I have a two channel audio file. It is a telephone conversation and each channel contains the audio of one speaker. Channel 1 audio is all nice and clean. But channel 2 audio contains the profile of channel 1 audio. Basically, the whole of channel 1 audio (much lower amplitude than the original signal due to attenuation in the telephone line) has been time shifted and added to channel 2 audio (leakage, echo in to the recording setup).
What is the best way to clean up my channel 2 audio?
In Audacity, click on the name of the 2 channel track and from the drop down menu select “Split Stereo Track”.
You will now have 2 single channel tracks.
Click on the name of one of the tracks and from the drop down menu select “Mono”, then repeat with the other track.
Now use the “Time Shift” tool (double headed arrow) to drag and slide one of the tracks left/right so that the “channel 1 audio” that is on both tracks matches as closely as possible.
This will not “remove” the leaked audio, but the clean recording of the “channel 1 audio” will now mask the leaked signal and it should sound much better.
Using the Equaliser effect it may be possible to improve the sound quality further.
Also, the Noise removal effect may help, but do not over-do it or you will hear the sound becoming sort of “metallic”. The noise removal effect in Audacity 1.3.x has been improved a lot since Audacity 1.2.6
I have two mono files after the split.
Let me see if I understand your suggestion, correctly.
I use “Time Shift” tool to move channel 1 audio and try my best to sync it with channel 2 audio. Then, I would have to attenuate the channel 1 signal using amplify and then invert it.
Finally, I can use quick mix to mix these two mono files?
I tried the above but it doesn’t work. I’m using audacity for the first time…wonder if I’m missing something here.
Yes, so that the “channel 1 audio” that has leaked over to channel 2 plays as closely as possible on both tracks - when you get it right there should be little if any echo effect noticeable.
No. What you are thinking here is an excellent idea, but unfortunately it won’t work. You are thinking of the “centre pan removal”/“vocal cancellation” type effect, where you can cancel out a sound by inverting it? This won’t work in this particular case, because for it to work, the signals that are to be cancelled need to be identical. Because the echo that you want rid of has been up and down a telephone line and through all manner of electronics, and possibly bounced off a few satellites on its trip round the world, there will be a considerable amount of difference between the two copies of the sound (in terms of frequency response, phase shift, and so on). They will be far from identical, and therefore cannot be cancelled out.
The best that you will be able to achieve is that the echo of track 1 that is on track 2 will be heard (almost) simultaneously with the sound of track 1. Because track 1 is much louder that its echo on track 2, it will hopefully mask the sound to a large degree so that the echo is not too noticeable.
If the sound of track one that has leaked through to track 2 is still too annoying, then you have a somewhat more difficult task. What you will then have to do is to reduce the playback level of track 2 in between each bit of voice 2 recording.
I’ll try and make that a bit clearer:
Let’s say - person A is recorded on track 1.
person B is recorded on track 2.
There is an echo of person A on track 2.
In between each bit of person B on track 2, we will reduce (mute) the playback level of track 2.
There are a couple of ways that this can be done - you can do it manually using the “Envelope” tool - this will be very time consuming. You may be able to do it using the “Auto Duck” effect that is available in Audacity 1.3.x - this will be tricky to set up accurately, but could be the best method. http://www.audacityteam.org/manual/index.php?title=Auto_Duck