Trying to massage spoken word to match existing speech

I’m helping my friend update his audiobook using portable Audacity 2.0.5 on Windows 7 (from

Four years ago, he recorded his book, That’s Smart! How to Improve Learning Habits at a professional studio. Now he has a home studio and has recorded some updates he wants added to the latest edition of the book, to be released shortly.

The problem is, the new material sounds a bit different to the original material–which I’m suspecting was highly processed–and I want to massage the new audio to sound as much like the original stuff as I can,

You can actually see the difference in the waveforms:

Original material (old stuff.wav)

New material (new stuff.wav)

I’m hoping someone would know what might work.



While it’s possibly to eventually get a reasonable match, it may take you a very long time. You should find someone with a good graphic equalizer in their sound system and play with that in real time. I bet you can a get a good match in under an hour.

Audacity has an equalizer, too, but it’s not real time. You have to change a setting, stop what you’re doing and see what happened, take the effect out and do it all again with different settings until you hit one. Kind of like driving a car by changing the steering wheel and then having to stop and see which way the car went.

Import first one track and then import the other. Audacity will stack them one above the other. Audacity will try to play them at the same time unless you tell it not to with the MUTE and SOLO buttons. Play the studio clip, stop, mute that and play the home clip.

Effect > Equalization.
The line in the middle is a rubber band and you can push it around with your mouse to boost or dip different pitch tones. Higher pitch tones and crispness to the right, low tones and rumble to the left.

I got some improvement with the illustration.
You have a new variation on “Make My Voice Sound Like…” question. This one is new. Make my voice sound like myself.

Screen shot 2014-01-31 at 7.13.10 PM.png

I’ll try that and tell you how it works!

One other thing that happens in home recording is “Room Tone.” That’s the noise that the room makes when you shut up. Studios don’t make any noise. Your house probably makes all sorts of little refrigerator/air conditioner/traffic noises. This may create problems in post production because you can’t do any processing that makes these noises louder.

Because of distortions built into the microphone(s), you can’t get a perfect match just by making some tones louder than others (Effect > Equalization). If you could, the market for expensive microphones would collapse, and it’s still doing just fine.

There’s another common misconception about microphones – that the more perfect they are, the better. Not always. The older, legendary microphones were far from perfect, but they made love to the human voice.


OK, it took two rounds, but I think I got it fairly well matched up.

Being a visual person by nature–it’s easier to understand if I see it, I used Analyze > Plot Spectrum to do a spectral plot of the voices I wanted to match:


Then I overlaid them with graphic software, making sure to line up the grid points.

I then used Equalize in graphic eq mode, setting the slider how I wanted to adjust each frequency, louder or softer. For future edits I’m going to use a log freq axis rather than a linear one to match the eq sliders more closely.

I knew going in I wouldn’t be able to match the voices perfectly; I just wanted to match them well enough so the edits aren’t that obvious. However, there’s a bit of a hiss in there I don’t like. any ideas how to get rid of it?

Did you try Effect > Noise Removal?


I did it before the last round of equalization.

Attach a WAV audio sample comprising a couple of seconds of the hiss only, followed by a couple of seconds of hiss with the signal mixed together. Please see: for how to make and attach a sample file.