View Full Version : Ouch - too loud
I recently acquired John Waite's 'Ignition' which *some* years ago was loaded in the cassette deck of my 2 litre Ford Escort - slightly modified - and belting out often. Some great tracks ('Change' being my favourite from this album) from this performer. I have to say that the CD - remastered - is a proud possession, but it is far too loud. Looking at the file in a wave editor clearly shows peaks 'flattened out', meaning clipping. Where do these engineers get off? I have read about this phenomenon, but only now have I been a victim of it. My chagrin came from what I refer to by reference to others in this forum who have named it, as 'listening fatigue'. Something made me wince occasionally when listening and I had to lower the volume, but that didn't really help. Bummer. A good album not destroyed so much, but damaged. Anybody else had a similar experience?
Clipping is actually exceedingly rare, since it adds horrible distortion (exactly what a fuzz-box does for an electric guitar).
The process is actually one of compression, whereby the dynamic range is reduced, effectively making the quiet bits louder, and then mastering the CD at maximun volume. It gives a very loud, very flat waveform, with no subtlety, and gets pretty boring after a while...
Mark Lanctot
2007-03-09, 08:49
This is very well-known and has been a problem on CDs for going on 10 years now.
http://en.wikipedia.org/wiki/Loudness_War
Everyone wants to make their CD sound louder than the other guy. When they reached the limit of the CD format, they decreased dynamics even further and increased compression.
I'm surprised that Khuli says it's rare - perhaps he's referring to analog clipping? What's happening here is "digital clipping" - take a look at the waveforms in Audacity, they hit 1.00 or -1.00 and stay there for many samples. This leads to a staticky, buzzy, distorted sound.
I first noticed this with Red Hot Chili Peppers - Californication which is a notorious example.
This is very common now on just about every CD made. If you use ReplayGain, take a look at some of the extreme values it's generating for modern CDs: -10, -11 dB. Crazy, and unfortunately it's recorded right on the CD, so there's no fix. Let's hope the masters aren't clipped like that or that music data is gone forever.
Anybody else had a similar experience?
There's a really good thread here where someone actually got a refund for defective goods due to this. Had that been a large label rather than BMG, perhaps it could be considered justification for initiating a class action suit (?). Surely the labels, in retaliation, would somehow make it completely illegal to rip a cd to HDD.
DON'T ACCEPT DEFECTIVE CDs, by Mark G Woodruff.
Exclusively available at:
http://forums.slimdevices.com/showthread.php?t=22115
Clipping is actually exceedingly rare, since it adds horrible distortion (exactly what a fuzz-box does for an electric guitar).
The process is actually one of compression, whereby the dynamic range is reduced, effectively making the quiet bits louder, and then mastering the CD at maximun volume. It gives a very loud, very flat waveform, with no subtlety, and gets pretty boring after a while...
As others have noted, clipping is not rare at all. The loudness available through straightforward compression was exhausted some years ago, and the only way to get even louder is to resort to digital clipping. This is precisely what has been going on in mainstream rock and pop releases for several years. It's now the norm rather than the exception to find rock and pop albums absolutely swimming in clipping. Typically there are *several thousand* instances of clipping *per track*.
nicketynick
2007-03-09, 10:19
Typically there are *several thousand* instances of clipping *per track*.
Seriously? Wouldn't several thousand imply that every single 'attack' (I may not have the terminology right here) is driven into the realm of digital clipping? Or even worse, are they compressing to the point that sustained sounds are clipping?
This is ridiculous - it certainly isn't sustainable - we can only hope that sanity will prevail.
ModelCitizen
2007-03-09, 10:22
Clipping......
The process is actually one of compression, whereby the dynamic range is reduced, effectively making the quiet bits louder, and then mastering the CD at maximun volume. It gives a very loud, very flat waveform, with no subtlety, and gets pretty boring after a while...
Some electronic music I love depends on huge amounts of compression. The compression is part of the music, almost an instrument. This music tends to be good for driving very fast and dancing wildly.
MC
Seriously? Wouldn't several thousand imply that every single 'attack' (I may not have the terminology right here) is driven into the realm of digital clipping?
Yes that's exactly what happens and result is usually an album that sounds bland and boring, and usually leaves me thinking 'I can't be arsed listening to this'
Craig
Mark Lanctot
2007-03-10, 14:30
With regards to the "digital clipping", what I hear sounds a lot like analog clipping.
A fuzziness/distortion in the peaks. No sharp popping like with analog clipping, but a staticy sound.
The "bland and boring" sound would be more closely related to the lack of dynamic range and high compression I'd think, not clipping per se.
This music tends to be good for driving very fast and dancing wildly.
MC
Not both at the same time, I hope....
Ceejay
With regards to the "digital clipping", what I hear sounds a lot like analog clipping.
A fuzziness/distortion in the peaks. No sharp popping like with analog clipping, but a staticy sound.
The "bland and boring" sound would be more closely related to the lack of dynamic range and high compression I'd think, not clipping per se.
Yes it's the lack of dynamics that causes the blandness, but it's likely the clipping is caused by overdoing the compression/limiting rather than trying to write +10db to a cd.
Here's a pic of a track picked at random.
Craig
Seriously? Wouldn't several thousand imply that every single 'attack' (I may not have the terminology right here) is driven into the realm of digital clipping? Or even worse, are they compressing to the point that sustained sounds are clipping?
This is ridiculous - it certainly isn't sustainable - we can only hope that sanity will prevail.
I'm serious. And you're right - it is ridiculous.
Let's take a typical example, pretty much at random. The track "Is It Any Wonder" from Keane's album "Under the Iron Sea". This track has 3518 separate flat-topped waves. I've attached a couple of screen shots. The first shows the general shape of the waveform. The second has markers added at the places where clipping occurs. (The different colours used in the markers indiciate the severity of clipping: green for clips up to 6 samples wide, blue for 7-10 samples wide, yellow for 11-14 samples wide, and red for 15 or greater. On this particular track there are 13 occurrences of clips that are 24 or more samples wide).
For the conspiracy theorists among us, note also that the clipping is at -0.3dB, not 0dB. It's almost as if they've deliberately perpetrated the clipping, then backed off the peak level so as to avoid casual detection.
Great minds think alike CliveB :-)
The radio stations are partly to blame for all this, they compress the hell of of the music to be the loudest station on the dial. People hear it and think I want my music to sound like it does on the radio so they compress the hell out of it. Then it gets played on the radio and they compress it again, it's a vicious circle.
Craig
For the conspiracy theorists among us, note also that the clipping is at -0.3dB, not 0dB. It's almost as if they've deliberately perpetrated the clipping, then backed off the peak level so as to avoid casual detection.
Aside from basically eliminating dynamics and killing what might otherwise be decent albums, what is this doing to our audio equipment?
Eric Seaberg
2007-03-11, 08:02
Great minds think alike CliveB :-)
The radio stations are partly to blame for all this, they compress the hell of of the music to be the loudest station on the dial. People hear it and think I want my music to sound like it does on the radio so they compress the hell out of it. Then it gets played on the radio and they compress it again, it's a vicious circle.
Craig
Actually, in his book on MASTERING, Bob Katz talks about a test that was done with 6 different 'mastered' versions of the same song sent through an Orban Optimod, the typical processor for radio transmitters. The 6-mastered mixes all had varying levels of dynamics, based on how the mastering engineer "heard" it, going from good dynamic range to not at all... much like your waveform above.
By the time all 6 mixes went through the Optimod, their dynamics level was IDENTICAL... almost NONE!! This proves that a CD can be mastered with a good, natural dynamics range and STILL pop out of the radio after going through the Optimod at the transmitter.
It was also proven that the mix with the least amount of dynamics sounded WORSE after the Optimod... another point for putting dynamics back into the music.
Tests have also proven that women are VERY sensitive to the added distortion and will, usually, change stations when it increases.
By the way, Bob is now a member of this user's group and, I'm sure, would love to give a little more detail if asked.
I've been in a commercial radio transmitter site and seen an Optimod reducing peaks by 18dB and that was after whatever was done in the studio.
Craig
Technically, I suppose we should call it "pseudo-clipping," or something like that, because if you really zoom in on the flattened waveforms, you'll see that have slightly rounded corners, which alleviates the infinite harmonics that would result in really grating clipping distortion.
But it still sucks.
Furthermore, if a large percentage of the "buying public" is willing to download their friends' MP3s for free or buy some flattened crap from iTunes rather than pay for a CD, what message does that send to the labels to put the effort into decent production?
Eric Seaberg
2007-03-12, 06:09
It's not really clipped on the CD, but is held at around -.2dBFS (2/10 of a dB below FULL SCALE). This is pretty typical of a CD mastering job in present day. Usually a brick-wall limiter is placed in the processing chain to keep it there. Consider it more like 'controlled' distortion!
BTW, it has been proven that distortion has the psychoacoustic effect of seeming louder to the listener... which is really what's happening here. You can't go any louder than the digital ceiling, yet the slight 'controlled' distortion makes it 'feel' louder.
It's not really clipped on the CD, but is held at around -.2dBFS (2/10 of a dB below FULL SCALE). This is pretty typical of a CD mastering job in present day. Usually a brick-wall limiter is placed in the processing chain to keep it there. Consider it more like 'controlled' distortion!
Hard digital limiting is exactly the same as clipping. Just because they've chosen to limit at -0.2dB instead of 0dB makes no difference to the fact that the wavetops are chopped off flat (ie. "clipped" - that's the origin of the word).
BTW, it has been proven that distortion has the psychoacoustic effect of seeming louder to the listener... which is really what's happening here. You can't go any louder than the digital ceiling, yet the slight 'controlled' distortion makes it 'feel' louder.
While it is true that distortion makes things sound louder, that is not the major factor at work here. The whole purpose of the clipping/limiting is to restrict the peak levels so that the average levels can be increased. It's this ability to increase the levels that contributes mostly to the increased loudness. The additional psychoacoustic effect of the distortion is minor in comparison.
Furthermore, if a large percentage of the "buying public" is willing to download their friends' MP3s for free or buy some flattened crap from iTunes rather than pay for a CD, what message does that send to the labels to put the effort into decent production?
You forgot to mention walking down the street listening to music on their loudspeaking phones!!
Maybe we're just old farts ;-)
Craig
Some electronic music I love depends on huge amounts of compression. The compression is part of the music, almost an instrument. This music tends to be good for driving very fast and dancing wildly.
MC
I'm guessing that anything which relies upon this sort of treatment to sound improved in any way, apart from louder, is up against it in the first place, notwithstanding the fact that I do like a lot of electronic music.....
ModelCitizen
2007-03-12, 23:53
I'm guessing that anything which relies upon this sort of treatment to sound improved in any way, apart from louder, is up against it in the first place, notwithstanding the fact that I do like a lot of electronic music.....
No, it's just a tool, one amongst many types of sound processors in an electronic musicians's effects rack, used to gain a certain quality to the sound.
To me, and obviously I'm not totally alone, compression can be used with great affect.
MC
Eric Seaberg
2007-03-13, 07:28
Hard digital limiting is exactly the same as clipping. .....
I've been in the mastering business since the late 70s. What I was trying to say is that there is a sound difference, which you obviously understand, too. Don't start gettin' technical on me, now.
I've been in the mastering business since the late 70s. What I was trying to say is that there is a sound difference, which you obviously understand, too. Don't start gettin' technical on me, now.
The only conceivable differences between hard limiting at less than 0dB and clipping at 0dB that I can think of are: (i) the limiter may operate in such a way that the chopped off wavetops aren't absolutely ruler-flat; (ii) if you limit at less than 0dB, you might actually make the flat-tops a little wider than had you let them clip at 0dB. I think the extremely subtle differences between these two cases pales into insignificance next to the gross distortion introduced by either mechanism.
By the way, since you're in the mastering business, perhaps you could take your colleagues to one side and give them a good slapping over this issue. And don't let them plead the "I was only following orders" defense - that doesn't work for war criminals, and nor should it apply to the much more heinous crime of ruining music.
I'm not a recording engineer. I don't even play one on TV. But just thinking about it from the standpoint of physics, the only thing the flat part of the waveform represents is silence. Voltage has pushed the speaker cone out to some point and it is held there for the duration of the flat portion of the wave.
Distortion is introduced at the corners of the flat portion. If the signal is really clipped, then there will be very sharp corners on the ends of the flat portion. As the speakers attempt to negotiate those corners, much hideous distortion will ensue because you're asking the cone to come to an immediate halt, which it can't because it has mass and inertia.
I believe the limiters employed by these dastardly recording/mastering engineers do not hard clip, however. They slightly round the corners on the edge of the flat portion of the waveform, which your reproduction equipment can negotiate with less distortion.
At the same time, I'm not saying it doesn't sound worse than a properly mastered recording.
To me, and obviously I'm not totally alone, compression can be used with great affect.
MC
I'm in that camp. Phase games are a valid stereo illusion as well, though certainly not a sound people would always want to listen to.
After bluedragon asked: 'how do you know the artist didn't do it intentionally?' in the last thread, I've kept an eye out for intentional flat topping. I haven't seen an intentionally clipped wave in an otherwise sanely leveled track, so that would be interesting if you find one. It seems like it's usually an all out assault, and the whole album clips, or none. I could be wrong.
EDIT: removed image because it was a dB waveform. Doh!
Eric Seaberg
2007-03-13, 11:15
I believe the limiters employed by these dastardly recording/mastering engineers do not hard clip, however. They slightly round the corners on the edge of the flat portion of the waveform, which your reproduction equipment can negotiate with less distortion.
Yes, commonly called "soft-limiting".
I'm not a recording engineer. I don't even play one on TV. But just thinking about it from the standpoint of physics, the only thing the flat part of the waveform represents is silence. Voltage has pushed the speaker cone out to some point and it is held there for the duration of the flat portion of the wave.
Well, I suppose you could define it as silence, for a very short period of time, and with an extreme DC offset. Another more traditional way of looking at it is that it's akin to a square wave, which of course is an aggregation of lots of sine waves at various harmonics. In other words, it's extreme distortion.
I believe the limiters employed by these dastardly recording/mastering engineers do not hard clip, however. They slightly round the corners on the edge of the flat portion of the waveform, which your reproduction equipment can negotiate with less distortion.
See the attached screen shot, which is a detail of some of the clips from the same song (Is It Any Wonder by Keane) that I posted about earlier in this thread. Those wavetops are simply chopped straight off. There is no subtlety going on here; it's pure audio vandalism.
As it happens, because the song has previously been subjected to massive compression, the wavetops that get chopped off are not far from flat in the first place, so the "corners" of the clipping are not as severe as they might otherwise have been. Is this an excuse? ("But officer, some of the windows were already broken, so I thought it was quite alright to smash the rest").
There's still a big difference between a square wave and these chopped-off wave forms. A pure square wave asks a speaker cone to move from a contracted resting position to an extended resting position instantaneously, which is of course, physically impossible.
At least these chopped-off wave forms ramp.
But don't get me wrong! I'd like to strangle the little monkeys that master albums like this!!
There's still a big difference between a square wave and these chopped-off wave forms. A pure square wave asks a speaker cone to move from a contracted resting position to an extended resting position instantaneously, which is of course, physically impossible.
At least these chopped-off wave forms ramp.
Yes, in my example the prior hypercompression does make the flat-tops less extreme. For a more extreme example, see the screen shot posted by Craig (posting #11 in this thread).
Powered by vBulletin® Version 4.1.12 Copyright © 2013 vBulletin Solutions, Inc. All rights reserved.