Improving an old recording

“Sticks and Stones” is my favorite so far. It has a real voice and relatively little processing.
Is one of those cuts s***-kickin’, line-dance music?

You need:

A. A producer.
B. Good speakers.

The tonal balances are so out of whack I’m goin’ with you can’t hear what you’re doin’. The bass line for “Sticks and Stones” is like somebody dumped a pint of maple syrup on it. Cal Tech gave it a 1.8.


OK, got it. I dusted off the tone controls [cough, cough]. Pump the treble to bring the brightness up and restore a little of the sting to the drums. Turn the bass 3/4 the way off.

OK. I can listen to that.

“But the words you said just tore my heart in two.”


Are you producing these songs using MP3 as mix clips? Stop that. MP3 screws up music and you shouldn’t use it at all until the very last thing before posting and never for masters. It’s making some of your instruments “honky” and cell-phony.

WAVs or Audacity Projects, although Projects without backup WAVs can be dangerous.


You’re all the voices on “When did you stop lovin’ me?” right? Koz

Why do the instruments on “Full Blood Country,” sound so much better? It’s like night and day. Did you mic them different? Koz

If the infrasound was not removed from the vocal-only track and compression was applied, that could account for the vocal volume abnormally fluctuating : the vocal waveform is superimposed on the inaudible low-frequency infrasound waveform. If dynamic-range-compression is applied to a track with high-amplitude infrasound there will be noticeably more attenuation of the vocal where the infrasound has a peak or a trough.

The first thing you should do to any of your recordings before processing them, (e.g. applying compression), is remove the infrasound with the equalization shown below …
Remove infrasound with equalizer demo #.gif
No-infrasound-in-Audacity.xml (187 Bytes)

Koz - “Sticks And Stones” was originally recorded 15 or so years ago on cassette then converted to digital.

Koz - No, I’m not using MP3 mix clips. I record directly into Audacity creating tracks as needed for each item. Then export to MP3.

Full Blood Country was recorded by me but “fixed” by a pro.

Koz - No I’m not all the voices, just the lead singing on “When Did You Stop Loving Me”. And this is another one that was recorded on a cassette and converted to digital.

Guys, obviously I’m not very good at recording and mixing but I’m trying to learn and do a much better job. I’ve always had somebody else (studios, etc) do this stuff. I just play and sing. So please be patient. And any help is greatly appreciated.

Trebor - So the workflow would be to add effects, etc then run the “no infrasound” then compress that particular track??

The infra-sound has to be removed before any compression or other processing is applied: unfortunately removing infrasound from the final mix won’t correct the fluctuating voice volume :¬(

For some reason there is a LOT of infrasound, (sound with frequency lower than 20Hz), on your recording of “shoot her memory down” which can be seen on Audacity’s frequency analysis , (click on the image below to see the animation in its entirety) …
Remove infrasound, (sound below 20Hz), before processing, (eg before compression).gif
The first thing you should do to a recording of a performance is to remove frequencies below 20Hz with the equalization shown, or similar low-cut / bass-roll-off filter. By doing that you’re not removing anything audible, you’re getting rid of the low-frequency amplitude wobble which will subsequently interfere with any processing effect which has a threshold setting, e.g. compression, expansion, noise-gate, etc.

But I should be able to go back to the project file and fix it there, right?

Yes if you have the vocal-only track before you applied compression.

No, not before compression. It’s all just in the project file.

It’s all just in the project file.

I’d probably export original performances as WAV and archive them – or make a copy Project. We recommend saving new projects as you edit with dates and times – remembering to use the ISO dates (2014-01-27).


Projects do not save UNDO. Once you save the project, close the computer and go home, that’s the end of going back. If you make an editing mistake, you have to sing and play it again if you don’t have safety backups.

The grownups do this all the time. That’s how they come up with difference mixes for the vinyl, CD, club/dance and karaoke. They’re all original mixes taken from recording masters.

Full Blood Country was recorded by me but “fixed” by a pro.

Did you get to watch them do it? That would be worth writing a check.


If the compression has already been applied to the vocal-track in the project then one time-consuming solution is manually undoing the compression on the vocal-track using the envelope tool , but would take a lot longer than it would take to re-record it. You could try removing the infrasound from the compressed vocal-only track and then compress it again, that could help reduce the abnormal fluctuations in voice volume.

[ You should archive copies the unmodified recordings of each singer & instrument, (in WAV or FLAC format, not mp3 format), so you have the option of reworking them in the future without loss of quality ].

Koz and Trebor - Yep, I should be saving the original and am going to start doing that.

Trebor - the no-infrasound EQ curve XML gives me an error when I try to import it. Attached is the error message.

Nevermind. I re-downloaded it and it imported.

I wonder what could give you super low rumble like that. Are you recording in a room near the rail yards? Somebody repaving the interstate near your place? What else could do that? Wind? Air Conditioner? Recording outside?