Compression Parameters

Feedback and Discussion Forum
1

first thing i wanted to do was start editing on the newer computer.

so i downloaded the current version of audacity. and when it did not recognize my m4a file, i was able to follow instructions, and download the ffmpeg (sp?) system. and now it opens my file

my computer and audacity version was probably 10 years ago or so

my latest project is going thru all my files and compressing them. this is so i can get a more equalized volume. as most of you probably know, any sort of amplification of the files always looks at all the peaks, and then amplifies only to the point that nothing goes past a certain point.

that is the weakness. if you have a music file with a few peaks in them (which is most common), the rest of the file will be curtailed. what compressing the file did, was to raise the rest of the wave form, such that the peaks were minimal, and then the whole file could be amplified, giving on average a music file that would play twice as loud

with this new version of audacity, however, the waveform only gets smaller, and it does not really end up removing the peaks. because if you amplify it again, the peaks are back to where they were before.

on the old computer, i always used the standard compress settings. and i did the same with this new version. although i think the questions are all different. in any case, i would have expected the standard settings to work basically the same. they are almost 180 degrees out of phase - LOL. since i dont know anything about the settings, i left them alone. as it basically did the job i wanted.

the old way was much better. cuz i could just continue to compress until such a point, where there were no significant peaks. i mean sometimes you can get a monster peak from the artist making one loud note.

any clues on how i can make this new audacity work more like the old one ? on the new one, threshhold = -10, make-up gain = 0, knee width = 5, ratio = 10, look a head = 1, attack = 30, release = 150

2

The old compressor and limiter are still there: Effect → Legacy → Legacy Compressor.

And you might want to experiment the Limiter which is like a fast-kind of compression intended to “push down” the peaks. Limiting tends to have fewer audible side-effects and fewer settings to mess-with.

3

thank you doug,

gosh, that was a great help. i had never used the limiter, before. the old legacy did just as it had done before. the limiter did seem to greatly reduce the peaks. at least at whatever the default settings are, it does all of its work with the first limit. applying it again, does almost nothing. that is not true with the compressor. after compressing it once, i will fairly often hit the repeat compressor once or more than once. and of course, if i think i did it too much, i can always get rid of any of the repeats. is there any place where i can read about the limiter ? cuz i am figuring that a good first step is to use the limiter. from my first use of it, and just looking at the new wave form, i should not have to use the compressor nearly as much. when compared to not using the limiter, and just the compressor. i keep thinking i am pretty good with audacity. i have certainly quite a few tools now in my audacity tool chest !! but i think i may be adding a new limiter tool - LOL.

4

It’s not too complicated. Here is something general about limiters.

Audacity’s limiter is a bit unusual because it has look-ahead (like a negative attack time). That means it can work without distorting the waveform and it doesn’t have to kick-in on a peak that’s only close to the limit. A traditional limiter starts pushing-down and rounding-over the before you hit the limit so you don’t get hard-clipping. (Of course that’s only possible with files. You can’t look-ahead with real time processing :wink: .)

5

thanks doug, i will play around with it, some more. one other question. did audacity make its player much better ? these same waveforms are twice as loud as they used to be. the player never used to be that good, compared to when i loaded the file into itunes. the itunes player used to sound much better than the audacity one. but this one really sounds highly improved. not just in loudness, but in clarity. especially with higher tones. it used to sound more muffled.

6

so the legacy compressor was not working the way it used to. so i logged onto the old computer, and took a look at the settings. everything was the same, but i had “compress based on peaks” checked. when i checked it on the new computer, it now seems to be working the way it used to. i had help from an expert the first time i set this up. i suspect he had me check that. now i know why.

then i checked the limiter on the old computer. the newer computer with the latest version of audacity has all the same parameters. input gain on left and right are both at 0. limit to is set at -3db. hold (ms) is set at 10. and no apply make up gain. on the right side there is a button that you can click for soft limit, hard limit, soft clip, and hard clip. all 4 of them have the above 5 settings exactly the same. and when you click on each of the 4 buttons, the result is somewhat similar. but not the same. they do different things. if there is something i can read about it, i will. if you have some input as to which of the 4 limiter buttons i should use, i would be happy to hear it. the first time i went to the limiter, it displayed “soft limit”. but if you try one of the other 4 buttons, it will default to the last one used. btw, i am nowhere near sophisticated enough to be doing any real time processing. this is all on files. i edit almost every song after i download it from my cd.

7

hi doug, did some research on soft and hard, and limiting and clipping. Soft Limit (recommended) reduces gain progressively as audio approaches the threshold, creating less distortion, while Hard Limit acts as a strict brickwall, creating harsh, abrupt flattening. from this research, it sounds like soft is always better than hard, for me. and limiting is better than clipping. so i think the choice is to use soft limiting.

Soft Clipping (Soft Clip)

  • Action: Waveshapes the signal, rounding off the tops of waveforms when they exceed a threshold.

  • Sound: Introduces subtle harmonic distortion or “saturation,” often described as “analog warmth”.

  • Usage: Best for retaining punch on drums, individual tracks, or bus channels.

  • Soft Limiting (Soft Limit)

    • Action: Attenuates (turns down) volume for peaks that exceed a threshold.

    • Sound: More transparent than clipping, but can make transients sound softer.

    • Usage: Ideal for mastering to control peaks while maintaining a transparent, clean sound.

8

ok doug,

i think i got my system down.

first compress (usually just need one)

then do a soft limit. this brings the file at roughly the same volume

then i have a separate program that puts all the files at the exact same volume

but using the compression/limiting, allows the individual file to have roughly the same value, without a lot of peaks to interfere with the eventual end product !!

9

boy, i just learned 3 more buttons that will be useful to me. fit song to the end of screen. zoom in. zoom out. the zooms will be helpful when i want to edit a passage. you can estimate starting and stopping points much better when you see a longer wave per time.

10

Have you installed the MusxFX free effects? Click on Effect > Get more effects . There’s a Compressor and a Master effect with one knob and lots of presets.

11

thanks. i read a little bit about it. i think that is going beyond my capabilities, and probably my needs. the music i listen to is almost all recorded in the 50s and 60s. some drift into the 70s. but it is almost all hits of one sort or the other. typically referred to as oldies. some of the major things i do are trim the front of it, so it starts out singing. trim the end of it, cuz lots of artists drag on the song, longer than needed. delete sections in a song, usually an instrumental that i dont think is particularly useful in the song. and then especially in older country, some of the crazy stuff they do, especially with the steel guitar. i am not interested in adding anything to a song, like reverb or stuff like that. what i am trying to do is take a melody that i like, and then getting rid of the stuff that i dont like. if it is a stereo recording, that allows me to get rid of stuff in just one channel at a time. that is very helpful in getting rid of certain sounds and instruments, because most of the extraneous stuff is usually just in one channel. so the singing and perhaps the lead guitar are central to both channels. and then the rest of it is in one or the other. lots of these older recordings though were released in just mono. with this recent help, my task has been much easier. i do a quick legacy compressor to bump everything up fairly high (sometimes a repeat), and then a limiter. that very quickly gets me to the point of listening to it, and then doing whatever further editing i may need to do. the zoom out button will be very helpful when i am trying to edit any section in the middle. sometimes i want to delete it, other times i want to just shorten it. with an extended wave form, i can get a much better guess at my stop and start points. i keep thinking i have learned everything i need to know. and then i find another tool to make my job easier - LOL.

12

after i get done with compress and limit, my screen is somewhat a solid green (i am using the high contrast option). it is amazing how all of the waveform comes back into play, when i zoom in 2-3 times. that is really helpful. it is easy to position the slider to within about 1/10 of a second. certainly 2/10. if i need to be more accurate than that, i can still hit the numbers to get probably within 2/100 of a second !!

13

I kinda’ had the idea you were making your own music.

Virtually all commercial music is dynamically compressed already and it’s a normal part of music production. But with modern music often it’s over-done (IMO) making music boring! (The Loudness War)

If you like it more compressed (more constant loudness) that’s fine… It’s YOUR music for YOUR enjoyment but music is SUPPOSED to be dynamic (with loud and quiet parts). It’s one of the things that makes live music better than recorded music.

…Somewhat related - I use ReplayGain to make all of my music library approximately equally-loud. ReplayGain makes ONE adjustment before the track starts so it doesn’t alter the dynamics. (But one potential downside is that it tends to lower the volume of most tracks and if you don’t have enough analog gain your music may be too quiet.)

There is also MP3Gain and WaveGain which make “permanent” loudness adjustments so the files work with any audio player. (I don’t know of an MP4 version.) Again, they will make most music quieter.

Apple has something similar called Sound Check.

Audacity’s Loudness Normalization works similarly, but you have to choose a target loudness yourself, and unlike the above solutions it doesn’t automatically check if you are boosting into clipping. (By default, ReplayGain and related won’t boost into clipping so in some cases where the loudness can be boosted, they just boost as much as possible without clipping.)

14

thanks doug, i cant recall the name of the program that i use. it is only on my music computer. but i had to do a lot of research to find one that will work. first, i store all the original files in wavpack. at the time there was only really wavpack and flac as main choices. audacity will accept either of these 2 as input choices. but i output the files from audacity in flac. there were definite reasons for that, but i dont recall for certain. at first i was only downloading mp3s into itunes. but once i got smart and decided i wanted to keep my play library lossless, my only real choice was to use alac, since i wasnt gonna change itunes. itunes is a wonderful program, with great playlists, etc. it was a million times better than its competition 15 years ago or so. almost none of the loudness gain programs work on alac files, because of apple proprietary reasons. sound check did not work for me. there was something about it that made it unsuitable for my needs. however, we are still missing the elephant in the room. all gain programs work with the same goal. all they do is place the highest peak of each song at the same decibel spot. so songs with a large difference between the highest peak and the average peak will play much more softly than a song whose peaks are more similar. which is exactly what compressing and limiting do. i dont prefer different levels of loudness in a song. i tend to like recorded music better than live music. but not because of the different loudness levels. they each do that. in a recording studio, you can take a bunch of takes. you can overdub. and you can probably have better control of the acoustics, echo, etc. i do wonder if all the overtones are captured as well when recorded. so i could see that as a definite reason for a noticed difference ? there is no musical instrument that sounds pleasing to me. it is only when it is transformed into a melody that i like, that i enjoy the sound. which plays a tremendous role in why i like older music. back then, melody was by far the dominant part of a song. and they had professional song writers all competing to make the best melodies. this current re-introduction to audacity has made my editing much better. especially because of the limiter and the zoom-in. neither of which i knew existed. the only other thing that would be absolutely wonderful is i think a possibility. but i dont know if we have that current capability, or we would still have to invent something to do so. but it would be stupendous if i could push a button, and have a musical instrument disappear from the recording. goodbye steel guitar - LOL. also if i could take a musical instrument and change its tone. i cant stand high notes on strings. and i do not like to hear individual strums of the strings. this happens all the time in old country music. but i am not complaining. there is still tons of songs that i thoroughly enjoy. i think i have over 10,000 songs in my library. and have lots of stuff that i havent even opened yet. thanks a lot to everyone who contributed to the thread. i very much appreciate all the help and comments.

15

hi doug, i looked up how they use the word dynamic in music.

Saying music is “dynamic” means

it features variations in volume, intensity, and power throughout a performance, rather than playing at a constant, flat level. It involves moving between quiet (piano) and loud (forte) sections, creating emotional contrast, tension, and shape within a piece to enhance its expressive quality.

my first comment is the words used to describe the situation. dynamic has a very positive connotation, while flat has a very negative connotation. while someone else may refer to my music listening as very flat and theirs as dynamic, i could very easily say the reverse to them. let me explain.

most of the songs from my listening era are already short. lets just say that the average song length is 2 minutes and 20 seconds. i almost always get rid of the intro, so vocals start right away. the ending of songs is often dragged out longer than necessary. and there is usually a portion in the middle that i will either shorten or delete. and then if it is a pretty good song, i will usually have at least keep the song at 2 different tempos.

now there are some longer ballads that i do very little editing to, so they will remain long. so i dont know what the average length of my songs are. but the median length is really more applicable. and that is almost certainly less than 2 minutes.
so in the same time that this longer piece of music has played loud parts and quiet parts, and changed instruments, i have listened to perhaps 4 or 5 catchy melodies, just one after the other. i find that to be way more dynamic.

melody and tempo are the two aspects of music that my mind seems to like. while i dont prefer too much volume changing, i do enjoy a change in tempo and melody. just not in the same song. but since my songs change way more frequently than the average listener, i get my dynamics in this way. if i am doing a fast workout, then most of my songs will be fairly uptempo. but if i am just listening, the tempo is very random. and with that many songs, i have no idea what is coming next - LOL.

16

Saying music is “dynamic” means

it features variations in volume, intensity, and power throughout a performance, rather than playing at a constant, flat level. It involves moving between quiet (piano) and loud (forte) sections, creating emotional contrast, tension, and shape within a piece to enhance its expressive quality.

Correct! Most people say “dynamic range” but I like the term “dynamic contrast”. There is a online dynamic range database of recordings. The philosophy of that website is that dynamics are good and compression is bad. But like I said, if you like more-compressed sound that’s OK and it’s the way modern music is made so you’re not alone!

Some of the dynamics come from the performance… The music and arrangement. Then the recordings are usually compressed during production, depending on what the producer and music company wants. You can’t “win” the loudness war without compression! :stuck_out_tongue: There is dynamic expansion but it’s almost never used in production,

Compression compresses (reduces) the dynamics. Ironically, when a recording is compressed to make it “louder” or “constantly loud”, people will often describe it a “more dynamic” simply because it’s louder.

But you can’t really define dynamics as a single measurement. There are short-term dynamics like a drum or cymbal hit, or a singer screaming for a moment, etc. Or short quiet parts or a pause or rest in the music for a beat or two. And there are longer-term dynamics like a song that starts-out quiet and ends-loud… And everything in-between.

And there are short peaks in the waveform that are too-short to be heard as loud. Those can be limited without affecting the sound of the dynamics and they are usually taken-out or minimized during mixing and mastering. If you make an MP3 (1) or a vinyl record, some peaks get higher and some lower without affecting the sound of the dynamics. So those new higher-peaks can make it measure better (more dynamic) without affecting the sound of the dynamics.

Classical music has a lot of both. Often when listening at “reasonable levels”, especially in a car with road noise, you can’t hear the quiet parts… So you turn it up and get blasted by the louder parts later. I don’t listen to classical much but I have some Broadway soundtracks that have me constantly reaching for the volume control in the car. When I’m listening “seriously” at home I enjoy the dynamic contrast!

Dynamic range also relates to the range of the equipment or format which is the range from between the noise floor and how loud it can go. The '“numbers” for equipment are far higher than music or regular program material, but that’s needed because you NEVER want to hear background noise, including during silence between songs.

(1) MP3 is file compression. It makes a smaller file. It does not compress the dynamics.

17

i guess it kinda depends on what sort of music one wants to listen to, and what sort of setting one is in. as a rule, i dont prefer a lot of dynamic contrast in a song. but i prefer dynamic contrast in various instruments in a song. while there are a few instrumental hits, these comments are all about vocals.

i will always prefer the vocal to sound the loudest. and usually this is true. whether compressed, limited, or nothing at all, the vocal still stays predominant. in fact, that is what i do like about the compression is that it tends to equalize the overall loudness, it does not change the varying instruments. if the voice was most predominant before, then it will be so, afterwards. it that was not true, then i would not compress the songs.

whenever i do have a chance, i will usually increase the vocals. i love acappella, if it is a good singer. there is often instruments that come in and out for just a second. if i could, i would get rid of all of that. i am not sure why they do that. but for me, the sounds are dissonant.

also, while there are times that i would like to delete an instrument, most of the time you could not delete an instrument without messing up the passage. so what would be better is to be like an organ. i want to change that steel guitar to sound like a fiddle, for example. that way it would still play the same notes, but it would no longer have the whining, and screaming, and absolute torture to my ears !!! why that was ever placed into recorded country music, i have no idea. it originated in hawaii - certainly not a part of traditional country music.

18

hi doug,

the reason i download in flac from audacity is the same reason. audacity can download only in aac, not alac.

and the gain normalizer i use is sound normalizer. at least years back, the only one i could find that worked on alac files.

20

Thanks for the info! I usually export my projects as FLAC in Audacity to keep the best quality. Because Audacity doesn’t export ALAC directly, I convert with Sound Normalizer afterward. It also helps me get consistent volume levels across all my files, and works well with FLAC, MP3, WAV, and AAC without hurting tags or quality.

21

does SN also convert ? never looked into that. i guess cuz i dont have a need.

my whole process, as of now - eac to download cd to wavpack. input wavpack file (or later an m4a file) into adaucity, output a flac file. input that file (really a cd worth folder) into foobar. use foobar to make any changes to text (like song name, track number, etc.). use foobar to convert to mono m4a. use sound normalizer to equalize sound to the standard decibel (is it 86 or 87 ?) (i set that to standard a million years ago, so i forget what the exact number is). then load that into itunes.

one thing i want to look into is that in looking something up in my conversation with doug, i see that foobar is also supposed to be able to do the sound normalizing. foobar is a very simple program, that is absolutely excellent. and since i already use it, it would probably be faster to use it, as opposed to sound normalizer. the convert process in foobar takes almost no time at all. 4-5 seconds to convert 20 songs or so. you are the first person that i have talked to, that is doing almost exactly what i am doing.