[Husky]

2. (...) various forms of distortion/lowered fidelity are ALWAYS employed and there would be no rock and roll or any other popular music genres since the 1950's without it.
3a. Why do you think compression was invented? Why do you think various forms of replay-gain exist? Why do most regions have TV broadcast loudness normalisation laws/requirements? Why does YouTube loudness normalise? Why do you think cinemas and film mix stages are level calibrated? Again, you're making insulting statements without thinking about or having any real understanding of the issue.
4. No you can't! OK, that's not absolutely true, in theory you could see compression and measure it, IF you had an un-compressed channel/track and then compared it with that same channel after compression (and no other processing). Do you have that/those uncompressed and then compressed (with no other processing) channels? No you don't, no one does!

2.
Analog guitar amplifier distortion is completely different from digital clipping of e.g. a female voice. The former is intentional for expressing music by the artist, the latter intentional for having a loud CD, sacrificing for sound quality.
So both are intentional, but one for the good of the listener, one for the bad. There may be artists who use hypercompression for artistic, not loudness reasons. But I don't see market compatibility (the wish to have the same - bad - sound as others) as a musical artistic choice.

3a.
So do you claim that the loudness war is unneccesary because of loudness normalisation? I hope that the increasing regulation will move the industry towards better sound quality.
Or at least to let the customer choose - why not make a more dynamic version available? Could be the unmastered mix, or if there is a vinyl version make it digitally downloadable.
And what is your understanding of the issue (other than everything is OK because the industry/artists think it's OK)?

4.
You only need both if you want to exactly measure compression. But you can compare any track against pink noise.
If you don't believe it, take ClippingAnalyzer and try it out on some tracks where you know what compression was applied. Be aware, that light compression can still show as 0% and uncompressed material may show >0%, but heavy compression is quite well recognized.
I'd be glad to support you with the analysis, you can give me some test tracks if you like.

 

[gregorio]

The SACD is mastered at about -35dB RMS, and I swear 20dB of that is just for this one 'ping' on a glockenspiel. Been wanting to set up a blind test on that one for some time to see how much can be taken off the top.

This raises another interesting point (beyond the one you are making), a point which is virtually always ignored but which makes a huge difference and that is; the difference between what an instrument actually produces and how it is heard (or expected to be heard). Metal perc instruments, particularly those struck by hard beaters (like a glock, gamelan, triangle, etc.), produce massive transients. However, even the front row of an audience is going to be at least 30 feet away from the glock, most of that massive transient (and it's substantial HF components) is simply not going to be there any more, it's going to be absorbed massively by everything, the floors, walls, intervening musicians and their chairs and even substantially by the air itself. What the instrument actually produces and what we would actually hear are two vastly different things however, what we record could be either, depending on where we position the mic. If we were to place a mic a foot or a few inches from the glock we'd (more or less) record exactly what it actually produces and therefore, we'd have to apply some processing to bring it in line with what an audience would expect to hear. So, should that 20dB glock peak even be in the master to start with? Is it merely a consequence of the mic/s positioning that could/should have been reduced/dealt with? Logically of course, a glock ping should not overpower an entire orchestra going full pelt (fortissimo) by 20dB. This is just one of numerous examples of where compression is (or could be) vital and not the "black and white" evil it's made out to be.

In addition to this, there's your point that a few dB of compression cannot generally be heard but how much cannot be heard entirely depends on what we're compressing and the type of compression used. And this brings up another overlooked/misunderstood point; compression/limiting isn't "a thing", it's a wide range of things with numerous different characteristics. They cost anywhere from $0 - $50,000 (and $0 does not mean bad, just different) and any completed commercial music product from the last 40+ years is going to have at least 2 different compressors applied, possibly as many as 8 or so, applied at various stages (throughout mixing, mastering and on older recordings, even during recording) and some of them used multiple times with different settings. So, suggesting no compression and then the consumer adds their own is nonsense.

I think the old school way of looking at peak was as an emotional climax point, with the expectation you were waiting for it.

There was no old school way of looking at peaks, you couldn't see them! Analogue (VU) meters had a measurement window of about 0.3 secs (300ms) but transient peaks are in the few (or few tens of) milli-secs range, so you either couldn't see them at all or their level was vastly under reported! With digital on the other hand our measurement window (with CD, 44.1kS/s) is 0.00002 secs (22ns or 0.02ms) and not only can we "see" the transients but we can "see" different parts/segments of the transients. Of course though, with analogue you didn't get digital clipping, just compression/saturation/breakup.

Your post though is a little confusing, it's not clear whether you're talking about physical peaks or musical structure, which raises yet another very important and overlooked/misunderstood point. I warned you, this is a rabbit hole!! Some of my answers may have appeared rather glib and that's because a full answer would take too long. Part of the problem is that from my perspective, I'm some way down the rabbit hole while many of the posters here are just looking in from the outside and haven't even noticed that there is a rabbit hole! Communication is therefore going to be an issue without some common ground, some idea of the basic layout of the rabbit warren and as this is the science forum, I'm presuming people are here to gain some understanding of what's really going on under the surface:
For this reason, I very strongly suggest reading this SOS article before we try to continue. It's the best I've seen on the subject because it's saying it as it really is, not the massively over-simplified soundbite/propaganda bilge which was necessary to get consumers interested in the issue! BTW, the article doesn't give all the answers, just a more reasonable place to start.

It's quite nonsense in my view to squeeze the sound to gain e.g. 15dB and then the material doesn't use the full dynamic range of the medium.

The reason your view is nonsense is because if we did "use the full dynamic range of the medium" not a single person would like it, not even an extreme audiophile! All we'd get is complaints and demands for refunds and that's why when we make very dynamic recordings we literally use about 1,000 times less than the full dynamic range available!

... if you want to hear the human voice in all its dynamic glory go to a church with 1/2 decent acoustics ...

Actually, I'd say the exact opposite, that a church is just about the worst place to go for "dynamic glory". In a church you've typically got a large space with very reflective surfaces and therefore a long, dense reverb with an RT60 of at least 4 secs and up to around 8 secs (and nearly double that in a large cathedral). This means notes overlap with the reverb of previous notes, there's no silence between notes and the dynamic range is reduced by very wet acoustics. You want to hear the dynamic glory of the human voice, try being about 3ft away from Placido Domingo in a broadcast studio. That's close enough to hear his jacket rustle as he breathes in and close enough to feel like you're in the middle of an artillery barrage! I'm not joking, the amount of volume was literally staggering (and the noise I made when I did stagger, made him aware of my presence and he stopped and apologised), there are very few mics which woudn't be damaged if you tried to place them where you normally would for a pop/rock singer. I wasn't even directly in front of him and it was easily over 100dB.

G

 

[RRod]

This raises another interesting point (beyond the one you are making), a point which is virtually always ignored but which makes a huge difference and that is; the difference between what an instrument actually produces and how it is heard (or expected to be heard).

Your points are exactly what I was hoping someone would expand on. Note my 20dB was an exaggeration, but the reality is that you can probably take 10dB off the peaks and no one would be the wiser. Now sure, as someone pointed out, on an SACD there's no need to change this, but it's certainly easier to amp these kinds of things then you've got 10dB less you need to pump out. For someone using something like wireless headphones that can make a real difference.

 

[gregorio]

Analog guitar amplifier distortion is completely different from digital clipping of e.g. a female voice. The former is intentional for expressing music by the artist, the latter intentional for having a loud CD, sacrificing for sound quality.

The former is intentionally making louder and distorting the sound quality for the purpose of expressing music by the artist and the latter is intentionally making louder and distorting the sound quality for the purpose of expressing music by the artist. There is NO DIFFERENCE, except that you personally don't like the latter. There were a great many who didn't like the sound of a louder, distorted electric guitar when that started being used too! It took Hendrix years to get accepted and he had to go to another country for his sound and talent to be recognised. I personally don't like clipping but I still absolutely defend those who employ it, I'm rather more forgiving of very heavy compression but in every case it's about appropriateness, artistic intent and what/how/when one chooses to express it.

Read the article to which I linked and then we can talk, until then most of what you're saying, particularly about my views, is incorrect or complete nonsense!

Your points are exactly what I was hoping someone would expand on.

And that was my hope with some of my glib comments but rather than expanding, some are more interested in digging their heels in and contracting, that's sad by itself and even more sad because it appears to be such a common audiophile response!

Note my 20dB was an exaggeration ...

Actually it's not, I've done a great deal of percussion recording and indeed it was my speciality for a number of years. With instruments like a glock (+ gamelan and some others) massive transient peaks of up to 20dB is certainly not beyond the bounds of reality. In practice though, depending on the circumstances, one would ideally not mic too closely, in order avoid capturing those transients so strongly.

G

 

[bigshot]

Do you have a volume knob in a classical or jazz live concert (without PA system)? No. And a big orchestra can have >100 dB.

Recorded music rarely exceeds a dynamic range of over 55dB. It's uncomfortable to listen to excessive dynamics. More isn't necessarily better. In order to hear all 100dB of dynamic range, you would have to raise the volume to the threshold of pain to overcome the noise floor in your listening room. The 100dB figure you're citing is based on sitting right in the middle of the orchestra, not the sound you would hear sitting in a seat in the audience. No music is made for dynamics as broad as you cite.

You want to hear the dynamic glory of the human voice, try being about 3ft away from Placido Domingo in a broadcast studio. That's close enough to hear his jacket rustle as he breathes in and close enough to feel like you're in the middle of an artillery barrage!

I can't even imagine that! I heard Pavarotti in recital, and I was over 50 feet from him and my ears were ringing when he hit the high notes. Caruso does that even on acoustically recorded 78s. Some voices really cut through!

 

[Husky]

(...) I personally don't like clipping but I still absolutely defend those who employ it, I'm rather more forgiving of very heavy compression but in every case it's about appropriateness, artistic intent and what/how/when one chooses to express it.

So we have in common that we don't like clipping and very heavy compression, for our personal taste, right?

I'm simply frustrated when I like a track because of the music (melody, singer, instruments, rhythm, arrangement etc.), but there is distortion that distracts from the musical quality. Maybe I'm more severely irritated than average listeners because I know how it could sound and I know the cause of the irritation.
My hearing of high frequencies isn't as excellent as it was 20 years ago, so I'm a bit less distracted by clippings nowadays - at least one positive side effect of getting older

Could you please explain to me why e.g. a female singer would choose to have distortion on her voice, only on the CD, not live? E.g. Amy Winhouse, Back to Black.
And why do many people prefer the vinyl version over CD version where the CD is clipped and much more compressed?

As a consumer I don't have to defend any choices of either the artist or any other part of the music industry. I understand that people in the business feel offended by critics, because somehow they want that their work is appreciated. But I feel very sad about pure ignorance of arguments against the loudness war and for better sound quality.

In your experience, do all artists really tell the engineer to brick wall with audible distortions because they feel it was musically better, comparing the versions listening with same perceived loudness? Not for reasons of "it has to be as hot as the market"?

Read the article to which I linked and then we can talk, until then most of what you're saying, particularly about my views, is incorrect or complete nonsense!

I know the article and read it again. I can't see much controversy to my statements, other than maybe the interpretation of music style and the use of loudness measures.
What and why exactly do you think is nonsense what I'm saying?

The 100dB figure you're citing is based on sitting right in the middle of the orchestra, not the sound you would hear sitting in a seat in the audience. No music is made for dynamics as broad as you cite.

In the orchestra it can be >120dB at fortissimo, resulting in >100dB in the audience (first rows). Of course this isn't the normal listening situation, but with good equipment at home it's possible and - in my view - desirable to get an experience nearest to live. (Edit: I prefer if the orchestra doesn't play that loud. Generally I don't like extremely loud concerts)

 

[bigshot]

In the orchestra it can be >120dB at fortissimo, resulting in >100dB in the audience (first rows). Of course this isn't the normal listening situation, but with good equipment at home it's possible and - in my view - desirable to get an experience nearest to live.

Do you know what 120dB sounds like? It's the threshold of pain. It's as loud as an ambulance siren at close range. Exposure for as short as 8 seconds could cause hearing damage. It's important to know what those numbers mean in real life sound.

 

[Strangelove424]

There was no old school way of looking at peaks, you couldn't see them! Analogue (VU) meters had a measurement window of about 0.3 secs (300ms) but transient peaks are in the few (or few tens of) milli-secs range, so you either couldn't see them at all or their level was vastly under reported! With digital on the other hand our measurement window (with CD, 44.1kS/s) is 0.00002 secs (22ns or 0.02ms) and not only can we "see" the transients but we can "see" different parts/segments of the transients. Of course though, with analogue you didn't get digital clipping, just compression/saturation/breakup.

Your post though is a little confusing, it's not clear whether you're talking about physical peaks or musical structure, which raises yet another very important and overlooked/misunderstood point. I warned you, this is a rabbit hole!! Some of my answers may have appeared rather glib and that's because a full answer would take too long. Part of the problem is that from my perspective, I'm some way down the rabbit hole while many of the posters here are just looking in from the outside and haven't even noticed that there is a rabbit hole! Communication is therefore going to be an issue without some common ground, some idea of the basic layout of the rabbit warren and as this is the science forum, I'm presuming people are here to gain some understanding of what's really going on under the surface:

I meant in terms of style. Particularly in the remasters, since those are the only examples where one can make a direct comparison of old vs new mastering on the same song. Though its impossible to get in the head of every musical engineer, the style of the older masters seems to save peaks for emotional climaxes rather than to keep that level constant throughout the song. There seems to be more emphasis in modern mastering with keeping a constant, homogeneous presence. It's also important to keep comments in context to genre or better yet specific material, so to be clear I am not addressing the often excessive range of classical, I am being very specific to rock (the previously mentioned Dire Straits and Fleetwood as examples).

 

[Husky]

Do you know what 120dB sounds like? (...) Exposure for as short as 8 seconds could cause hearing damage. It's important to know what those numbers mean in real life sound.

Yes, I know and I don't like it that loud. But still there are concerts in every genre that are way too loud, with peaks around 120dB in the audience. And it's legal, e.g. in Switzerland peaks (LAFmax) of 125 dB(A) are allowed. The hourly mean level (LAeq1h) is allowed to be 100dB(A), if the audience is informed about it and free ear plugs are available. Personally I think it's perverse if we need ear plugs in the audience (that diminish sound quality) in order to prevent ear damage.
Sadly, ear damage is widespread, especially among musicians.

 

[Strangelove424]

Just as an aside: It occurred to me a while ago that these measuring tools were only good at coming up with numbers, not interpretations, or much less decisions. If you like the dynamics and sound of a song, you really don't need any waveforms or numbers to prove it to you. That makes these tools sort of redundant. Which means their only purpose is to prove to others what you personally might be hearing in a song, which in turn means they are fodder for forum arguments. I think it was partially this realization that resulted in my reduced participation in this thread (in addition to many bans that were happening at the time.) I am just thinking out loud. Carry on.

 

[bigshot]

Mixing and mastering is a creative process. It's important to have a foundation in consistency and technical stuff, but that isn't what makes a mix great.

Yes, I know and I don't like it that loud. But still there are concerts in every genre that are way too loud, with peaks around 120dB in the audience.

Weren't you saying that classical concerts hit 120dB? Of course amplified concerts do, but I would still stick cotton in my ears before subjecting myself to that. Recorded music flat out doesn't require that kind of dynamic range. Ears can only hear about 45dB at a time. They need time to adjust if you are going to extend beyond that. Music doesn't generally allow for that long of a period between dynamic contrasts. That's why most music is between 40 and 50dB. Go beyond that and you have to listen to it too loud to be comfortable.

 

[castleofargh]

someone will always be unhappy about the mastering choices of a specific album. I understand those who somehow hope to get the emotion of a loud and very dynamic experience at home. crank the amp up and imagine being at a live event. I also understand that if I had only that type of albums with how long I spend each day listening to music, my ears today would be absolutely ruined and I'd be one of those guys with an almost permanent headache. or I would play it cool and fail to hear half of the songs because with the dynamic it would be too quiet.
the general loudness war was/is a monstrosity, pretty much everybody agrees. but first, there's a reason why it was used so much, mainly because it worked. let's blame stupid subjectivity instead of the people who found out how to make use of it(which is exactly what they were paid to do all along!!!!!!!). and second, it never ever meant that compression in general was bad or evidence of a sound engineer who doesn't know his job. the polarized view about compression needs to go away even more than the loudness war itself. I don't know what I would do if all my favorite bands started releasing only albums with uncompressed sound. that would be the real nightmare for me.

 

[gregorio]

[1] So we have in common that we don't like clipping and very heavy compression, for our personal taste, right?
[2] I'm simply frustrated when I like a track because of the music (melody, singer, instruments, rhythm, arrangement etc.), but there is distortion that distracts from the musical quality.
[3] Could you please explain to me why e.g. a female singer would choose to have distortion on her voice, only on the CD, not live? E.g. Amy Winhouse, Back to Black.
[3a] And why do many people prefer the vinyl version over CD version where the CD is clipped and much more compressed?
[4] I understand that people in the business feel offended by critics, because somehow they want that their work is appreciated. But I feel very sad about pure ignorance of arguments against the loudness war and for better sound quality.
[5] In your experience, do all artists really tell the engineer to brick wall with audible distortions because they feel it was musically better, comparing the versions listening with same perceived loudness? Not for reasons of "it has to be as hot as the market"?
[6] I know the article and read it again. I can't see much controversy to my statements, other than maybe the interpretation of music style and the use of loudness measures.

1. No, that's not exactly what I said! Firstly, they are two different things, you can have very heavy compression without clipping and vice versa. My reason for disliking clipping isn't that I necessarily dislike it per se but that the result of clipping is somewhat unpredictable. Meaning, that the effect of clipping heard in the studio may not produce that effect on another system, IE. On a particular consumer system the clipping could sound worse. While a mix (and the master) are tested on a number of different consumer systems because of the esoteric design philosophy of some/many audiophile manufacturers, the results of clipping could be worse on an audiophile system than on a professional studio system or a cheaper consumer system. As far as very heavy compression is concerned, whether I like it or not depends on what it's applied to and for what purpose. For example, very heavy compression on most classical music I generally wouldn't like, although there are some exceptions. With some other genres and some processing or certain instruments, then very heavy compression is pretty much a necessity. It's like saying do I like a lot of sugar on my food; on a steak, certainly not but baklava without a lot of sugar wouldn't be baklava. There's no black and white answer.

2. You're basically just repeating the same thing over and over just using different wording. Rather than just doing that, stop for a minute and try to think about it because what you are saying is exactly what has been said countless times throughout the history of recorded music. The argument at one time was: "Why does an electric guitar need to be so massively distorted that it detracts from the music quality?", which is effectively the same as; "it's such a shame, I could like or at least appreciate Hendrix (for example) if he didn't ruin it with so much distortion". Today we (consumers) are much more accustomed to very heavily distorted guitars and we generally have a better understanding/appreciation of it, we realise that the distortion and the musicality are not two different things which could (and should) be separated, so much so in fact, that today we don't even question it. Of course, some people never really "got it", my mum eventually accepted that people thought Hendrix was a genius and there were some bits that she would have liked but then at some point in the song he would always ruin it with some "screechy, horrible noise", the very thing that every really good musician learnt to avoid. Like you, my mum was "frustrated"!

3. That's a tough question to answer in a few words, as there are numerous potential reasons. One is a technical reason, some effects require very fine adjustment, in the case of heavy compression or brickwall limiting tenths of a dB can make a big difference. That's not a problem in the studio when mixing or mastering because as it's already recorded we know precisely what level/s we got, not so in a live situation where we've only got a vague idea of what the input signal levels are going to be.
Back to Black (Amy Winehouse), is a good example but again, a potentially very long answer. So again, I'll restrict myself and give just give a couple of short ones. 1. What is Winehouse's musical influences? Obviously jazz but also others, such as hip-hop. What is the structure and mix style of hip-hop? 2. Much of this particular song has a sort of angry disappointment feel but there's a section near the end where it changes to more pure sadness. Angry disappointment is quite harsh and she achieves that in places by constricting her throat and producing a more nasal tone, similar to that harshness you sometimes hear in a Brooklyn accent. This can be accentuated in the mix in various ways, the way they've chosen to do it is probably not how I'd have done it but then I'm maybe a bit more old school. I'd have perhaps relied more on some subtle EQ to create the harshness and created more contrast with the sad section but that's just me. They've used a more new school approach and applied more compression to her voice than I would have, although I'd still have used quite a lot because she has a voice which often needs it.
3a. How many more millions of people preferred the vinyl?

4. If you're so sad about the "pure ignorance", then why don't you do something to try and cure your ignorance rather than just keep coming out with nonsense statements? Who do you think it was who first started complaining about the loudness war? It was already a hot topic of debate amongst engineers when I entered the business in 1991, how do you think the "critics" got to hear about it years later?

5. In effect, yes.

6. Well those two exception would be a great place to start!!

In the orchestra it can be >120dB at fortissimo, resulting in >100dB in the audience (first rows). Of course this isn't the normal listening situation, but with good equipment at home it's possible and - in my view - desirable to get an experience nearest to live. (Edit: I prefer if the orchestra doesn't play that loud. Generally I don't like extremely loud concerts)

You are confusing peak level with dynamic range. Yes, in the first row of an audience you may get 100dB but then it's unlikely the noise floor is lower than about 40dB, so that would be a dynamic range of 60dB, which is 1,000 times less than the dynamic range of SACD or CD!

G

 

[bigshot]

I've been to many concerts and operas, and I can't think of any of them that had peak levels of 100dB or more. Even the immolation in Gotterdammerung sounded like it wasn't hitting 90. And I don't know why anyone would want to sit in the front row in a classical concert unless it was a soloist. It seems that the justification for crazy technical requirements has very little to do with real world music making.

 

[RRod]

I've been to many concerts and operas, and I can't think of any of them that had peak levels of 100dB or more. Even the immolation in Gotterdammerung sounded like it wasn't hitting 90. And I don't know why anyone would want to sit in the front row in a classical concert unless it was a soloist. It seems that the justification for crazy technical requirements has very little to do with real world music making.

I think of the charts that a certain someone posted where the 'ideal' listening position for a rock concert hit 130dB SPL (at which point peak/rms is pretty meaningless). Where the heck are people getting such crazy levels? When I absolutely BLAST my stereo system with something like the opening of the fifth door, my sound meter gives me in the 90-100dBC range for max fast RMS. There's no way the crest factor is 20-30dB over a 125ms time frame.