• Welcome to the all-new HUG. All content has been converted from the old system, and over the next days we will re-style HUG in a more familiar way.

Getting to grips with "clipping" - an ongoing research project

A.S.

Administrator
Staff member
Listening analytically for ....

Listening analytically for ....

...
To illustrate the difference between well-recorded and over-compressed music, I've picked a couple of tracks:

  1. Dire Straits - Money For Nothing. Taken from the original release (824 499-2)
  2. Red Hot Chilli Peppers - Scar Tissue. Taken from my wife's collection :)
These were chosen more or less at random - picked because they are well-known. I used Exact Audio Copy to extract them, and placed 20 second excerpts together. I've done nothing to adjust levels - this is how they come off the CD. Set the Dire Straits to a comfortable level, but be aware that the RHCP is much louder!

Having listened and satisfied yourselves that the average (RMS) levels are very different, what do you think about the peak levels? Do they vary by a similar amount? And what might that tell us about the power requirements for both tracks?

Just to be clear about the things you are encouraging us to listen for ....

Evaluation 1: loudness

If, as you say, what humans perceive as the loudness of an audio signal is related to its RMS value (another way of saying the heat energy in the music) then we must reach an opinion about how loud the two excerpts are

Evaluation 2: Presence (or absence) of peaks in level

Setting aside the loudness of the excerpts, we must listen to see if we can hear peaks in loudness

Evaluation 3, (follows from Evaluation 1 & 2): ratio of the peaks to the average loudness


We must try and determine the shades of loudness between peak white and average grey in the sound clips; the signal level never drops to black silence: there is always some sound. If the tracks were monochrome photographs, would there be lots of fine graduations of grey or would the sound picture be hard and stark? In other words, we must arrive at an impression of the peak to mean ratio in the music excerpts.

Is this correct?
 

mhennessy

Member
Listening approach

Listening approach

Is this correct?
I think so.

There's no doubt that excerpt 2 is much louder than excerpt 1. But the peak levels are what interest me: I don't think we're able to subjectively judge their size, which is what makes this interesting. Much more to come...

But as an initial thought, it's tempting to assume that the Dire Straits was simply recorded at too low a level. Many pressings from the early days of CD are accused of being cut at a 'conservative recording level' well within and never approaching the full loudness potential of the CD - as if engineers of the day didn't understand the medium - but as always, the real answer is rather more subtle than this. Any thoughts?

Mark
 

tmokbel

Member
Listener feedback -1

Listener feedback -1

I think so.

But as an initial thought, it's tempting to assume that the Dire Straits was simply recorded at too low a level. Many pressings from the early days of CD are accused of being cut at a 'conservative recording level' well within and never approaching the full loudness potential of the CD - as if engineers of the day didn't understand the medium - but as always, the real answer is rather more subtle than this. Any thoughts?

Mark
I just played both tracks on i-tunes to a DAC and then to my preamp/amp/speakers. The Dire Straits track gives a wonderful amount of gain on my pre amp being able to listen anywhere from 9oclock to 1 oclock on the volume pot and sounds natural, but with the RHCP track, I can't go above 10 o'clock and it's already too loud.
 

A.S.

Administrator
Staff member
Feedback -

Feedback -

...There's no doubt that excerpt 2 is much louder than excerpt 1. But the peak levels are what interest me: I don't think we're able to subjectively judge their size, which is what makes this interesting. Much more to come....
I think we'd all agree that the second excerpt is much louder than the first, probably uncomfortably so.

What I hear is that the Dire Straits (first clip) has a naturalness about it. There is some air around the performers. Yes, there is the repetitive drum beat which gives an impulsive sound, and there is reverb around that impulse which makes it seem believable, as would be heard live. The impulsive drum beats are obviously loud, quite a lot louder than the average for the song, but the sonic contrast between the sharp beats and the rather bland tune makes the overall sound interesting. The impulses have an energy and tonality which blends with the music. If they were akin to deep scratches on a gramophone record, they would be distracting because they would be atonal, and their perceived loudness greater to accompany their irritation factor. The fact that the impulses blend with the music rather than fight it, makes them seem softer and lower in level than they probably are.

The second clip has no air around it at all. The vocalist seems to be right up at the microphone, as are the musicians. There is no sense of space or air: the sound is very dry, as would be recorded outside. The entire piece is revving on the red line: there is no dynamic range, no sense of light and dark: it's all shouted at the listener.

I guess what you are interested in is our perception of the levels of the impulsive drum beats in the first clip compared to the (squashed) peaks in the second? You might surprise us by saying that, counter intuitively, those peaks in the Dire Straits track are of similar level to the perceived (high) levels in the RHCP clip. In which case, it would suggest that the ear is not at all good at determining peak to average levels; test equipment (like your 'scope) is excellent at doing just that. And if that is true, then, as you have been suggesting, if we are not really capable of hearing peaks in music buth our power amplifiers are obliged to try and follow them (until it runs out of PSU volts aka dynamic headroom) then our power amps are most likely to be driven into clipping more often than we realise: and we may be oblivious to that.

I must say that this concept of the dynamic range, the peak to average ratio of music is fascinating. I'd never really appreciated how significant it is for the humble audio electronics. And the human ear, really not at all reliable as a level instrument, as one hundred years of research shows.

To answer your question, I suggest two factors ...

1. The listener would probably turn up his amplifier when listening to the DS track, because the average loudness is low, and, especially after a glass or two and getting into the groove (!), up would go the volume control .... and
2. The listener would probably be so fatigued by the machine-gun intensity of the RHCP clip that he would turn down his amplifier when listening to it

And that leads to, again, a counter intuitive deduction (I suppose) ....

From the amplifier's perspective, playing the RHCP (second) clip is much less taxing on its PSU because a more or less continuous power drain is occurring, where as in the DS (first) clip, under real-world home listening (after a glass), whilst the average power drain between the drum impulse is tolerable for the PSU, even played loud, those impulses command a huge rush of volts to the speaker, which the amp may or may not be able to deliver from its already hard working PSU. If that is so, playing the Dire Straits excerpt at a decent volume with full fidelity and absence of clipping, is going to be a much more challenging prospect for the amplifier.

Does that sound credible?
 

mhennessy

Member
When we crank it up ....

When we crank it up ....

Hi Alan,

Yes, you've nailed it!

I've attached a screen shot of the clip, and you can see the dramatic differences between them. Not surprisingly, the second excerpt, the RHCP piece is right up against the end-stops, whereas the Dire Straits is much more open and dynamic. For a "rock" recording, it's amazingly well produced - it's the sort of track that makes you want to "crank it"; you can't help but turn up! But sadly, the RHCP track has exactly the opposite effect.

But yes, both tracks are peaking to nearly the same level - there is just 3dB in it.

When you compare the RMS levels (by watching the light green part of the level meter on Audacity) you see that the RMS level for the Dire Straits track varies between -30 and -24dBFS. For the RHCP excerpt, it's between -12 and -9dBFs. Quite different... As a test, I decided to reduce the level of the RHCP excerpt such that it sounded the same as the Dire Straits. I found that I needed to bring down the level by 16 decibels to make the two clips appear to have the same subjective loudness (to me, at least). But of course, despite seeming to have the same perceived loudness, the RHCP track sounded dull and lifeless compared to the Dire Straits.

You must be registered for see images attach


(As an aside to all this, broadcasters are moving across to loudness meters, rather than near-peak meters that are currently used. In an equal-loudness world, we hope that producers will realise that they need to re-introduce dynamic range into their recordings if they want them to stand out.)

So, what is the implication for power amplifiers?

As you say, the Dire Straits track will be a challenge for the amplifier, and clipping is quite likely. When I tested this a week or two ago using a 50 watt amplifier, I was actually quite surprised at how readily it clipped. I still wanted more level! Whereas when I tried to "crank" the RHCP track, I gave up at about 10 volts peak, which was only about a third of the available voltage.

While we as humans perceive loudness as a result of RMS levels, amplifiers must be concerned with peak voltages and currents. For high quality recordings, amplifier headroom is extremely important.

The next question is this: how does clipping sound?

In short, it depends! Shall we explore this further?

All the best,

Mark
 
Attachments only viewable to members

tedwin

New member
Yes! Keep going!

Yes! Keep going!

Shall we explore this further?

All the best,

Mark
Please do! An amplifier designer in open conversation with a speaker designer discussing what's behind the fidelity of music reproduction in the homes of hifi consumers at a technical, yet understandable level. Open to questions along the way, what more could you ask for?

Ted.
 

weaver

New member
Yes - go!

Yes - go!

The next question is this: how does clipping sound?

In short, it depends! Shall we explore this further?

All the best,

Mark
There is much here that I feel I am finally beginning to gain an understanding of, so thank you and yes please.
 

anonymous

New member
The practicalities of amplifier power needs

The practicalities of amplifier power needs

Absolutely. Please continue, Mark. But at some point I hope to find out what this means for someone like me who has speakers with the following specifications:

Impedance: 6 ohms - easy to drive.
Sensitivity: 83.5dB/1W/1m
Amp. suggestion: Works with a wide range of amplifiers, ideally from 15W/channel.
Power handling: 50W programme

Admittedly, I am getting a bit lost with some of the technical details, and I may be misunderstanding, but I find myself thinking that the next time I buy an amplifier I should just look for something with at least 300 watts that weighs 100 lbs just so I don't have to be concerned about headroom. I can't help reading all of this and thinking that "Too much is better than not enough." Even with my imperfect ears.

At some point in the discussion, it would be very helpful if Alan would explain Mark's findings in relation to specific Harbeth speakers.

1. What does "power handling of 50 watt programme" mean?
2. And would someone using 15w/channel be in a constant state of clipping with high quality recordings?

I may have wondered about this in the past, but the fact that Mark's 50 watt amplifier is clipping readily with a speaker of similar sensitivity and high quality recording brought these questions to mind with a bit more vivacity.

I'm really trying to understand this, and it's not coming easily to me. Of possible interest, I came across the following while mulling this all over:

http://sound.westhost.com/amp-sound.htm

It seemed useful for the summation of measurable performance characteristics, but I'll leave it to Mark and Alan to decide.
 

mhennessy

Member
More examples of clipping

More examples of clipping

OK - many thanks for the encouragement :)

There are two sides to the clipping coin:

  1. The likelihood of clipping occurring, which obviously depends on how powerful your amplifier is; how efficient your loudspeakers are; and, as we've seen, the dynamic range of the source material. The more dynamic the music, the more likely you are to run into clipping.
  2. How noticeable is clipping when it does happen. Again, this is strongly dependant on the material being played, but it's actually really complicated.
We've probably dealt with side 1 sufficiently well. The other side needs careful consideration now. I suggest that we have already seen the two extremes:

  1. Earlier we discussed sine waves, and concluded that it is very easy to visually notice and hear even very slight clipping with a pure sine wave.
  2. And with the snare drum in the Dire Straits track, it's actually pretty hard to hear something that you'd identify as clipping.
Initially, I wrote a long explanation about clipping sine waves verses clipping musical instruments, but decided that this was too much for now. I've saved it, and can regurgitate appropriate bits of it as the need arises. Instead, I decided that it would be better to create some audio files.

To start with, we have the Dire Straits excerpt - 3 times in a row this time. (See Post #55 for screen cam)

  1. Unprocessed, as it came off the CD
  2. I've applied hard-limiting to -12dBFS, which gives measurable, but fairly benign clipping
  3. Hard-limited to -18dBFS, which is clearly distorted
I'd be interested in your comments, and especially regarding 1 verses 2. If you can put the differences into some sort of context, that would be really interesting. Have you heard similar differences before? Have you heard more or less subtle differences when comparing equipment? Try to use a range of listening levels and equipment, and don't be shy - I'm interested in subjective opinions, and don't necessarily expect them to align with mine.

Mark

PS: Bear in mind that these tests aren't perfect indications of real-world clipping because real amplifiers do add their own "signature" when clipping, but there are advantages of distorting known recordings solely in the digital domain: anyone can repeat them, and we can also produce perfect "difference tracks" to highlight the alterations to the signal.
 
Attachments only viewable to members

mhennessy

Member
Practical advice

Practical advice

Hi Anonymous,

Admittedly, I am getting a bit lost with some of the technical details, and I may be misunderstanding, but I find myself thinking that the next time I buy an amplifier I should just look for something with at least 300 watts that weighs 100 lbs just so I don't have to be concerned about headroom. I can't help reading all of this and thinking that "Too much is better than not enough." Even with my imperfect ears.
Hopefully we'll be able to provide guidance/reassurance about this as we progress. Certainly, a mini-monitor ought not need a super-amp! But so much depends, and it's really hard to provide a simple answer. The lazy answer might be "get an oscilloscope and learn how to use it". But I can't imagine this going down terribly well! The trouble is, there is such a large range of listening rooms, user preferences, and source material out there. I just wish manufacturers would fit clipping indicators to their amplifiers!

Of possible interest, I came across the following while mulling this all over:

http://sound.westhost.com/amp-sound.htm
Yes, that is interesting. I have studied a lot of the articles on that site, but had missed that one. The sections about clipping are definitely worth reading. Thanks for that!

Cheers,

Mark
 

tedwin

New member
Just not sure...

Just not sure...

I'd be interested in your comments, and especially regarding 1 verses 2. If you can put the differences into some sort of context, that would be really interesting. Have you heard similar differences before? Have you heard more or less subtle differences when comparing equipment? Try to use a range of listening levels and equipment..
Hi Mark

Am unable to devote as much time to this as would like to right now, but after repeatedly listening on my pc via relatively good headphones (incidentally I never actually listen to music like this, I only have the headphones for when at my desk at work, ideally I'd listen at home on my stereo like normal)

Anyway, at first I thought I could hear maybe more 'air' and maybe a longer 'decay' in the first clip. This is most likely because I just had a flick through the link provided by anonymous which showed wave forms squared off and 'sticking to the rail' When given visual clues to what clipping is its easy to apply that to what I think I'm hearing.

BUT the fact of the matter is that after a few goes through I can't tell the difference between one and three leave alone one and two! I think I would have to somehow be able to hear them at the same time (obviously not really possible) to tell any difference, it seems I have the audio memory of a gold fish :)

Alan's earlier examples with the full orchestra coming in were very obvious, this, not so much...

Ted.
 

A.S.

Administrator
Staff member
Preparatory listening step

Preparatory listening step

Hi Mark

Am unable to devote as much time to this as would like to right now, but after repeatedly listening on my pc via relatively good headphones (incidentally I never actually listen to music like this, I only have the headphones for when at my desk at work, ideally I'd listen at home on my stereo like normal)

Anyway, at first I thought I could hear maybe more 'air' and maybe a longer 'decay' in the first clip. This is most likely because I just had a flick through the link provided by anonymous which showed wave forms squared off and 'sticking to the rail' When given visual clues to what clipping is its easy to apply that to what I think I'm hearing.

BUT the fact of the matter is that after a few goes through I can't tell the difference between one and three leave alone one and two! I think I would have to somehow be able to hear them at the same time (obviously not really possible) to tell any difference, it seems I have the audio memory of a gold fish :)

Alan's earlier examples with the full orchestra coming in were very obvious, this, not so much...

Ted.
We seem to have two parallel threads running covering the same issue from different perspectives.

May I suggest that the clue to answering Mark's post #49 in this thread, is revealed in my post on the other thread, here.

If you answer my question about the "digital" v. "analogue" sound (both in quotes; they are merely commonly used adjectives in audiophile parlance, both clips are digital) then you are equipped to answer Mark's question. Some preemptive ear training is necessary. Few of us are working with such fine-grained audio issues as a matter of routine.
 

Kumar Kane

New member
Soft-clipping amp

Soft-clipping amp

I just wish manufacturers would fit clipping indicators to their amplifiers!
My NAD amplifier used for a 2 channel HT set up has a soft clipping indicator - most NAD amps have this or a switch to turn soft clipping on/off. I have never bothered to find out what this does, but here is something about it on the net that you might be able to make sense of:
http://www.ecoustics.com/electronics/forum/home-audio/11863.html
 

tedwin

New member
Listening

Listening

OK, so now tried it at home on my stereo. Differences now seem apparent.

I would describe version two as less punchy and thinner. Maybe as much or more to do with the bass guitar than the drum. But the high frequency part doesn't seem to change so much. Although that is where I was listening for a difference at first for some reason.

Ted.
 

A.S.

Administrator
Staff member
More explanation

More explanation

OK, so now tried it at home on my stereo. Differences now seem apparent.

I would describe version two as less punchy and thinner. Maybe as much or more to do with the bass guitar than the drum. But the high frequency part doesn't seem to change so much. Although that is where I was listening for a difference at first for some reason.

Ted.
Ummm. This is interesting. It's a pity nobody has addressed by question here (the other thread) because to my ears, Mark's audio examples of current remastering comprehensively explain the "digital" v. "analogue" sound which is an essential to comprehending the example he has posted in this thread. Same old observation though: not enough feedback for us to adapt our presentation to help comprehension. We just cannot be expected to create this stuff late into the night and then answer our own questions: that would be kinky.

It suggests one of three things: either without training the ear there will be a wide range of possibly random opinions about sound (non necessarily aligned with the objective facts), that we truly do need to abut audio event A and B to give ears a chance to make sense of what we are hearing, or that listener preference is counter-intuitive. Without feedback for the various audio clips, Mark and I are a bit lost as to how to proceed.

I've taken Mark's audio clip from his post #49 above (read for more info) and made a screen cam. I've added a forth excerpt - the opening ex-CD excerpt has been added on to the end, so the clip opens and closes with the raw ex-CD excerpt.


To my ears, as we move from zero limiting (the CD source) through two stages of limiting, the sound becomes '"fatter", "warmer", "easier on the ears", "analogue-like" and when the CD source returns in the forth excerpt, the sound is so much "brighter", "sharper", "harder", "fatiguing", "not so easy on the ears", "irritating" ... "digital".

How does it sound to you?
 

anonymous

New member
The "analogue sound"?

The "analogue sound"?

Ummm. This is interesting. It's a pity nobody has addressed by question here (the other thread) because to my ears, Mark's audio examples of current remastering comprehensively explain the "digital" v. "analogue" sound which is an essential to comprehending the example he has posted in this thread. Same old observation though: not enough feedback for us to adapt our presentation to help comprehension. We just cannot be expected to create this stuff late into the night and then answer our own questions: that would be kinky.This explanation about (non-loudspeaker issues) is not in any sense core to our business.

It suggests one of two things: either without training the ear there will be a wide range of possibly random opinions about sound or that we truly do need to abut audio event A and B with no gap to give ears a chance to make sense of what we are hearing. I guess Mark and I are going to have to make the effort again ...
.. But don't you think that the second clip in the example (repeated below) sounds so much "warmer" and 'nicer' and those are the very attributes which we so often read that listener's seek out in their home audio? In contrast, the opening clip sounds so much "harsher", "colder", "analytical", "thin", all summarised by the dreaded words . . . "digital sounding".

If you acclimatise to the softer second example in the video above, you would almost certainly reject the opening example as being unnatural.

Hence, at a stroke, the explanation of the "digital" v. "analogue" sound?
This is interesting. I actually consider the much maligned remastering more "analogue" sounding, but was wondering if it had to do with some limitation of dynamic range or frequency response. Looking at the remastered waveforms, it was reminiscent of things I'd read about the process of cutting vinyl records, and the limitations inherent in the medium.
 

A.S.

Administrator
Staff member
A digital world ... not for all

A digital world ... not for all

This is interesting. I actually consider the much maligned remastering more "analogue" sounding, but was wondering if it had to do with some limitation of dynamic range or frequency response. Looking at the remastered waveforms, it was reminiscent of things I'd read about the process of cutting vinyl records, and the limitations inherent in the medium...
Thanks for the response.

It's a pity I had to show the screen cam of the audio waveform, because if the audiophile ear is, as we read, so finely tuned, it is completely unnecessary to see the wave. But we have to move the discussion on.

Does it actually matter what the technical reason (limiting, compression, distortion, frequency response change etc. etc.) behind the "analogue" sound is? There is, obviously, a comprehensive technical appraisal of the potentialities of analogue v. digital recording and delivery. Despite the use of digital recording by all classical labels I'm aware of, and excepting some who just have to be different for marketing reasons, all pop studios, plus all TV and radio broadcasting, the public telephone and mobile networks, the internet... a certain demography of listener finds "the analogue sound" very appealing. In a generation, that will be different, but that's how things are now.

The suggestion is that it should be possible to take any digital recording and "analoguise" it, to make it palatable for the analogue-sound enthusiast. We can already see some of the steps involved involving signal and dynamics compression.

P.S. To keep this thread tightly on track, could you please create a fresh post "Training for the ear" and re-post it. We'll then edit that subject and your links out of your post #56. Thanks.
 

anonymous

New member
Analogue - a preference, not a solution

Analogue - a preference, not a solution

Does it actually matter what the technical reason (limiting, compression, distortion, frequency response change etc. etc.) behind the "analogue" sound is? There is, obviously, a comprehensive technical appraisal of the potentialities of analogue v. digital recording and delivery. Despite the use of digital recording by all classical labels I'm aware of, and excepting some who just have to be different for marketing reasons, all pop studios, plus all TV and radio broadcasting, the public telephone and mobile networks, the internet... a certain demography of listener finds "the analogue sound" very appealing. In a generation, that will be different, but that's how things are now.

The suggestion is that it should be possible to take any digital recording and "analoguise" it, to make it palatable for the analogue-sound enthusiast. We can already see some of the steps involved involving signal and dynamics compression.
I believe that there are numerous plugins for recording engineers to shape the sound in the digital domain, obviating the need for expensive hardware of limited use. I remember hearing someone talking about some expensive tape machine that cost many thousands of dollars. I looked and found that there was a software plug in that. I can't speak to the difference between the original machine and the plug in, but given the ear, I can't imagine it'd be too bad. It was this:

http://www.uaudio.com/store/special-processing/studer-a800-tape-recorder.html

Hopefully in a generation, analogue-sound enthusiasts will be able to identify their preference as just that...a preference. Their preferences are not absolute standards, by any stretch of the imagination or any technical standard. Then they can enjoy their analogue sound and leave the rest of the world to enjoy digital, without any hectoring.
 

A.S.

Administrator
Staff member
The sound of real life

The sound of real life

... It suggests one of three things: either without training the ear there will be a wide range of possibly random opinions about sound (non necessarily aligned with the objective facts), that we truly do need to abut audio event A and B to give ears a chance to make sense of what we are hearing, or that listener preference is counter-intuitive. Without feedback for the various audio clips, Mark and I are a bit lost as to how to proceed.

I've taken Mark's audio clip from his post #49 above (read for more info) and made a screen cam. I've added a forth excerpt - the opening ex-CD excerpt has been added on to the end, so the clip opens and closes with the raw ex-CD excerpt....
It's interesting: we have been playing this clip (post #55) in the office over my plastic PC speakers (Trust brand, pretty good actually for $30 including subwoofer (!) when stuffed with foam) and I have some more feedback. Opinions, feedback are crucial to making sense of what we hear and sharing experiences. In audio, there are no rights and wrongs, no black and white, no absolutes: everything is in the middle, the grey zone.

My colleague listened to the above and thought that the compressed excerpts sounded "softer, almost muffled, as if someone had placed a cloth over the front of the speakers.... even more so on the third, more limited example... also the level limited examples sounded a little quieter ...".

I then asked him if he had ever heard a snare drum in real life. He did, once in a noisy club over a PA system, which he found too hard on the ear. In fact, if you do hear a real drum set, unamplified, the transients are really astonishingly loud. Even the ex-CD excerpts bear no real relationship to the live sound which has a far greater dynamic range, but would be uncomfortable and would not blend into the music unless very carefully level adjusted - the art of the sound recording engineer.

So starting from the position that what we consider to be, at home, the 'marvellous true to like 16 bit/44k sound' it is just a poor shadow of the dynamics of reality: the best we can do for artistic (mainly) and technical reasons (secondary). So when I hear even less dynamic range than the CD excerpt, the -12 and -18dBFS excerpts, it moves me further and further from what I would have expected to hear live in the studio.

The primary requirements for a high fidelity recording, distribution and replay system must be wide dynamic range, adequate frequency range and adequately low distortion. What the physical system looks like is irrelevant. If the nature of the recording, the mastering, the distribution medium, the replay equipment in any way significantly restricts the dynamics, the frequency response and the freedom from distortion that a healthy ear would hear at the live event, its right to call itself a 'high fidelity system' must be challenged.

Since music is an art form, the listener is entitled to prefer the rendition of sound over one system to another. But to claim that a system that is of intentional or technically unavoidable restriction (compared with what you would hear standing at the microphones) in dynamics, frequency range or low distortion is of superior fidelity is, logically, a nonsense. But as we have seen, listeners to seem to like the softer sound.
 

weaver

New member
Was clear - getting confused

Was clear - getting confused

Firstly - I disagree with Alan as far as his "digital vs analogue sounding" summing up goes - that's not the way I hear it here at all.

Secondly - I find myself confused at present and will try to think it through during the course of the day but so far the sources of that confusion seem to be:

Alan has demonstrated on a number of occasions how boosting a particular frequency range on a speaker can give an impression of liveliness, can make the sound appear exciting and that for the purposes of a quick A-B dem at a dealers we may be drawn to the more exciting sounding speaker. The current clipping discussion is to do with reducing the amplitude of the most energetic parts of a passage of music [I'm not clear whether the current samples also raise the less energetic parts such that all parts fall within a certain range or whether it is purely to do with limiting the louder parts].

To my ears, on my office system [passive pre/English solid state power/1970s-BBC type big box speakers] the passage from uncompressed to compressed bears as as many similarities to a boosted-presence-range to flat-response speaker simulation as it does to a digital to analogue one.

Alan has often stated that there is very little energy in the upper frequencies of a music signal, that the majority of the 'sound' is actually lower down; if there is so little energy in the upper frequencies then why does reducing the dynamic range appear to have such an effect on the upper frequency range?

To return to Alan's statement (#55) that the sound when compressed becomes "fatter", "warmer", "easier on the ears" - to me, here, it actually sounds emaciated, cold and dull.

Lastly to come back to Mark at #49 - I think I might have a problem with the word distortion. To me distortion (in audio) suggests something actively altered, not necessarily unpleasant (eg. Mike Oldfield's two slightly distorted guitars).

To give a visual example: I would not think of a black and white version of a colour photograph as distorted, though in the sense that it is not the original it clearly is distorted, whilst a photoshopped mosaic effect of the same image I would think of as distorted even though the colour content may have remained the same. The compressed versions of the music files do also appear to have had the colour drained out of them, but to the extent that they haven't had anything added I wouldn't immediately hear that as distortion. Going back to the sine wave in #16 it does sound there as if the clipping distortion has added to the sound.

Having written this down I'm starting to think that the illustrations of clipping are fine but as soon as we try to associate them with external, individual experiences then confusion creeps in. Clearly what we are all hearing on our own computers and speakers isn't what either Mark or Alan are hearing on theirs, but that needn't matter as long as we are simply comparing sample A with sample B, uncompressed with compressed. But as soon as Alan says (to paraphrase) doesn't this sound like your record player and doesn't this sound like your CD player (granted, that's isn't what he actually said but it is certainly one way of reading it) then some of us will disagree and (as is the case from my point of view) be confused as to why he would even suggest it.

Conversely when Alan presents a sample of music with a flat frequency response then boosts certain areas and contrasts it with that, I can hear what the samples are doing relative to each other.
 
Top