[old tech]

Oooh nice. I'll be attending the AES conference in Los Angeles and they'll be having a panel called "Recording & Production: RP5 - Mastering for Vinyl." Maybe I'll learn a thing or two about the mastering process!


...except it's at the same time as the "Listening Tests—Understanding the Basic Concept" pane. : (

Ian Sheppard has a podcast on mastering for vinyl.
 
http://themasteringshow.com/episode-13/

 

[Hi-Fi'er]

Truth about Audio: https://www.youtube.com/watch?v=Z5S_DI99wd8&feature=youtu.be

 

[tkteo]

hello everyone. This is a very interesting discussion thread.
 
earlier on there was mention that form genres such as EDM, there probably would need to have more dynamic range compression that, say, classical.
 
Would the knowledgeable folks here be able to explain to me why certain genres would be better served by relatively more DR compression compared to other genres? And why, when a track or album is too compressed, it would be quite uncomfortable listening to it?
 
Sincerely.

 

[pinnahertz]

  hello everyone. This is a very interesting discussion thread.
 
earlier on there was mention that form genres such as EDM, there probably would need to have more dynamic range compression that, say, classical.
 
Would the knowledgeable folks here be able to explain to me why certain genres would be better served by relatively more DR compression compared to other genres? And why, when a track or album is too compressed, it would be quite uncomfortable listening to it?
 
Sincerely.

Some genres actually depend on dynamics to convey a sense of emotion or drama.  Some classical especially, but you can find dynamic content in pretty much any genre if you go looking. 
 
However, the application dynamics compression is more about what someone in the production chain thinks is necessary for the music's end use.  Dynamic compression makes music sound louder, often quite a bit louder, so if you want your song to be louder you'd master with lots of processing.  Since that's done a lot in some genres, we have the "loudness war", and that means if you don't process excessively your song will be quieter than most others in that genre, and that's viewed as a bad thing.  
 
Compression, when appropriately applied with the real end-use in mind, can be a good thing.  Some classical music is so dynamic that it's impossible to listen to it comfortably in many situations, like in the car or as background music.  Intelligent compression applied somewhere in the playback chain (not the recording) can make a positive difference in audibility above environmental noise.
 
Compression does remove a lot of emotional impact, however.  From a standpoint of impact and listener involvement, it never does any good, in any genre.  
 
When a track is overly processed (there are many forms of dynamics processing included broadly in "compression"), it can become uncomfortable in several ways.  Dynamics processing often creates distortion, particularly intermodulation distortion, making music sound grungy and edgy.  Rapid changes in gain caused by fast attack and recovery of a dynamics processor cause unnatural modulation of surrounding sounds.  An example found in thousands of tracks would be a loud kick drum that causes the rest of the mix to momentarily "duck" each time it occurs.  That would never happen normal acoustic "life", and it sounds very unnatural, a form of distortion.  Highly compressed music sounds very "energetic" and "dense", which is mostly uncomfortable.  When taken to extreme, compression takes the form of peak limiting, and a hard "brick wall" limiter is very similar to clipping in terms of its effect on the envelope and distortion.  A piece devoid of any variation in peak energy, or having a very low peak-to-average ratio, sounds very loud, but at the same time restricted and distorted.  
 
The human hearing system has within it a form of natural compression that occurs when presented with very loud sounds.  When we hear artificially compressed music, we hear the same type of effect, but not at the volume we would expect it to occur.  The result is a perception of the sound that screams "unnatural" and "distorted". 
 
 All of these negatives are why listeners, when given a choice, almost always prefer a less processed version.  The "loudness war" in recorded music is why compression and all dynamics processing is over-applied.  

 

[gregorio]

  [1] However, the application dynamics compression is more about what someone in the production chain thinks is necessary for the music's end use.  [2] Dynamic compression makes music sound louder, often quite a bit louder, so if you want your song to be louder you'd master with lots of processing. 
 
[3] Some classical music is so dynamic that it's impossible to listen to it comfortably in many situations, like in the car or as background music.  Intelligent compression applied somewhere in the playback chain (not the recording) can make a positive difference in audibility above environmental noise.
 
[4] Compression does remove a lot of emotional impact, however.  From a standpoint of impact and listener involvement, it never does any good, in any genre.  
 
[5] Dynamics processing often creates distortion, particularly intermodulation distortion, making music sound grungy and edgy.  [6] Rapid changes in gain caused by fast attack and recovery of a dynamics processor cause unnatural modulation of surrounding sounds.  An example found in thousands of tracks would be a loud kick drum that causes the rest of the mix to momentarily "duck" each time it occurs.  [7] That would never happen normal acoustic "life", and it sounds very unnatural, a form of distortion.  [8] Highly compressed music sounds very "energetic" and "dense", which is mostly uncomfortable.  [9] When taken to extreme, compression takes the form of peak limiting, and a hard "brick wall" limiter is very similar to clipping in terms of its effect on the envelope and distortion.
 
[10] The human hearing system has within it a form of natural compression that occurs when presented with very loud sounds.  When we hear artificially compressed music, we hear the same type of effect, but not at the volume we would expect it to occur.  The result is a perception of the sound that screams "unnatural" and "distorted". 
 
 [11] All of these negatives are why listeners, when given a choice, almost always prefer a less processed version.

 
Interesting viewpoints on what compression is, some of which I agree with, some of which I strongly disagree with or are just incorrect:
 
1. This is incorrect. Compression isn't just applied according to end use. In the mixing phase it's applied for creative, positional, tonal, etc., reasons. IE., It's about the music itself rather than the end use.
2. Actually, compression lowers the level (of the peaks exceeding a user defined threshold). Of course, lowering the level of the peaks provides headroom to add more gain, should one wish to do so.
3. I agree entirely. Except that it's obviously far better applied by those mastering the recording than by some dumb device in the playback chain which has no idea of the musical intentions!
4. This is completely incorrect! In some genres, acoustic genres like most classical music, you would be correct, except in the case of #3 as you mentioned. However, in virtually all popular music genres it's use is essential ingredient in the expected sound. Many genres would be almost a laughable parody without it!
5. True, it can do, if so desired. A point you seem to be entirely missing is what happens in the case of someone wanting to create "grungy and edgy" music? Surely then such compression is the absolute perfect tool? Some vintage compressors are extremely highly prized for their pleasing/desired distortion.
6. Ducking and other effects such as pumping are a musical decision, applied to create impact or otherwise improve the listener experience of certain genres, EDM being a good example of such a genre.
7. That's a rather bizarre statement. Popular genres of music deliberately moved away from the "normal acoustic life" in the 1960's. That's largely what defines the difference between, acoustic genres such as classical/jazz and genres such as rock and most of those which followed. Sure compression is a form of distortion, so is an electric guitar amp! Take the distortion (inc. compression distortion) out of popular genres and make them "natural" and you kill the genre! Ever heard an electric guitar without distortion? The natural sound of an electric guitar is unrecognisable as an electric guitar. The same is largely true of a rock drum kit and to a lesser extent of the live un-compressed/unprocessed vox recorded in the studio.
8. Again, that's a bizarre thing to say. Energetic and dense is exactly the aim with many genres of music. I can appreciate that you personally might not like heavy metal, EDM and most other popular genres and find them "mostly uncomfortable" but far larger numbers of people like and purchase music of those genres than of the acoustic "natural" genres.
9. Brickwall limiting can look similar to clipping distortion when looking at a graphical representation of a waveform but usually sounds completely different.
10. This is incorrect. Sure it's possible to over-apply compression/brickwall limiting so that it sounds unnatural and distorted (and often that's desirable, as mentioned above) but it certainly doesn't have to and the vast majority of the time it's largely transparent.
11. This is blatantly incorrect, in fact the exact opposite is true. Virtually without exception, ALL TV and radio is compressed and brickwall limited, by law in many countries (inc. the USA). You appear to be arguing against yourself as this point contracts your point #3?! The only time your statement has been demonstrated to be true is when all the following conditions are met: In critical listening situations, with good quality playback equipment being used in a decently quiet listening environment and when the audio being listened too has been over-processed in the first place.
 

  [1] earlier on there was mention that form genres such as EDM, there probably would need to have more dynamic range compression that, say, classical. Would the knowledgeable folks here be able to explain to me why certain genres would be better served by relatively more DR compression compared to other genres? [2] And why, when a track or album is too compressed, it would be quite uncomfortable listening to it?

 
1. EDM as an example, uses very different sound sources than acoustic genres such as most classical music. The bass drum sounds used and indeed most of the sounds do not exist in the natural world, they are manufactured and only exist in the studio (hence the term EDM!), and compression is an essential tool used in the manufacture of those sounds. Commonly, the only relatively natural sound in an EDM mix is the lead vocal but even that has to be significantly compressed (and processed in other ways) in order to be successfully integrated into the mix. Additionally, as the name also implies, EDM is primarily designed for use in extremely noisy listening environments (clubs, arenas, etc.) and therefore requires a small dynamic range. Although there are some genre exceptions, a small dynamic range is not generally perceived as very exciting, so heavy compression is often used as it reduces the energy level differences while still maintaining some of the key indicators the brain uses to differentiate dynamic range. To take full advantage of this, the music has to be composed and orchestrated with this in mind. Obviously the vast majority of classical music isn't, it's designed for acoustic instrument performance, whereas EDM is specifically created/designed for this type of processing.
 
2. Generally it wouldn't necessarily be uncomfortable as such, you could just turn your amp down if it were too loud. Most of the energy in music mixes is in the lower frequencies. When applying broadband compression it's therefore these lower frequencies which get more compressed (lowered/attenuated), resulting in the mix having relatively more mid and high frequency content, which makes it sound louder and more present but also potentially harsh (and therefore potentially uncomfortable).
 
G

 

[pinnahertz]

   
Interesting viewpoints on what compression is, some of which I agree with, some of which I strongly disagree with or are just incorrect:
 
1. This is incorrect. Compression isn't just applied according to end use. In the mixing phase it's applied for creative, positional, tonal, etc., reasons. IE., It's about the music itself rather than the end use.
2. Actually, compression lowers the level (of the peaks exceeding a user defined threshold). Of course, lowering the level of the peaks provides headroom to add more gain, should one wish to do so.
3. I agree entirely. Except that it's obviously far better applied by those mastering the recording than by some dumb device in the playback chain which has no idea of the musical intentions!
4. This is completely incorrect! In some genres, acoustic genres like most classical music, you would be correct, except in the case of #3 as you mentioned. However, in virtually all popular music genres it's use is essential ingredient in the expected sound. Many genres would be almost a laughable parody without it!
5. True, it can do, if so desired. A point you seem to be entirely missing is what happens in the case of someone wanting to create "grungy and edgy" music? Surely then such compression is the absolute perfect tool? Some vintage compressors are extremely highly prized for their pleasing/desired distortion.
6. Ducking and other effects such as pumping are a musical decision, applied to create impact or otherwise improve the listener experience of certain genres, EDM being a good example of such a genre.
7. That's a rather bizarre statement. Popular genres of music deliberately moved away from the "normal acoustic life" in the 1960's. That's largely what defines the difference between, acoustic genres such as classical/jazz and genres such as rock and most of those which followed. Sure compression is a form of distortion, so is an electric guitar amp! Take the distortion (inc. compression distortion) out of popular genres and make them "natural" and you kill the genre! Ever heard an electric guitar without distortion? The natural sound of an electric guitar is unrecognisable as an electric guitar. The same is largely true of a rock drum kit and to a lesser extent of the live un-compressed/unprocessed vox recorded in the studio.
8. Again, that's a bizarre thing to say. Energetic and dense is exactly the aim with many genres of music. I can appreciate that you personally might not like heavy metal, EDM and most other popular genres and find them "mostly uncomfortable" but far larger numbers of people like and purchase music of those genres than of the acoustic "natural" genres.
9. Brickwall limiting can look similar to clipping distortion when looking at a graphical representation of a waveform but usually sounds completely different.
10. This is incorrect. Sure it's possible to over-apply compression/brickwall limiting so that it sounds unnatural and distorted (and often that's desirable, as mentioned above) but it certainly doesn't have to and the vast majority of the time it's largely transparent.
11. This is blatantly incorrect, in fact the exact opposite is true. Virtually without exception, ALL TV and radio is compressed and brickwall limited, by law in many countries (inc. the USA). You appear to be arguing against yourself as this point contracts your point #3?! The only time your statement has been demonstrated to be true is when all the following conditions are met: In critical listening situations, with good quality playback equipment being used in a decently quiet listening environment and when the audio being listened too has been over-processed in the first place.
 
 
1. EDM as an example, uses very different sound sources than acoustic genres such as most classical music. The bass drum sounds used and indeed most of the sounds do not exist in the natural world, they are manufactured and only exist in the studio (hence the term EDM!), and compression is an essential tool used in the manufacture of those sounds. Commonly, the only relatively natural sound in an EDM mix is the lead vocal but even that has to be significantly compressed (and processed in other ways) in order to be successfully integrated into the mix. Additionally, as the name also implies, EDM is primarily designed for use in extremely noisy listening environments (clubs, arenas, etc.) and therefore requires a small dynamic range. Although there are some genre exceptions, a small dynamic range is not generally perceived as very exciting, so heavy compression is often used as it reduces the energy level differences while still maintaining some of the key indicators the brain uses to differentiate dynamic range. To take full advantage of this, the music has to be composed and orchestrated with this in mind. Obviously the vast majority of classical music isn't, it's designed for acoustic instrument performance, whereas EDM is specifically created/designed for this type of processing.
 
2. Generally it wouldn't necessarily be uncomfortable as such, you could just turn your amp down if it were too loud. Most of the energy in music mixes is in the lower frequencies. When applying broadband compression it's therefore these lower frequencies which get more compressed (lowered/attenuated), resulting in the mix having relatively more mid and high frequency content, which makes it sound louder and more present but also potentially harsh (and therefore potentially uncomfortable).
 
 
 
G

"Interesting viewpoints on what compression is, some of which I agree with, some of which I strongly disagree with or are just incorrect:
 
1. This is incorrect. Compression isn't just applied according to end use. In the mixing phase it's applied for creative, positional, tonal, etc., reasons. IE., It's about the music itself rather than the end use."
Actually, not incorrect, but we are both correct. These are not opposing concepts, both occur.  But compression on a per-track or per-source basis is not responsible for the overall loudness of a piece.  Yes, per-track compression is about the music.  Mastering loudness compression is more about the impression the piece has up against other competitors.
 
 2. Actually, compression lowers the level (of the peaks exceeding a user defined threshold). Of course, lowering the level of the peaks provides headroom to add more gain, should one wish to do so.
 
I never said otherwise, but the type of compression you're referring to is specifically peak limiting, fast attack and recovery, infinite slope or close to it.  It controls peak maximum, as you say, but that permits the total volume to be pushed up to close to 0dBFS, as much as 12dB or so closer than it would without peak limiting.  
 
3. I agree entirely. Except that it's obviously far better applied by those mastering the recording than by some dumb device in the playback chain which has no idea of the musical intentions!
 
Not so obviously...some mastering compression may be beneficial, but its not tuned to specific listener environments.  The needs for compression in forground vs background applications are vastly different.  Background systems I design and install all include some form of general music compression so levels can be consistent.  The average level in background music is very touchy, a little variation is percieved as wrong.  Very different demands for compression.  However, if it's over-done in mastering, there's no way to un-do it for any application.
 
4. This is completely incorrect! In some genres, acoustic genres like most classical music, you would be correct, except in the case of #3 as you mentioned. However, in virtually all popular music genres it's use is essential ingredient in the expected sound. Many genres would be almost a laughable parody without it!
 
This is a matter of opinion, so "incorrect" would be a bit of an over-statement.  The drive to heavily processed popular music is well documented in the loudness war.  It's not about listener preference or artistic expression at all.  There is so much about the loudness war published, I don't feel compelled to detail it here.
 
5. True, it can do, if so desired. A point you seem to be entirely missing is what happens in the case of someone wanting to create "grungy and edgy" music? Surely then such compression is the absolute perfect tool? Some vintage compressors are extremely highly prized for their pleasing/desired distortion.
 
IMHO, the desire for moderate grunge may be artistic expression, over use of compression is simply misguided.  I'm intimately familiar with what the popular vintage compressors do having used them personally before they were "vintage".  In many cases they aren't responsible for the objectionable effects I'm referring to.
 
6. Ducking and other effects such as pumping are a musical decision, applied to create impact or otherwise improve the listener experience of certain genres, EDM being a good example of such a genre.
 
The examples I hear I find objectionable.  I actually like and listen to some of the newest music, but have even tried to remove some of that effect with very limited success (pardon...). Again, this is a matter of opinion, though I firmly believe that most listeners, if given the choice of ducking or not in a level matched test would prefer the not.
 
7. That's a rather bizarre statement. Popular genres of music deliberately moved away from the "normal acoustic life" in the 1960's. That's largely what defines the difference between, acoustic genres such as classical/jazz and genres such as rock and most of those which followed. Sure compression is a form of distortion, so is an electric guitar amp! Take the distortion (inc. compression distortion) out of popular genres and make them "natural" and you kill the genre!
 
Not at all.  Acoustic life is all around us.  It's the reference, like it or not.  Genres have adopted certain styles, that's true, but I don't think severe loudness processing defines the style.  It could be backed of 6dB or more without affecting the style or genre while vastly improving the clarity and impact of the piece.
 
Ever heard an electric guitar without distortion? The natural sound of an electric guitar is unrecognisable as an electric guitar. The same is largely true of a rock drum kit and to a lesser extent of the live un-compressed/unprocessed vox recorded in the studio.
 
You're confusing compression as applied as an (abused) loudness tool with processes that define an instrumental sound.  The are not at all the same nor comparable.
 
8. Again, that's a bizarre thing to say. Energetic and dense is exactly the aim with many genres of music. I can appreciate that you personally might not like heavy metal, EDM and most other popular genres and find them "mostly uncomfortable" but far larger numbers of people like and purchase music of those genres than of the acoustic "natural" genres.
 
How would you or anyone else know that?  Those that buy music have no choice of the final processing.  They buy it for what it is, processing is beside the point.
 
In radio we've experimented for many decades with different types and amounts of processing, and yes, there's been a loudness war on radio for 50 years.  What the evidence of audience ratings has shown is audience preference is for the content, and processing and loudness doesn't matter.  That's why an AM talk station in a major market can beat all of the FM music stations, processed or not.  The AM signal has issues, it's lower quality, and this particular station (yes, it's a real example) was actually one of the least processed, less loud AM stations on the dial.  They still won the ratings.  You wouldn't know what music buyers would choose, because they don't have any choice in the matter.  However, there is research to show that less compression is generally preferred.  That research has prompted the development of highly sophisticated broadcast processors that do what they do with less compression artifact.  
 
If you combine that with the admittedly grass-roots movement against the loudness war in recorded music, to me, that confirms that less compression is the preferred version.
 
9. Brickwall limiting can look similar to clipping distortion when looking at a graphical representation of a waveform but usually sounds completely different.
 
Yes, but the point is moot.  Both are used to varying degrees in recorded music.  Both add distortion of several kinds.
 
10. This is incorrect. Sure it's possible to over-apply compression/brickwall limiting so that it sounds unnatural and distorted (and often that's desirable, as mentioned above) but it certainly doesn't have to and the vast majority of the time it's largely transparent.
 
Respectfully have to disagree that the majority of the time it's transparent.  The popular music I hear is all way over processed, easily heard.
 
11. This is blatantly incorrect, in fact the exact opposite is true. Virtually without exception, ALL TV and radio is compressed and brickwall limited, by law in many countries (inc. the USA).
 
What makes you think that the choices made by broadcasters reflect listener preference?  I've been a broadcast engineer for 45 years, I can tell you all about broadcast processing.  I've designed broadcast processors.  Your concept of a brick-wall limiter is also blatantly incorrect.  Pre-emphasis and assymetrical modulation limits (AM) preclude the typical brick wall limiter from being effective.  True, at the final transmitter input there must be absolute peak control, but with pre-emphasis and de-emphasis involved, it doesn't come out that way at the receiver audio outputs.
 
A broadcast processor includes absolute peak protection as the final process in a complex chain of other processes.  This is generally some form of clipping, but must act on the final pre-emphasised and bandwidth-limited signal.  To over-simplify a bit, that clipper is preceded by a fast peak limiter and in the case of analog AM and FM, a high frequency limiter that response to the specific pre-emphasis curve in use.  That's preceded typically by a multi-band limiter, at minimum two band, at maximum 8, but typically 4 or 5.  Each band is handled by a complex gain control algorithm that includes variable attack and release, several different composite time constants, bandwidth controls, re-mix controls, threshold controls and release-time gating.  That's typically preceded by a broadband compressor, again with variable attack and release, but fairly slow, and with release time gating.  All of it is so incredibly adjustable and variable that no two stations are likely to be processed identically.  TV and radio are processed differently too, as is each radio format.  The goals fundamentally are to sound as loud as the competition without sounding bad, and that translates to sounding as loud as possible without sounding too bad, all the while keeping within legal modulation limits.  But "brick wall" doesn't even scratch the surface, nor does it describe the same sort of thing as used in music mastering.
 
Yet in the development of audio processors, has been plenty of research that proves that listeners, if given the choice, prefer less processing and compression.  Again, that's why over the last 40 years broadcast processing has become less damaging for the same comparative loudness.  The problem is that when the tools got better, there was also more opportunity to abuse them.
 
You appear to be arguing against yourself as this point contracts your point #3?! The only time your statement has been demonstrated to be true is when all the following conditions are met: In critical listening situations, with good quality playback equipment being used in a decently quiet listening environment and when the audio being listened too has been over-processed in the first place.
 
Actually, I'm not arguing against myself at all.  My statements in point #3 are specific to an end-use situation.  My statements are not based on the conditions you outline, in fact, the conditions that make classical music difficult or impossible to listen to most often are found in a moving car with high background noise and medium quality equipment.  Those are conditions where processing is necessary and beneficial.  
 
Actually, my point was that if the end use is well understood, then the proper processing can be applied.  To over-process generally creates a signal that has no opportunity for ever being played with any better dynamics, and is actually mastered for the lowest common denominator.  That, I feel, is a mistake.

 

[sikki-six]

Compression does remove a lot of emotional impact, however.  From a standpoint of impact and listener involvement, it never does any good, in any genre.   

 
Actually, compression does the opposite in many cases. It often improves the emotional impact. Many (at least popular) styles of music need some extra punch and a sense of presence from the performance. Especially drums, guitars and most of all bass benefits from compression. The thing I know about production, rock and metal mostly: Bass guitar is all over the place in the lows, without compression. These styles need compression - always - to have their intended sound. There aren't professionally recorded drum-sets or basses (jazz and classical not included) without some compression.
 
Compression is a necessary tool for almost every recording.
 
I'm not saying it's not used wrong, it is (esp on drums, they sound like ass with too much compression). Brickwall-limiting is even worse, though.

 

[pinnahertz]

   
Actually, compression does the opposite in many cases. It often improves the emotional impact. Many (at least popular) styles of music need some extra punch and a sense of presence from the performance. Especially drums, guitars and most of all bass benefits from compression. The thing I know about production, rock and metal mostly: Bass guitar is all over the place in the lows, without compression. These styles need compression - always - to have their intended sound. There aren't professionally recorded drum-sets or basses (jazz and classical not included) without some compression.
 
Compression is a necessary tool for almost every recording.
 
I'm not saying it's not used wrong, it is (esp on drums, they sound like ass with too much compression). Brickwall-limiting is even worse, though.

I can agree with this if we're talking about a per-source, per-instrument, or per-track basis. In that case compression is absolutely necessary, and fundamental to getting the performance with expression into the mix.  However, when aggressively applied to the final mix compression squashes emotion out because the total dynamics have been reduced or removed.  
 
Let's not confuse the discussion with the use of processing on sources and tracks within a mix.  That's a very different thing than overall loudness processing in mastering.

 

[sikki-six]

I find compression to be a necessary tool also in the mastering stage. When used in moderation, it "glues" everything together. When used wrong it screws up drums especially, makes them sound like a wimpy ass. (Limiting does that even more, it's a separate discussion, I guess.) That good kind of compression makes the music more punchy and exciting.
 
There is such a thing as "too much dynamics" too, it happens. Things get too loud sometimes, without compression. You need good levels all the way, without the need to cut or boost your hifi's volume mid-way of a piece. It's all about balance. Compression helps lots.

 

[VNandor]

Looking at the popularity of EDM, compression and little dynamic range does not neccesarily sound bad. People seem to like it in some cases.

 

[pinnahertz]

  Looking at the popularity of EDM, compression and little dynamic range does not neccesarily sound bad. People seem to like it in some cases.

Got an statistics or surveys?

 

[pinnahertz]

  I find compression to be a necessary tool also in the mastering stage. When used in moderation, it "glues" everything together. When used wrong it screws up drums especially, makes them sound like a wimpy ass. (Limiting does that even more, it's a separate discussion, I guess.) That good kind of compression makes the music more punchy and exciting.
 
There is such a thing as "too much dynamics" too, it happens. Things get too loud sometimes, without compression. You need good levels all the way, without the need to cut or boost your hifi's volume mid-way of a piece. It's all about balance. Compression helps lots.

In this case compression is the automated cheat to a good mix.  Manual or fader automation would be another option. 
 
We're talking here like it's either compression or not, there are infinite variations with continuously variable parameters. It's not like its on or off.

 

[gregorio]

  "Interesting viewpoints on what compression is, some of which I agree with, some of which I strongly disagree with or are just incorrect:
 
1. This is incorrect. Compression isn't just applied according to end use. In the mixing phase it's applied for creative, positional, tonal, etc., reasons. IE., It's about the music itself rather than the end use."
Actually, not incorrect, but we are both correct. These are not opposing concepts, both occur.  But compression on a per-track or per-source basis is not responsible for the overall loudness of a piece.  Yes, per-track compression is about the music.  Mastering loudness compression is more about the impression the piece has up against other competitors.
 
 2. Actually, compression lowers the level (of the peaks exceeding a user defined threshold). Of course, lowering the level of the peaks provides headroom to add more gain, should one wish to do so.
 
I never said otherwise, but the type of compression you're referring to is specifically peak limiting, fast attack and recovery, infinite slope or close to it.  It controls peak maximum, as you say, but that permits the total volume to be pushed up to close to 0dBFS, as much as 12dB or so closer than it would without peak limiting.  
 
3. I agree entirely. Except that it's obviously far better applied by those mastering the recording than by some dumb device in the playback chain which has no idea of the musical intentions!
 
Not so obviously...some mastering compression may be beneficial, but its not tuned to specific listener environments.  The needs for compression in forground vs background applications are vastly different.  Background systems I design and install all include some form of general music compression so levels can be consistent.  The average level in background music is very touchy, a little variation is percieved as wrong.  Very different demands for compression.  However, if it's over-done in mastering, there's no way to un-do it for any application.
 
4. This is completely incorrect! In some genres, acoustic genres like most classical music, you would be correct, except in the case of #3 as you mentioned. However, in virtually all popular music genres it's use is essential ingredient in the expected sound. Many genres would be almost a laughable parody without it!
 
This is a matter of opinion, so "incorrect" would be a bit of an over-statement.  The drive to heavily processed popular music is well documented in the loudness war.  It's not about listener preference or artistic expression at all.  There is so much about the loudness war published, I don't feel compelled to detail it here.
 
5. True, it can do, if so desired. A point you seem to be entirely missing is what happens in the case of someone wanting to create "grungy and edgy" music? Surely then such compression is the absolute perfect tool? Some vintage compressors are extremely highly prized for their pleasing/desired distortion.
 
IMHO, the desire for moderate grunge may be artistic expression, over use of compression is simply misguided.  I'm intimately familiar with what the popular vintage compressors do having used them personally before they were "vintage".  In many cases they aren't responsible for the objectionable effects I'm referring to.
 
6. Ducking and other effects such as pumping are a musical decision, applied to create impact or otherwise improve the listener experience of certain genres, EDM being a good example of such a genre.
 
The examples I hear I find objectionable.  I actually like and listen to some of the newest music, but have even tried to remove some of that effect with very limited success (pardon...). Again, this is a matter of opinion, though I firmly believe that most listeners, if given the choice of ducking or not in a level matched test would prefer the not.
 
7. That's a rather bizarre statement. Popular genres of music deliberately moved away from the "normal acoustic life" in the 1960's. That's largely what defines the difference between, acoustic genres such as classical/jazz and genres such as rock and most of those which followed. Sure compression is a form of distortion, so is an electric guitar amp! Take the distortion (inc. compression distortion) out of popular genres and make them "natural" and you kill the genre!
 
Not at all.  Acoustic life is all around us.  It's the reference, like it or not.  Genres have adopted certain styles, that's true, but I don't think severe loudness processing defines the style.  It could be backed of 6dB or more without affecting the style or genre while vastly improving the clarity and impact of the piece.
 
Ever heard an electric guitar without distortion? The natural sound of an electric guitar is unrecognisable as an electric guitar. The same is largely true of a rock drum kit and to a lesser extent of the live un-compressed/unprocessed vox recorded in the studio.
 
You're confusing compression as applied as an (abused) loudness tool with processes that define an instrumental sound.  The are not at all the same nor comparable.
 
8. Again, that's a bizarre thing to say. Energetic and dense is exactly the aim with many genres of music. I can appreciate that you personally might not like heavy metal, EDM and most other popular genres and find them "mostly uncomfortable" but far larger numbers of people like and purchase music of those genres than of the acoustic "natural" genres.
 
How would you or anyone else know that?  Those that buy music have no choice of the final processing.  They buy it for what it is, processing is beside the point.
 
In radio we've experimented for many decades with different types and amounts of processing, and yes, there's been a loudness war on radio for 50 years.  What the evidence of audience ratings has shown is audience preference is for the content, and processing and loudness doesn't matter.  That's why an AM talk station in a major market can beat all of the FM music stations, processed or not.  The AM signal has issues, it's lower quality, and this particular station (yes, it's a real example) was actually one of the least processed, less loud AM stations on the dial.  They still won the ratings.  You wouldn't know what music buyers would choose, because they don't have any choice in the matter.  However, there is research to show that less compression is generally preferred.  That research has prompted the development of highly sophisticated broadcast processors that do what they do with less compression artifact.  
 
If you combine that with the admittedly grass-roots movement against the loudness war in recorded music, to me, that confirms that less compression is the preferred version.
 
9. Brickwall limiting can look similar to clipping distortion when looking at a graphical representation of a waveform but usually sounds completely different.
 
Yes, but the point is moot.  Both are used to varying degrees in recorded music.  Both add distortion of several kinds.
 
10. This is incorrect. Sure it's possible to over-apply compression/brickwall limiting so that it sounds unnatural and distorted (and often that's desirable, as mentioned above) but it certainly doesn't have to and the vast majority of the time it's largely transparent.
 
Respectfully have to disagree that the majority of the time it's transparent.  The popular music I hear is all way over processed, easily heard.
 
11. This is blatantly incorrect, in fact the exact opposite is true. Virtually without exception, ALL TV and radio is compressed and brickwall limited, by law in many countries (inc. the USA).
 
What makes you think that the choices made by broadcasters reflect listener preference?  I've been a broadcast engineer for 45 years, I can tell you all about broadcast processing.  I've designed broadcast processors.  Your concept of a brick-wall limiter is also blatantly incorrect.  Pre-emphasis and assymetrical modulation limits (AM) preclude the typical brick wall limiter from being effective.  True, at the final transmitter input there must be absolute peak control, but with pre-emphasis and de-emphasis involved, it doesn't come out that way at the receiver audio outputs.
 
A broadcast processor includes absolute peak protection as the final process in a complex chain of other processes.  This is generally some form of clipping, but must act on the final pre-emphasised and bandwidth-limited signal.  To over-simplify a bit, that clipper is preceded by a fast peak limiter and in the case of analog AM and FM, a high frequency limiter that response to the specific pre-emphasis curve in use.  That's preceded typically by a multi-band limiter, at minimum two band, at maximum 8, but typically 4 or 5.  Each band is handled by a complex gain control algorithm that includes variable attack and release, several different composite time constants, bandwidth controls, re-mix controls, threshold controls and release-time gating.  That's typically preceded by a broadband compressor, again with variable attack and release, but fairly slow, and with release time gating.  All of it is so incredibly adjustable and variable that no two stations are likely to be processed identically.  TV and radio are processed differently too, as is each radio format.  The goals fundamentally are to sound as loud as the competition without sounding bad, and that translates to sounding as loud as possible without sounding too bad, all the while keeping within legal modulation limits.  But "brick wall" doesn't even scratch the surface, nor does it describe the same sort of thing as used in music mastering.
 
Yet in the development of audio processors, has been plenty of research that proves that listeners, if given the choice, prefer less processing and compression.  Again, that's why over the last 40 years broadcast processing has become less damaging for the same comparative loudness.  The problem is that when the tools got better, there was also more opportunity to abuse them.
 
You appear to be arguing against yourself as this point contracts your point #3?! The only time your statement has been demonstrated to be true is when all the following conditions are met: In critical listening situations, with good quality playback equipment being used in a decently quiet listening environment and when the audio being listened too has been over-processed in the first place.
 
Actually, I'm not arguing against myself at all.  My statements in point #3 are specific to an end-use situation.  My statements are not based on the conditions you outline, in fact, the conditions that make classical music difficult or impossible to listen to most often are found in a moving car with high background noise and medium quality equipment.  Those are conditions where processing is necessary and beneficial.  
 
Actually, my point was that if the end use is well understood, then the proper processing can be applied.  To over-process generally creates a signal that has no opportunity for ever being played with any better dynamics, and is actually mastered for the lowest common denominator.  That, I feel, is a mistake.

 
We're maybe not so far apart as I first thought from your previous post but there are still a few points:
 
1. I take it you would agree that on the basics that compression affects the level of a signal? If you do, then applying compression on an individual track changes the level of that track and increases it's perceived volume by the amount of make-up gain. In order to create a balance with this track we have to increase the other tracks accordingly. So I'm going to have to disagree, applying compression on a per track basis does indeed affect the overall loudness of the piece. In some pieces/genres compression is applied to almost every individual track and is largely responsible for the overall loudness.
 
2. No, I'm not talking about any specific type of compression but all compression. On all compressors a threshold level is set, any part of the signal/waveform which peaks above that threshold level is compressed, attenuated by the amount defined by the ratio setting or limited to that threshold in the case of a brickwall limiter.
 
4, 5, 6 & 7. Respectfully I have to say that you are the one "confused". You are confusing compression with extreme amounts of make-up gain applied on the master buss with a brickwall limiter, which is just one specific type and use of compression. That is the tool/process of the loudness war, a phenomena I was discussing 25 years ago with other audio engineers. However, even without the excessive master-buss brickwall limiting which defines the loudness war, there is still a considerable amount of compression applied to all popular music genres, even quite extreme amounts in some genres like EDM. Ducking for example, is virtually a requisite for some of the harder-core EDM and not using very significant amounts of compression on the electric guitar and vocals of say thrash metal would most probably change it's genre, it would no longer be thrash but some other sub genre.
 
8. And this answer maybe explains some of your other answers. We know that "far larger numbers of people like and purchase music of those genres than the acoustic genres" because for many decades we've had the sales figures! Of course consumers have a choice in the final processing, they can choose to buy another piece of music with different processing. Yes, they buy it for what it is BUT, processing is not besides the point, processing is what defines "what it is" in the first place! This fact, which trumps any number of studies, is not only what has caused the loudness war but the very evolution of the many popular music genres themselves!
 
10. "Largely transparent" was perhaps the wrong term, unobjectionable would perhaps be better. And, I wasn't talking of just music recordings but of all commercial audio.
 
11. What makes you think that the choices made by broadcasters reflect listener preference? Nothing at all, in fact quite the opposite, which is why the choice was taken away from the broadcasters and is now defined by federal law (in some countries). For this reason, the "goals" you talk about no longer exist in TV and haven't for a few years. "Your concept of a brick-wall limiter is also blatantly incorrect" I've used one almost every working day for more than 2 decades, so I don't think so, not to mention that I'd loose my job if my printmasters were not precisely limited to -1dBTP (or -2dBTP for my North American clients).
 
The studies I'm aware of, which support a preference for a greater dynamic range, are all contingent on the conditions I detailed previously. In other conditions, in a car as you mentioned and these days also in a bus, train, plane, cycling, when walking the dog, doing the homework, jogging and numerous common other conditions, this would obviously not be the case.
 
Lastly, this thread isn't specifically about the extreme overuse of compression which characterises the loudness war and tkteo was not specifically asking about the loudness war but just the use of compression. The amount of compression, both per track and master-buss compression, which is not only acceptable but actually indicative of EDM would generally be obscene if applied to a genre such as classical music. I'm certainly no supporter of the loudness war, quite the contrary, I've been discussing and complaining about it since the beginning of the '90's but I'm also concerned by the complete opposite, extremist view I've seen quite often on this board, essentially that compression is an evil which should be eradicated. This view is even more potentially damaging than the loudness war itself, it seems to be based on ignorance. The loudness war is not about compression, it's about the abuse of compression!
 
G

 

[Muinarc]

Yes we need some compression in the process. From what I've been reading in here it's the customer's fault and not the engineers. Can the engineers do anything to educate their customers at all to get back to "responsible" compression practices or is that a lost cause?

 

[pinnahertz]

   
We're maybe not so far apart as I first thought from your previous post but there are still a few points:
 
1. I take it you would agree that on the basics that compression affects the level of a signal? If you do, then applying compression on an individual track changes the level of that track and increases it's perceived volume by the amount of make-up gain. In order to create a balance with this track we have to increase the other tracks accordingly. So I'm going to have to disagree, applying compression on a per track basis does indeed affect the overall loudness of the piece. In some pieces/genres compression is applied to almost every individual track and is largely responsible for the overall loudness.
 
2. No, I'm not talking about any specific type of compression but all compression. On all compressors a threshold level is set, any part of the signal/waveform which peaks above that threshold level is compressed, attenuated by the amount defined by the ratio setting or limited to that threshold in the case of a brickwall limiter.
 
4, 5, 6 & 7. Respectfully I have to say that you are the one "confused". You are confusing compression with extreme amounts of make-up gain applied on the master buss with a brickwall limiter, which is just one specific type and use of compression. That is the tool/process of the loudness war, a phenomena I was discussing 25 years ago with other audio engineers. However, even without the excessive master-buss brickwall limiting which defines the loudness war, there is still a considerable amount of compression applied to all popular music genres, even quite extreme amounts in some genres like EDM. Ducking for example, is virtually a requisite for some of the harder-core EDM and not using very significant amounts of compression on the electric guitar and vocals of say thrash metal would most probably change it's genre, it would no longer be thrash but some other sub genre.
 
8. And this answer maybe explains some of your other answers. We know that "far larger numbers of people like and purchase music of those genres than the acoustic genres" because for many decades we've had the sales figures! Of course consumers have a choice in the final processing, they can choose to buy another piece of music with different processing. Yes, they buy it for what it is BUT, processing is not besides the point, processing is what defines "what it is" in the first place! This fact, which trumps any number of studies, is not only what has caused the loudness war but the very evolution of the many popular music genres themselves!
 
10. "Largely transparent" was perhaps the wrong term, unobjectionable would perhaps be better. And, I wasn't talking of just music recordings but of all commercial audio.
 
11. What makes you think that the choices made by broadcasters reflect listener preference? Nothing at all, in fact quite the opposite, which is why the choice was taken away from the broadcasters and is now defined by federal law (in some countries). For this reason, the "goals" you talk about no longer exist in TV and haven't for a few years. "Your concept of a brick-wall limiter is also blatantly incorrect" I've used one almost every working day for more than 2 decades, so I don't think so, not to mention that I'd loose my job if my printmasters were not precisely limited to -1dBTP (or -2dBTP for my North American clients).
 
The studies I'm aware of, which support a preference for a greater dynamic range, are all contingent on the conditions I detailed previously. In other conditions, in a car as you mentioned and these days also in a bus, train, plane, cycling, when walking the dog, doing the homework, jogging and numerous common other conditions, this would obviously not be the case.
 
Lastly, this thread isn't specifically about the extreme overuse of compression which characterises the loudness war and tkteo was not specifically asking about the loudness war but just the use of compression. The amount of compression, both per track and master-buss compression, which is not only acceptable but actually indicative of EDM would generally be obscene if applied to a genre such as classical music. I'm certainly no supporter of the loudness war, quite the contrary, I've been discussing and complaining about it since the beginning of the '90's but I'm also concerned by the complete opposite, extremist view I've seen quite often on this board, essentially that compression is an evil which should be eradicated. This view is even more potentially damaging than the loudness war itself, it seems to be based on ignorance. The loudness war is not about compression, it's about the abuse of compression!
 
G

1. I take it you would agree that on the basics that compression affects the level of a signal? If you do, then applying compression on an individual track changes the level of that track and increases it's perceived volume by the amount of make-up gain. In order to create a balance with this track we have to increase the other tracks accordingly. So I'm going to have to disagree, applying compression on a per track basis does indeed affect the overall loudness of the piece. In some pieces/genres compression is applied to almost every individual track and is largely responsible for the overall loudness.
 
Except that you don't have to increase all of the other tracks, you can lower the compressed track to re-establish balance.  Per-track compression need not affect overall loudness.  
 
2. No, I'm not talking about any specific type of compression but all compression. On all compressors a threshold level is set, any part of the signal/waveform which peaks above that threshold level is compressed, attenuated by the amount defined by the ratio setting or limited to that threshold in the case of a brickwall limiter.
 
That's an overly simplistic view of compressors and limiters.  A compressor may also have separate attack and release time adjustments, and the transition around the threshold can be made a hard corner or soft entrance into the compression ratio curve.  Even more importantly a compressor may have one of three different types of envelope detectors, peak, average or RMS.  An RMS, mild slope, slower attack and very slow release compressor is an entirely different animal than a peak limiter function, which is defined by using a peak envelope detector, very fast attack (sometimes actually negative, in the case of "look ahead" limiters), fairly fast release (required, or excessive overall gain changes result from very short, nearly inaudible peaks), and an infinite compression slope.  It's a special case, and fairly singular relative to just about any other form of compressor. 
 
4, 5, 6 & 7. Respectfully I have to say that you are the one "confused". You are confusing compression with extreme amounts of make-up gain applied on the master buss with a brickwall limiter, which is just one specific type and use of compression. That is the tool/process of the loudness war, a phenomena I was discussing 25 years ago with other audio engineers.
 
 
No, I don't think so. I'm well aware of what make-up gain is, how and why it's applied.  And a sure know what a brick wall limiter is (though, again, you've grossly generalized, there are many different way to handle a firm, maximum peak limiter). 
 
However, even without the excessive master-buss brickwall limiting which defines the loudness war, there is still a considerable amount of compression applied to all popular music genres, even quite extreme amounts in some genres like EDM.
 
No argument there.
 
Ducking for example, is virtually a requisite for some of the harder-core EDM and not using very significant amounts of compression on the electric guitar and vocals of say thrash metal would most probably change it's genre, it would no longer be thrash but some other sub genre.
 
I doubt that reducing the amount of extreme processing would change the genre, if the tune was a good one in that genre to begin with.  From experience in broadcasting, changing processing doesn't attract or drive away listeners, they listen for content.  Processing may make something more easily audible in some conditions, or less attractive in others, but it never changes the content.  I've run stations through radical processing changes that should have impacted TLS, Cume and total numbers if listeners detected a shift in genre.  It never did, not even a tiny bit.
 
8. And this answer maybe explains some of your other answers. We know that "far larger numbers of people like and purchase music of those genres than the acoustic genres" because for many decades we've had the sales figures!
 
That has nothing to do with compression, it has everything to do with what the music is.  Nice quote, but it's not mine.
 
Of course consumers have a choice in the final processing, they can choose to buy another piece of music with different processing.
 
Like Mom used to say, "You have two choices for dinner: take it or leave it."  That's not a choice.  If I want Song 15 by XYZ group, I have no choice but to accept what they've released.  The ONLY possible option may be a vinyl release where the choice of less processing may have been made.  Otherwise, you have no choice for a given track.
 
Yes, they buy it for what it is BUT, processing is not besides the point, processing is what defines "what it is" in the first place! This fact, which trumps any number of studies, is not only what has caused the loudness war but the very evolution of the many popular music genres themselves!
 
I'm afraid we disagree on this, and I don't see us coming together on it.  I don't accept that extreme processing defines a genre.  I've already explained why I feel this way.
 
10. "Largely transparent" was perhaps the wrong term, unobjectionable would perhaps be better. And, I wasn't talking of just music recordings but of all commercial audio.
 
Now you've expanded it to "all commercial audio"?  We'll disagree again on that one, unless film isn't commercial audio.  Unobjectionable or not is in the ear of the listener, very subjective.  Processing research shows people favor less processing across the board.  Less doesn't have to mean none.
 
11. What makes you think that the choices made by broadcasters reflect listener preference? Nothing at all, in fact quite the opposite, which is why the choice was taken away from the broadcasters and is now defined by federal law (in some countries). For this reason, the "goals" you talk about no longer exist in TV and haven't for a few years. "Your concept of a brick-wall limiter is also blatantly incorrect" I've used one almost every working day for more than 2 decades, so I don't think so, not to mention that I'd loose my job if my printmasters were not precisely limited to -1dBTP (or -2dBTP for my North American clients).
 
So, do you insure that your printmasters are precisely limited to -1dBTP by driving the mix 10dB above threshold of a fast attack/fast recovery brick wall limiter?  Or is it done tastefully as appropriate to the mix and end use?  
 
The studies I'm aware of, which support a preference for a greater dynamic range, are all contingent on the conditions I detailed previously. In other conditions, in a car as you mentioned and these days also in a bus, train, plane, cycling, when walking the dog, doing the homework, jogging and numerous common other conditions, this would obviously not be the case.
 
You of course know that the total dynamic range can be effectively reduced and be made appropriate to any of those situations without slamming into a brick wall.  Or is that the fix for everything?
 
Lastly, this thread isn't specifically about the extreme overuse of compression which characterises the loudness war and tkteo was not specifically asking about the loudness war but just the use of compression.
 
Perhaps that's at the core of our "problem".  I can late to the thread, assumed he was asking about compression that was objectionable.  
 
The amount of compression, both per track and master-buss compression, which is not only acceptable but actually indicative of EDM would generally be obscene if applied to a genre such as classical music. I'm certainly no supporter of the loudness war, quite the contrary, I've been discussing and complaining about it since the beginning of the '90's but I'm also concerned by the complete opposite, extremist view I've seen quite often on this board, essentially that compression is an evil which should be eradicated.
 
Yes, eradication of all compression would be a terrible mistake.  
 
This view is even more potentially damaging than the loudness war itself, it seems to be based on ignorance. The loudness war is not about compression, it's about the abuse of compression!
 
Agreed!!!!!
 
Again, late to the thread, perhaps I should have read the whole thing first before offering explanations. Sorry.