Skype? That’s normally an act of Congress/Parliament.
I did spend several hours going through and reducing the amplification in Bill’s channel whenever Mark was talking, but I noticed that the volume of Bill’s voice was getting less and less as time went on.
There were only a very few instances of Bill and Mark talking at the same time, except for Mark saying “uh huh” or something like that. I was able to take out the peaks of sound by zooming in real close and selecting only those peaks. That usually improved things, though sometimes Bill’s word would get lost. But the loss of volume of Bill’s voice grew gradually worse and worse, and I was not pleased with the overall effect.
I do not yet understand about “automatic ducking.” That sounds like what I was doing with the manual reduction of amplification in Bill’s channel whenever Mark spoke. I have been hoping that there is some method of improvement that can be applied at once to the whole channel.
I noticed that the volume of Bill’s voice was getting less and less as time went on.
That’s one reason I wanted you to talk about how you’re recording Skype. That’s not easy. I can’t account for that volume change other than serious instability when you cross sound channels by accident. You also have overload and distortion in Bills channel, too, right?
There’s ways to fix gradual volume change, but it’s not simple and we keep bumping up against the simplicity of just recording Bill over again.
You don’t have to do it from a printed script. You can use a very old simul-read trick. Play the interview into your headphones and announce Bill’s words as you hear them a second or so late. That sounds insane and it takes a bit of getting used to, but it lets you rip through the interview in nearly real time. A little timing shift here and there and it will produce a nearly perfect interview.
How long is the interview? Ten minute podcast or two hour stream of consciousness? If it’s a monster, that may be the end of the show.
I do not yet understand about “automatic ducking.”
You key a music track to a voice, so that the music volume “ducks down” whenever the announcer speaks, for example. It takes two tracks, the key announcer and the background.
Your key voice and the track to be patched are the same track. So that may be the end of Auto Ducking.
By what method, exactly? Recording Skype is a common question on the forum. There are some brute-force methods, but nothing easy from one computer. Are you following an instructional posting or YouTube video?
We should remember this is a forum, users helping each other, not a Help Desk. Can you help with this?
I got Skype recording to work and sound perfect, but I did it with two computers, a quiet room and a sound mixer.
There is an analysis tool in Audacity called contrast which can be used to measure the perceived volume.
It suggests you need to turn up the Skype-side (Mike) by 4dB to be equally loud with the studio (Bill).
Also I’d be tempted to include some crossover between left & right channels for those listening on headphones …
