I’ve inherited a REALLY bad recording that I need to fix before publishing the podcast. It seems that the audio crew set up two mics and not the mixer I had asked for, so the recording is picking up feedback from the second mic. I’ve been able to isolate the sound into a separate track, but am unable to remove it from the original track. Not sure what I’m doing wrong. Could someone share how to remove the sound in one track from another track?
If I’ve done this correctly, attached is an MP3 of the full audio recording, and a second MP3 of the sound I want to remove. Oh…I’m using the Windows version 2.1.0
Any assistance you can provide would be much appreciated. I’ve been using Audacity for about a year now. Great program. But I’ve never had to deal with a recording this bad before!
If you could identify the sounds enough to isolate them, why didn’t you just press delete, or maybe use that as the profile for noise reduction?
You can’t use cancellation if you only have a similar sound. Cancellation is tricky and it fails more often than not. The two sounds have to be surgically perfect. It doesn’t work with MP3s.
I’ve been able to isolate the sound into a separate track
You should probably think of what you would do if we can’t fix the show. You appear to be in violation of #4.
The Four Horsemen of Audio Recording (reliable, time-tested ways to kill your show)
– 1. Echoes and room reverberation (Don’t record the show in your mum’s kitchen.)
– 2. Overload and Clipping (Sound that’s recorded too loud is permanently trashed.)
– 3. Compression Damage (Never do production in MP3.)
– 4. Background Sound (Don’t leave the TV on in the next room.)
So it would be good to tell us exactly what the sound is and how it was recorded. If it’s constant background sound like a fan or air conditioner, we maybe able to help with Noise Reduction. If it’s other conversations or traffic noises, there are still tools available, but they’re difficult to use and can leave a funny sounding show.
I hear you! Unfortunately, I had little control of the audio (there’s not a shot I can publish as is).
Okay, so attached are 1 minute portions of the original audio and the sound I want to remove. The original file is 1:40:29 and is too large to attach as a file. If you can fix this (and I would be eternally grateful!), please let me know the steps you took so I can recreate on the entire file.
Thanks a million. Anxiously awaiting your expert advice.
I think you misunderstand what happened. That is extreme echo cancellation and compression distortion. It’s what happens when someone gives a speech over chat, possibly to a conference, and leaves their speaker and microphone open at the same time. If you pay attention you can hear the last phrase of the last word of each sentence coming back multiple times in a bell-like echo.
If the system didn’t echo-cancel, the channel would start screaming in feedback (…eeeeEEEEEEEEE) and do that over the whole show.
I think it’s too far gone. I had high hopes for using the noise sample on Noise Reduction as the Profile, but it didn’t work. Also Noise Gate didn’t work. It only gets rid of the quiet dialog distortion but leaves the trashed words.
Past not using the recommended equipment, the recordists also didn’t wear headphones, even briefly, to make sure they were getting a good recording.
I tried a notch filter to take out the worst of the bell tones. See what you think.
The Desperation Method is transcribe the words (if you can understand them) to paper and have an actor read them into a good microphone.
It’s also possible you trashed the forum submissions when you made the MP3…
That’s why we discourage using MP3 for anything except listening at the beach.
It’s better. Ugh! This is the worst recording EVER. Could you tell me the steps you took to clean it up? The file was originally an MP4. I can try to export as a WAV. Do you think that might help?
I can do that when I get back. I can make out one or two more of the words. That may be enough to write it all down and have an announcer read it for the sound submission. Or if you have a good Text To Speech program. The original performance is trash.
I didn’t publish initially because it’s a lot of work for nearly no benefit.
– Import the Background noise clip into Audacity
– Select the whole clip or show by clicking just above MUTE.
– Effect > Normalize > [X]Remove DC, [X]Normalize to -1 > OK
– Drag-select the first 20 seconds or so.
– Effect > Noise Reduction: Profile
– Import the show.
– Select the whole clip or show by clicking just above MUTE.
– Effect > Normalize > [X]Remove DC, [X]Normalize to -1 > OK
– Select the whole clip or show by clicking just above MUTE.
– Effect > Noise Reduction: Settings 18, 6, 6 > OK
– Select the whole clip or show by clicking just above MUTE.
– Effect > Notch Filter: Frequency 537, Q 3 > OK.
Aaaaand. We’re done. No, I can’t make out all the words, either. The notch filter takes out one of the more obnoxious honky tones.
I’m curious how you shot that clip. At the beginning you suggested something simple like the recordist set out the wrong microphone or something. This is far more serious than that. This is like a compressed camcorder in a large room version of an insanely bad conference call.
And as before, you stabbed yourself repeatedly by not having WAV or other uncompressed format anywhere in the process. Each pass through MPEG (M4A, etc) or MP3 causes additional distortion and makes the rescue effort harder and harder … or impossible.
There was a fantasy story of a guy that had a zombie as a housekeeper instead of a regular person. The zombie was, of course, living dead, so they spent all of their time cleaning up after parts of their body fell off. Very little of the actual housekeeping got done.