Compression vs. Limiter and MP3 kbps Value on Export

Audacity Help Forum Windows
1

Good afternoon (Koz? :slight_smile: )

Hope this finds you doing well and having a super weekend! Have a new question for you. I’ve read in an ACX article/post that they actually “prefer” that we use “Limiter” vs. “Compression”. Would you be so kind as to expound on your feelings/knowledge about either/or? I’m attaching a .png of what my Audacity default Limiter settings are – note that it always came up on the “Hard Limit” selection. I’ve also attached a .png of my Compression settings that should match ones I got from your site. Oh, and can you clarify if Normalize should be set to -3.2dB or -3.5dB – I seem to remember having seen it both ways only in different posts.

I’m being able to pretty much consistently get ACX Check “passes” on my raw recordings. Have really had to play around with my AT-2035 mic to get it just right in positioning. Honestly, it’s been a bit disappointing and frustrating to me that it appears to be so finicky. Then again, it could be me! :slight_smile:

Oh, and I do have one more question. I read a post you posted back in June to someone in response to them saying that their recording met the ACX Check test requirements (passed), but they thought their MP3 sounded pretty awful. I tend to find that mine seem somewhat muffled – or, just not as crisp and clear as I would have expected. Anyway, you commented to them about the 192 kbps output to MP3 is a “minimum” and that you can go to 256 kbps or 320 kbps as long as you do not exceed the 120 minutes/chapter and 170MB file size. So, did that mean you recommend going higher if one can meet the other time-frame and file size limits?

OK, that’s enough words (maybe too many) for one post (ha! ha!). Oh, and as a reminder for you, I am using the following:

Windows 7 Pro - Audacity 2.1.2
Mic = AT-2035
Focusrite Scarlett 2i2 audio interface (1st generation)

Thanks so much for your help, as always, and I hope you have a terrific day!

Sheila
Compressor_Settings.png
Limiter.png

2

Disclaimer - I am not an audio book expert.

Every situation is different so nobody can give you the best compression/limiting presets. It has to be done by trial-and-error, and by ear. The new/current limiter is quite good and I’d tend to favor it over “regular” compression, or at least I’d try it first.

Oh, and can you clarify if Normalize should be set to -3.2dB or -3.5dB – I seem to remember having seen it both ways only in different posts.

I believe the spec is -3dB, right? Since MP3 is lossy compression, some peaks will increase and others will decrease. So overall, you’ll usually end up with some higher peaks after MP3 compression, especially if you’ve used compression or limiting. You can re-open the MP3 and check the new peaks* after MP3 encoding, and if it’s too far off, go back to the uncompressed original and try again with a different normalization setting.

tend to find that mine seem somewhat muffled – or, just not as crisp and clear as I would have expected.

Try boosting the high frequencies (with the Equalization effect). That will help to bring-out the “T” and “S” sounds. But, be aware that when you boost anything you also boost any existing background noise and any background hiss will be boosted along with the highs.

Anyway, you commented to them about the 192 kbps output to MP3 is a “minimum” and that you can go to 256 kbps or 320 kbps as long as you do not exceed the 120 minutes/chapter and 170MB file size. So, did that mean you recommend going higher if one can meet the other time-frame and file size limits?

As you may know, bitrate (kilo_bits_ per second) is related to file size and quality. If you know there are 8 bits in a byte, you can calculate file size from the bitrate and playing time. MP3 is lossy compression and the lower the bitrate the smaller the file and the more data you’re throwing away. So higher bitrate generally means better sound quality (and of course, bigger files).

With voice (and with most music) you probably won’t hear any difference between 192kbps and anything higher. But even if you can’t hear any difference, I’d use the highest bitrate allowed unless you want smaller files. The ONLY advantage to lower bitrate is smaller files.


\

  • I use the Amplify effect to check the peaks. i.e. If Amplify defaults to +3dB, the current peak is -3dB.
3

Wow! Thanks so much for all your info. Perhaps if Koz reads this, and if he has anything to add, I hope he will. It’s not that I don’t value what you provided to me, I have just gotten the feeling from him that he may have more experience with the ACX stuff? Then again, I could be ever-so-completely WRONG! :slight_smile: I’ll play around with the Limiter stuff. I definitely don’t need to emphasize my “Sss”, as I tend to speak my words very specifically . . . and sometimes with a not-so-appreciated sibilance. :slight_smile: I’m working on actually learning to change the way I pronounce my “S’s” and “C’s”-- thus, hopefully reducing (and eventually) eliminating my sibilance.

I did take a relative small sampling (a book intro) and exported it to MP3 using 3 different bit rates (192, 256, 320). You are right about not being able to tell much difference in the sound, but definitely the file size. Again, will keep that in mind depending on any size file I may, hopefully, eventually have.

Thanks so much again and have a super day!

Sheila Q.

4

There’s an app for that : a de-esser , e.g. Spitfish (free) plugin which works in Audacity on Windows.

You don’t want to eliminate sibilance completely : you’ll sound like you’ve got a lisp, [ or more accurately a lithp].

5

Hi Trebor,

Thanks for your reply. Yes, I’m aware of the D-essers, but I know ACX does NOT approve of using them. My sibilance is not horrible, I just try to always improve on my delivery. :slight_smile: Gives me something new to learn. Plus, as Koz always recommends, I’m striving to get as close to perfect in my raw file, thus minimizing the “effects” I have to use in my mastering. I’m getting ready to post a new question, but first want to search to see if it’s been asked, then I will create it under a new topic. Figure that way, if it is of interest to someone else, it won’t be buried in this thread. Thanks so much again!

Sheila

6

No-one will be able to tell that a de-esser has been applied, unless it’s grossly overdone to the point of giving the speaker a lisp.

7

Hi Trebor,

Hey . . . good to know. I’m beginning to think that “some” of the effects can be used with minimal to unidentifiable extents as long as your raw recording meets the ACX Check test. Maybe I’m overthinking it (smile - well, I AM a woman!! ha! ha!) and getting overly anxious about ACX’s requirements that some of those types of things not be used. Maybe they just mean “to an extreme” to where, like you mentioned, it begins to be noticeable on the final output.

Thanks again for your input . . . and I have seen references to the Spitfish plug-in, so . . . I’ll keep that in mind. Hope you have a great rest of your day/weekend!

Sheila

8

Paul-L’s De-Esser plugin [at the bottom] is very good, see … https://forum.audacityteam.org/t/updated-de-clicker-and-new-de-esser-for-speech/34283/37
But SpitFish is easier to operate … https://www.youtube.com/results?search_query=audacity+Spitfish

9

Thanks so much Trebor. I’ll check both of them out. I practiced some today and I feel it went quite well. Only had 2 or 3 sharp “S’s” (smile). Played around with getting rid of them manually . . . and well, I don’t think I did so bad for just trying the process out. :slight_smile: Again, will check out your recommendations. Have a great rest of your evening and hope you also have a great week!

Sheila