Help matching acoustics between two voice recordings

I’m editing together an audio lecture for one of my classes in Audacity. I unfortunately had to record different pieces of the lecture in two different rooms and the ambience of the resulting recordings differs somewhat. Assuming I don’t have a chance to re-record, I’m wondering if there’s a way to make the two existing recordings sound more alike using filters or effects. Attached to this message are samples of the two different recordings.

To be clear, the recordings aren’t of a live lecture recorded in front of people, but rather me sitting in a quiet space in front of a mic.

And here are my version numbers:
Audacity 2.1.0
OS X 10.6.8

Many thanks!

That’s not dreadful. I was all set to rain on you for including echoes in your performance, but this seems to be a straight equalization job (it says here).

Let me see what I can do. If I get a working filter, I’ll publish so you can do it, too.

Koz

I assume Sample 2 is the ratty one only because it sounds ratty.

Is that close enough? I can’t get any closer without tons of work. You didn’t just change rooms did you? You went from an expensive studio to your inexpensive living room and you want us to make them sound the same.

That roaring in a cardboard tube sound in the background is your air conditioner or your computer fan. That’s after hefty Noise Reduction. In the raw pitch correction it’s much worse. I can get sorta close match with custom equalization (pitch change), volume compression (denser) and noise reduction (background sound). You didn’t just move the recorder to a new location. Everything is different. You left the studio, got in your car and went home.

It’s not a simple process and it’s very typical of what happens when someone tries home recording.

Let me know.

Koz

You can make this worse. My example sound file is a producer who decided they didn’t need that darn ol’ studio and they were going to record at home.

http://kozco.com/tech/audacity/clips/EchoSample.mp3

This person will always sound like she recorded in her mum’s kitchen. There’s no filters for that, and no, it doesn’t match the rest of the promotional announcement at all.

Koz

Thanks so much for your help (and the stern advice). I’ll try these out and let you know how they go; whatever the outcome, I really appreciate the help and the learning experience. I did in fact change rooms, but the first was an office space that, while not optimal, I could at least shuffle around for recording purposes and the second was literally a closet. The background is that I’m moving for work. I start my new job in a couple weeks, and the office building has an actual recording space, so with luck I won’t be facing issues like these too frequently going forward.

My previous reply seems to be held for moderation for reasons beyond me. Thank you for the help and the stern advice. If it’s not too much trouble, I would appreciate a plugin to replicate the changes you’ve made. The move between the two recordings was from an office space, which wasn’t ideal, but which I was at least able to rearrange for recording, to what was literally a closet. The context is that I’m in the process of moving between jobs. The new gig will have a studio recording space, so hopefully these are all issues I can avoid in the future.

Thank you, immensely, again, for your willingness to tinker with this!

Here’s an equalization-curve which will convert sample2 into sample 1 …
UnRat300 equalization curve.png
Rather than redraw that curve , (or an approximation of it), you can import that “UnRat300” curve into Audacity’s equalizer via the xml data file below …
UnRat300.xml (52.9 KB)
UnRat300 is close to the desired result , but a final tweak is required …
Final Tweaking EQ using Bass&Treble.png
Here’s a before-after …

then a Limiter.

Exactly. I couldn’t do it without compression, either, which is why I thought everything changed, not just the room. The whole character changed.

That’s a more comprehensive version of my manual curves, and I used Effect > Compressor, not Effect > Limiter, but if Limiter worked, go with it.

Koz

My previous reply seems to be held for moderation for reasons beyond me.

We didn’t like the colour shirt you were wearing.

Koz

The curve UnRat300 was primarily created via Steve’s experimental plugin “SpectrumToEqCurve.ny” , but needed a manual adjustment: a peak at 300Hz, ( because the data-points on the curve generated with Steve’s experimental plugin are too sparse at the low end , even when using the maximum 2048 points ).

And as well as that, you have not made enough approved posts yet, which is the real reason.

Gale

Thanks to you all! Sorry for the comment about the post being held for moderation, that was totally snippy of me and unwarranted. And FWIW my shirt was, indeed, pretty hideous. :blush:

Trebor’s solution is hugely helpful and kozikowski’s hugely informative. I really appreciate the assistance and instruction from both of you. I do hope to get a chance to re-record some of this stuff, but your help means that if I don’t get that opportunity I can at least make the recordings work together passably. And, as always, I came to the help forum to learn something and did, thanks to you, for which I’m also very appreciative.

There’s still one other parameter between the recordings which doesn’t match : the stereo-image.
Your recordings at the different locations are stereo, which gives-away they were made in different places.
A quick solution is make all the recordings mon[o]aural, they would then sound spatially the same.
You can do that by applying “Stereo to mono”, (which is in the in the Audacity “Tracks” menu ).

Excellent suggestion, thanks!