Improve "weak" audio for training course

(UPDATE 2018-11-04: A consolidated posting of effects I ended up applying can be found in this post Voice improvement - consolidated effects.)

Hi, I’ve learned Audacity to improve audio for a project management course, and have applied 100 Mhz hum removal, normalization, de-clicker, and some tempo improvement.

The audio is much better, but still sounds “weak” to me. This might be an unfixable feature of the narrator :wink: However, audacity has so much power, I am wondering if there are other things that can be done. Could those with expertise with Audacity give a listen to the following, and provide your advice on what other processing would be recommended to improve it?

https://soundcloud.com/user-524851913/audio-test

Thanks so much.

I can’t listen to your file 'cause I’m at work.

and have applied 100 Mhz hum removal

There’s no MHz in audio. :wink: Power line hum is 50Hz or 60Hz, depending won where you live, and related harmonics (multiples of the power line frequency). A 100Hz high-pass filter will reduce everything below 100Hz, so if that’s what you did, fine.

Normalization uses the highest peak in the file for it’s reference, so if you’ve got one “loud” spot that will limit any increase you might get from normalization. You can try the Limiter, set to Hard Limit with make-up gain (It’s the make-up that actually makes it louder, after limiting). Maybe try limiting to -6dB, and if that’s not enough run it again.

Sorry, I meant to say I applied the equalization effect “100Hz rumble”.

The weakness comes across as a quavering and shakiness. I was hoping there was a remedy for this, something that would boost the energy in the voice.

quavering and shakiness.

Nope. Not here. I hear someone with little or no experience reading, but it doesn’t seem to be broken. Listen on a different device or computer.

That’s theatrical presentation—acting. Instructional presentation isn’t just slowing down. Here’s your clip contrasted with a CGP Gray instructional video track (attached). He is cheating. He’s carefully slicing off the interstitial bumps between sentences, but nobody said you have to mirror real, live reading (except ACX AudioBook).

That’s not to say you can’t edit your material and have it come out almost perfect. It can’t take more than a couple of months for the first segment or two. You should without question get the performer to do this. Nothing spurs self-correction faster than having to clean up your own bad reading.

Of course, if he’s paying you enormous sums of money, leave the arrangement the way it is.

Koz

Like that. I can do better. I did that on a dead run.

Koz

The original is 9 seconds long. The revision is 7 seconds and I could take even more air out of with a little effort. It still drags a bit.

Koz

I couldn’t resist. I took some of the air out of “proooooooject” and shortened the whole thing up again. It’s down to 6.2 seconds and starting to sound human.

Koz

There are pulses of bass-iness, which isn’t helping intelligibility.
You could just remove more bass, like the 100Hz rumble filter, but, say 250Hz instead.

Or, even better, use a multi-band compressor which will reduce the bass & DeEss at the same time, see …

G-Multi-band compressor in action.gif
project.xml (1.52 KB)

Is this sound work the track behind a video? Power Point? You didn’t say so, but if it is, then all the condensing and speeding up isn’t going to work because the sound sync will fall apart.

Still, listening to instruction at one-third speed is pure torture. If the production is portable, then each viewer may stop, pause and replay anything they didn’t understand.

Koz

I had separated the audio from the video for the course using Camtasia, the best inexpensive video creation software imho.

When I speed up the video with Camtasia, and then add back the audio sped up in Audacity, it syncs quite well.

I am playing with the various options - thanks for the tips.