The presentation has “ice pick in the ears” sibilance and “essing.” I can well imagine what listening to that on headphones would be like.
You also have compression (talking into a milk jug or wine glass) distortion.
This is quite a low quality 56kbps export.
Oddly, that’s not it. You can produce a perfectly acceptable non-stereo presentation at 56 and nobody will notice. The mono limit is somewhere in the 30s. However, you can’t produce your voice track in MP3. Every time you edit or do corrections or production in MP3, the sound quality goes down and you can’t stop it. If you do enough careful editing and cleaning, the voice will start to sound like bad cellphone.
Record your voice and export a safety copy as WAV (Microsoft) 16-bit. Mono WAV quality is somewhere up in the 700s, and it doesn’t change. Do all the editing in Audacity Projects or WAV and then, way at the end when you’re getting ready to ship the final product, then make the MP3. So that should solve the honky sound. That, not coincidentally, is the recommendation for audiobooks.
Sorry, once it does that, the track is trash There is no good way to take that kind of damage out.
Are you using a home USB microphone such as the Blue Yeti? Those are famous for overly crisp, sharp, gritty, piercing sound. I think the makers assume that sounds “professional.” It makes my ears bleed.
There is a DeEsser which can help.
I have some notes on how to use it here, somewhere.
The DeEsser defaults are not the best and that’s the set I use.
One shortcoming of the DeEsser is the need to produce the voice track first, before you apply the tool. That’s what the audiobook mastering suite is doing. That’s not a bad way to stabilize your track volume for your products, although it does less well with music. Do the voice corrections first, then add music.
That Audacity de-esser plugin, (By Paul-L), does not have a “linked stereo” option:
it will add a weird spatial [sic] effect if applied to voice which has had stereo reverb added.
You should apply that de-esser to mono vocal before any reverb (or music) has been applied.
There is a simpler faster but less-precise de-esser plugin here, (by Steve).
It always has “linked stereo” which avoids the unintentional spatial effect.
Thank you for your reply. I’m not entirely sure what ‘linked stereo’ is. Is it “joint stereo”?
What I have is four stereo tracks. One has my voice. One has music. And the other two are my voice but one just in the right and one just in the left.
When I exported it to MP3 before, I used “stereo” not “joint stereo” because I had read that was better when you want something to clearly be in one ear or the other (like my last two voice tracks).
SO
At the moment, open in my audacity programme, are those 4 tracks.
What I’m not sure that I understand, is the exact process I should go through. You say it should be applied to a mono voice… do you mean to split my main vocal stereo track into mono and desser it… or can I de-esser it while its still stereo.
Do I need to delete the music and secondary voice tracks and then de-esser it, because it’s going to apply it to the WHOLE 4 tracks at once, and I only want it on the main vocal… or can I just select the main track and de-esser it.
I’m sorry. I had no idea this would be this tricky. I sailed through putting it together, was happy with it, then listened to it on headphones and EEEEEEEEEEK!
I’m trying to install the de-esser. I can’t even figure that out. I’ve followed a link in this forum to a page in a wiki that says to install it using Tools>nyquist plugin installer…
But I don’t have that under Tools. (Only “nyquist prompt” - which I don’t know how to use - I tried opening the desser.ny file in that and it loaded up what looked like a text file, that when I ‘ran’ it, just came back with a dialogue box saying De-essed Nyquist returned the value 10, but didn’t seem to actually do anything… What is that???)
The de-esser effect is threshold (i.e. volume) dependent: it can have no effect if the threshold is set too high.
It’s a trial & error process …
Lower the de-esser threshold until it’s removing too much ess: i.e. until it sounds lispy, (or should that be lithpy) ,
then raise the threshold a bit until the lisp is gone.
If possible only use de-esser effect on vocal track only, as it can distort music.
Blue snowball is a mono (not stereo) microphone.
There can be a stereo-like pair of tracks from it, but they are absolutely identical, (a/k/a dual-mono).
Making a low bitrate stereo mp3 version can add a pseudo-stereo effect, particularly on sibilance.
That explains why the original mix sounded OK, but the 56kbps version is “EEEEEEEEEEK!”.
The only way I ever got the classic DeEsser to work and sound reasonable is to run a mono (one blue wave) voice track through our audiobook mastering suite first.
The suite’s job is to horse whatever I recorded into perfect loudness and condition to submit for audiobooks if I wanted to. They have pretty strict rules about this. Then run the DeEsser and I don’t have to chase the volume of the performance all over the place because the mastering tools got there first.
If you don’t do something like that, then as Trebor above, you have the performance volume and the corrections chasing each other to get the effect you want.
I did miss the description of why you needed four tracks. If you are playing with stereo effects in the show, it could be a career move straightened it all out.
Blue makes the Snowball and they make the Yeti microphones. So you have a snowball. About as big as a softball with the cable out the back and you speak into the company name.
I’m not sure which direction to go with this. I’ve never had to unscrew a complex production this far along.
