This picture show the music waveform taken at the line level
output of a Marrantz CD5000 CD player. At the instant when the
shot was taken, the music is going through some loud passages.
The vertical display of the scope trace has been adjusted so that
only 30% of the waveform is shown.
I dont know if this clipping is due to jitter or other digital
anomalies, but my guess is that it could be in the 1's and 0's
that comes from the CD, ie it was recorded that way.
Would the squared off waveforms explain why sometimes when you listen to music and it gets to a loud passage, you ears
can't stand the music and you instintively want to turn it down.
So the fact that you want to turn it down is not because it
is loud but because the playback music was distorted during
the loud passages. This distortion take place at the source
so even if your amp is not clipping, it is amplifying a clipped
waveform. This is a relatively new recording (done in the last
couple years or so) I assume using better recording technology etc.. because the CD sounds better/clearer than others that are 10 -20 years old. But
it seems that there are still room for improvement at the recording
studios. Granted, I only scoped this one CD, and may be other
better recorded CD dont have this problem. And since only
one or two wave forms out of hundreds (in that instant of time) are clipping the resultted music sounded good, with just a hint
of harsheness. So when you are listening to your CDs and encounter some harshness, especially when it gets loud, then
before you start ripping your amp and preamp and speakers system apart trying to tweak it,
put a scope on the CD output, the problem may be in the music!
output of a Marrantz CD5000 CD player. At the instant when the
shot was taken, the music is going through some loud passages.
The vertical display of the scope trace has been adjusted so that
only 30% of the waveform is shown.
I dont know if this clipping is due to jitter or other digital
anomalies, but my guess is that it could be in the 1's and 0's
that comes from the CD, ie it was recorded that way.
Would the squared off waveforms explain why sometimes when you listen to music and it gets to a loud passage, you ears
can't stand the music and you instintively want to turn it down.
So the fact that you want to turn it down is not because it
is loud but because the playback music was distorted during
the loud passages. This distortion take place at the source
so even if your amp is not clipping, it is amplifying a clipped
waveform. This is a relatively new recording (done in the last
couple years or so) I assume using better recording technology etc.. because the CD sounds better/clearer than others that are 10 -20 years old. But
it seems that there are still room for improvement at the recording
studios. Granted, I only scoped this one CD, and may be other
better recorded CD dont have this problem. And since only
one or two wave forms out of hundreds (in that instant of time) are clipping the resultted music sounded good, with just a hint
of harsheness. So when you are listening to your CDs and encounter some harshness, especially when it gets loud, then
before you start ripping your amp and preamp and speakers system apart trying to tweak it,
put a scope on the CD output, the problem may be in the music!
Attachments
I'd like to know what scan rate your scope was set to. The waveform shouldn't look like that in the first place. Even if it was clippin, that doesn't even slightly resemble what we are looking for.
Also, the input on your scope might have not been attenuated enough. I'd imagine that the output on your marantz is fine.
Check over the scope settings.🙂
Also, the input on your scope might have not been attenuated enough. I'd imagine that the output on your marantz is fine.
Check over the scope settings.🙂
scan rate
was set to 0.2 ms per division. The clipping waveform in the top picture
is about 1 division wide so that makes it about 5000 Hz.
The bottom one is about 3 division so that makes it about
1.6 kHz
was set to 0.2 ms per division. The clipping waveform in the top picture
is about 1 division wide so that makes it about 5000 Hz.
The bottom one is about 3 division so that makes it about
1.6 kHz
Clipping?
Which part of your picture is clipping ? I can't see what we normally know as clipping.
Unless your CD player is defective I think you should never be able to clip any of the analog components. With about +/- 12 volts ( typically ) and about 2 volts at 0 db output , you could never clip it with any signal from the CD. And clipping off the CD data stream ? - I have never heard of it !
Clipped individual instruments - I have never seen this but that might happen at the recording stage , not while transfer to the digital domain.
In any case I can't see any clipped waveform on your screen. You should use a slower scan rate .
Cheers.
Which part of your picture is clipping ? I can't see what we normally know as clipping.
Unless your CD player is defective I think you should never be able to clip any of the analog components. With about +/- 12 volts ( typically ) and about 2 volts at 0 db output , you could never clip it with any signal from the CD. And clipping off the CD data stream ? - I have never heard of it !
Clipped individual instruments - I have never seen this but that might happen at the recording stage , not while transfer to the digital domain.
In any case I can't see any clipped waveform on your screen. You should use a slower scan rate .
Cheers.
I reccon.. I mean... what are we looking at? 🙁 shouldn't you burn a sign wave sweep onto a CD and play that?
It is quite possible that the digital signal is "clippling" as any ADC with over-driven inputs will create a stream of 1's.
Anyone have an idea how to determine if the CD itself contains a stream of 1's? I'm thinking that there must be software that can show the binary stream as audio "plays" off a disc.
🙂ensen.
Anyone have an idea how to determine if the CD itself contains a stream of 1's? I'm thinking that there must be software that can show the binary stream as audio "plays" off a disc.
🙂ensen.
For me the image taken by the scope is no surprise . I read about that a good article in Elektor 2 years ago or so .
The maker of the article did same experiments with a scope and took some of the album from Gloria Estefan passages .
It showed clearly the same waveforms , and the author experimented same thing on some of Jean Michel Jarre CD recordings .
So where is the fact ? There are many places the clipping can have been made . Sometimes it starts at very beginning of recording for each track , in case of a multi track recording , but mostly the evil is done later on into the recording process .
The worst stade is at mastering sessions , before the engineerer sends the final tape to be mastered . The use of very sofisticated technology is never the assurance of a clean result , and every waveform is not analized so not to ensure lack of that problem .
In fact , there are some tech rules for them saying such clipping is not HEARABLE if shorter than x milliseconds . That is not true of course but every mastering engineerer wants to maximize impact and high output levels so that sound is huge on the CD , very loud though . So , the resulting tape is a compromise . Here it could not result of analogue tape distortion but clearly indicates that the max input level in digital has been exceeded in the upper parts of the waveforms , in order to maximise output level . This is good for mass audio production , where some music destination seems not an audiophile system but popular mini-systems , where that problem would not be heard as such , regarding the indigence of dynamics,bandwith,transcient response and phase of such systems ( think about cafes, department stores, clothes shops etc etc . )
Of course , all that is really nasty for us , poor audiophiles , who take extra care and labour to really get clean results with expensive equipement - or extra diy labour ,
The maker of the article did same experiments with a scope and took some of the album from Gloria Estefan passages .
It showed clearly the same waveforms , and the author experimented same thing on some of Jean Michel Jarre CD recordings .
So where is the fact ? There are many places the clipping can have been made . Sometimes it starts at very beginning of recording for each track , in case of a multi track recording , but mostly the evil is done later on into the recording process .
The worst stade is at mastering sessions , before the engineerer sends the final tape to be mastered . The use of very sofisticated technology is never the assurance of a clean result , and every waveform is not analized so not to ensure lack of that problem .
In fact , there are some tech rules for them saying such clipping is not HEARABLE if shorter than x milliseconds . That is not true of course but every mastering engineerer wants to maximize impact and high output levels so that sound is huge on the CD , very loud though . So , the resulting tape is a compromise . Here it could not result of analogue tape distortion but clearly indicates that the max input level in digital has been exceeded in the upper parts of the waveforms , in order to maximise output level . This is good for mass audio production , where some music destination seems not an audiophile system but popular mini-systems , where that problem would not be heard as such , regarding the indigence of dynamics,bandwith,transcient response and phase of such systems ( think about cafes, department stores, clothes shops etc etc . )
Of course , all that is really nasty for us , poor audiophiles , who take extra care and labour to really get clean results with expensive equipement - or extra diy labour ,
first thing encountered after a microphone is often/always a compressor/limiter.
So.... what do you expect?
A horn system (genre klipsch, avantgarde) can 'restore' a bit of the leftover dynamics, only because most modern studiomonitors (mostly B&W) sound a bit 'lazy' (in comparision) by nature.
There are always some signs of 'life' left, even in the greatfull death.
BTW marketing of bass boost systems and other spatializers, dsp-programs need a cause, don't you think.
A healty patient is a bad doctor's client.
So.... what do you expect?
A horn system (genre klipsch, avantgarde) can 'restore' a bit of the leftover dynamics, only because most modern studiomonitors (mostly B&W) sound a bit 'lazy' (in comparision) by nature.
There are always some signs of 'life' left, even in the greatfull death.
BTW marketing of bass boost systems and other spatializers, dsp-programs need a cause, don't you think.
A healty patient is a bad doctor's client.
Hi
Just a idea but I did this:
made a audio CD with some test-tracks with 'goldwave' software to test my DACs.... some -90dB, -60dB, -20dB and 0dB tones, some squares (really square), etc etc ... The 0dB tone reveiled that the CD-rom (analog out this time) clipped assymmetrical... this was not the case with my DACs or CD-players... so it must have been the output stage of the CD-rom...
Maybe you could do the same thing to see what is really happening...
Regards,
Thijs
Just a idea but I did this:
made a audio CD with some test-tracks with 'goldwave' software to test my DACs.... some -90dB, -60dB, -20dB and 0dB tones, some squares (really square), etc etc ... The 0dB tone reveiled that the CD-rom (analog out this time) clipped assymmetrical... this was not the case with my DACs or CD-players... so it must have been the output stage of the CD-rom...
Maybe you could do the same thing to see what is really happening...
Regards,
Thijs
Thanks all for your replies. For those who could not see the
clipping waveform, I appologize for the fuzzy pictures. The pictures
I took was high res, but the forum software limits the size of the
of the image file so I had to resize it and losing resolution in
the process. The clipped waveform I am referring to in the first
picture is the one in the middle, the tallest waveform has a
flat top instead of rounded. The one in the 2nd picture is
also in the middle and extended all the way to near the top
of the picture , it also has somewhat wavy flat top.
clipping waveform, I appologize for the fuzzy pictures. The pictures
I took was high res, but the forum software limits the size of the
of the image file so I had to resize it and losing resolution in
the process. The clipped waveform I am referring to in the first
picture is the one in the middle, the tallest waveform has a
flat top instead of rounded. The one in the 2nd picture is
also in the middle and extended all the way to near the top
of the picture , it also has somewhat wavy flat top.
Check out the Jayhawks - Rainy Day Music cd. Nearly the whole cd is recorded well into clipping. Ethan Johns should know better!
Nothing unusual here. I have analyzed some discs using Cool Edit and found literally *thousands* of instances of digital clipping on one three minute track! Also some CDs are compressed so severely that the average signal level is -4db relative to max signal!
This is a result of the current "louder is better" view of many artists and record companies. Mastering engineers who refuse to do this sort of butchery will soon find themselves with no work.
This is a result of the current "louder is better" view of many artists and record companies. Mastering engineers who refuse to do this sort of butchery will soon find themselves with no work.
Mission impossible
So, now what? Do we have to figure out how to make speakers that can deliver DC?
🙂ensen.
So, now what? Do we have to figure out how to make speakers that can deliver DC?
🙂ensen.
I did the same test with Billy Corgan's CD of his new group 'Zwan' ..
I suspected him to manipulate the hell out of it sound-production wize .... but behold: the 4th track does reach the highest or second highest sample-value many times, but no longer than one or a couple of samples long.. I wouldn't wory too much ... how about a classical music recording.....
so no DC-reproduction needed yet 😉
Regards,
Thijs
PS
when I practise guitar, on overdriven tube amp offcourse, my dynamic range probably is less that 4 dB.....
I suspected him to manipulate the hell out of it sound-production wize .... but behold: the 4th track does reach the highest or second highest sample-value many times, but no longer than one or a couple of samples long.. I wouldn't wory too much ... how about a classical music recording.....
so no DC-reproduction needed yet 😉
Regards,
Thijs
PS
when I practise guitar, on overdriven tube amp offcourse, my dynamic range probably is less that 4 dB.....
tschrama said:
The release in question did not have an overdriven guitar in it. It wasn't head banging metal or loud rap. It should have had a much more normal dynamic range.
It's interesting to see that others also have experienced the problems with digital clipping... About a year ago I upgraded from some cheap consumer stuff to DIY tri-amped speakers & amp. I was really happy with my new system, but after a while I noticed that there seemed to be a problem, i e bad electrical contact, to my tweeters, since certain music produced harsh "chirp" sounds. What I found out many troubleshooting hours later was that this is the distinct sound of digital clipping.
Recently I bought a decent soundcard for my PC, RME DIGI 96, and with this card comes some pretty amazing piece of software called "Digicheck". With this software I can now view peak, RMS, crest values et al for a digital source, and also look at something called a "Goniometer" which makes it very easy to see which recordings that are bad; the recordings that make my tweeters sound awful are the same that show clear signs of digital clipping.
Although there are exceptions, I'd say that when it comes to mainstream music, the newer the recording, the worse the sound. I've seen examples where the maximum level was -0.1 db, but this was nothing else that clipping at 0.1 db below the CD:s max limit, probably some lousy engineer had some equipment limiting the max amplitude to -0.1 db and thought that by using it he would never run in to clipping problems.
Unfortunately, I don't know of any way to figure out the sound quality of a recording other that buying it, and I think that trying to return a CD to a store claiming digital clipping as the defect, would just make the salesperson think that I was the defect one...
BTW: My tweeters are Vifa D25AG-35-06's. The recording that
clips at -0.1 db is "The Brand New Heavies: Shelter".
Recently I bought a decent soundcard for my PC, RME DIGI 96, and with this card comes some pretty amazing piece of software called "Digicheck". With this software I can now view peak, RMS, crest values et al for a digital source, and also look at something called a "Goniometer" which makes it very easy to see which recordings that are bad; the recordings that make my tweeters sound awful are the same that show clear signs of digital clipping.
Although there are exceptions, I'd say that when it comes to mainstream music, the newer the recording, the worse the sound. I've seen examples where the maximum level was -0.1 db, but this was nothing else that clipping at 0.1 db below the CD:s max limit, probably some lousy engineer had some equipment limiting the max amplitude to -0.1 db and thought that by using it he would never run in to clipping problems.
Unfortunately, I don't know of any way to figure out the sound quality of a recording other that buying it, and I think that trying to return a CD to a store claiming digital clipping as the defect, would just make the salesperson think that I was the defect one...
BTW: My tweeters are Vifa D25AG-35-06's. The recording that
clips at -0.1 db is "The Brand New Heavies: Shelter".
- Status
- Not open for further replies.