|
|||||||||
|
Thread Tools | Search this Thread |
April 18th, 2005, 12:24 PM | #1 |
Posts: n/a
|
Sound Forge and Conversation (newbie)
Hi there!
I'm just starting out with audio editing, and I'm fiddling with Sony Sound Forge. I've got some DV tapes of interviews. Is there an automatic way to separate the interviewer's voice from the interviewee's voice? I heard a rumor that sound forge could do that, but rumors are rumors.. Thanks in advance, Rachel |
April 18th, 2005, 02:35 PM | #2 |
Trustee
Join Date: Jul 2003
Location: Burlington
Posts: 1,976
|
It depends on how the tapes were recorded, and what you want to accomplish with the project. Sound Forge is a two-track editor, meaning it is designed primarily to work with individual two-track files. It can be used to do what you are asking, IF the tapes were recorded properly OR the interviewer doesn't talk or comment while the interviewee is speaking. But it would be like using a scalpel to cut up a tree. It wouldn't be your first choice in tools unless all you need is one little piece.
Generally a multi-track editor, like Sony Vegas, is used for these purposes because it allows you to build a final program out of lots of two-track and single-track audio and video files. More like using a chain-saw to cut up a tree and then glue the lumber back together in a useful arrangement. It's much better for organizing many clips, which is usually what comes from an original interview recording. |
April 19th, 2005, 08:09 AM | #3 |
Posts: n/a
|
Thanks very much--that was very helpful!
The tapes are generally recorded on miniDV tapes using the internal (camera) mic. Thinking towards the future, what would be a better way to record the voice for these purposes? Sometimes both people do speak at the same time, and I was hoping there was an automatic way to split them (e.g., using frequency). |
April 19th, 2005, 09:07 AM | #4 |
Trustee
Join Date: Jul 2003
Location: Burlington
Posts: 1,976
|
If they were recorded with the camera's internal mic and they speak with some overlap, then there's no way to separate them during that overlap. You can work around this perhaps, but you'd have to get lucky as far as content and what you used in the final product. In other words there's not really a technical solution to fix this, you have to build a program by choosing what you can make work in the end, either with or without the overlap included.
Even if the interviewer and interviewee were recorded on two separate mics to two separate channels of the camera, you can't totally solve a problem that occurs due to overlap. It helps to have these two tracks separate, but the only total solution is for the interviewer to refrain from any comments during a response. I even warn people that there will only be head nodding and not the usual verbalizations that go along with listening to a response. It really is THAT important not to have overlap and to encourage the interview subject to give complete answers that can stand without any question being included in the final product. If you get a great response but it can't stand on its own, you can always re-record the question later if needed for better sound quality, or build your narration and the rest of your program so that any great answers make sense and are in context. Even if you isolate the interviewee with a single directional mic and don't attempt to record the interviewer, that mic will still pick up some sounds from the interviewer that can't be separated if they occur during a response. The same is true with telephones ringing, people talking outside the room, air conditioner and traffic noise, etc. EQ can help in these cases, but it can't totally fix the problem. |
| ||||||
|
Thread Tools | Search this Thread |
|