Here's the situation. I recorded a VoIP call of mine, and instead of using the wave-out for the capture, I captured a mono direct mic feed. As a result, I'm faced the following issues:
1) Motherboard interference (integrated soundcard)
2) Barely audible second person. (audio was captured from headphones)
3) TOO audible "ME".
I've tried several trial plug-ins and haven't had much luck with any of them so far, most probably due to the fact that I lack experience in this type of setting. The audio as a whole is very important to me, and re-recording is not an option. I am willing to go to whatever lengths neccessary in order to improve the audibility of the second voice.
I've included a small sample to demonstrate the above issues.
http://mewlaw.googlepages.com/MediaSample_1.wav (http://mewlaw.googlepages.com/MediaSample_1.wav)
I realize that it might be a long shot, but is the vocal track recoverable? If so, can anyone advise on a good restoration suite/plug-in, and if possible, some good settings to try and adjust for better audibility?
Thanks for your attention.
Maybe you could run it through noise reduction then quite a hard compressor, to bring the levels into check. Having said that, I had a go with your sample and I still could not hear what the other person was saying clearly.
I didn't listen to your sample, because I'm at work.
Compression is the right process. Dynamic compession makes the loud parts quieter and/or the quiet parts louder. You need to set-up the compressor to increase the quiet parts. (GoldWave[/u] (http://www.goldwave.com)[/url] has an Increase Quiet Parts preset.)
Noise reduction is worth trying, but it may actually make things worse. If the signal is "lost in the noise", the noise reduction will remove some of the low-level signal along with the noise.
I am in complete agreement with the previous posts. I would recommend processing your voice and the other person's voice separately. I would start with a bandpass filter, maybe 300-4000 Hz. Then a judicious amount of noise reduction. You maybe be able to improve the sound enough for you to understand the words. If you can, make a written transcription of the conversation. That and the processed sound file is as good as you're going to get. I wish you luck.
Woody
I also agree with Woody,
I downloaded your sample and played with it for some time using 'Cool edit'
I tried notch filters, noise reduction, compression and a combination of all and yes there is some improvement but still much of the conversation is lost below the noise level and probably irretrievable.
If you don't have audio editing software you could simply play the file in Winamp and adjust the EQ as Woody suggested boost the mid range frequencies.. 310hz 600hz, 1k and 3 k, and minimize the rest.
Audicity provide a good audio editor which is free. http://audacity.sourceforge.net/ (http://audacity.sourceforge.net/)
At the end of the day the brain is the best sound processor , in my opinion some very simple filtering provided best intelligibility.
Good luck.