Re: Re: Re: Re: Re: Re: Re: nw_avphile
Man, can you tap-dance!
First you ramble on about differences in amps quoting lots of THD figures in relation to what you can and cannot here, now you take an 180 degree saying ah, THD really doesn't tell it, its the jitter really (in CDPs). Then again, I understand that the difference is only heard with certain speakers? Wow, I hope all those amp designers out there read your posts, otherwise they may really turn out crappy amps!
Jan Didden
Pan said:
Man, can you tap-dance!
First you ramble on about differences in amps quoting lots of THD figures in relation to what you can and cannot here, now you take an 180 degree saying ah, THD really doesn't tell it, its the jitter really (in CDPs). Then again, I understand that the difference is only heard with certain speakers? Wow, I hope all those amp designers out there read your posts, otherwise they may really turn out crappy amps!
Jan Didden
"Man, can you tap-dance!"
No, I can´t tap dance.
"now you take an 180 degree saying ah,"
Absolutely not, if you think that maybe you can clarify your post?
"THD really doesn't tell it, its the jitter really (in CDPs)"
What I say and mean, is that even with distortion levels as low as in good CDP´c (like 0.001% or so) there is a clear audible difference. Simply means that simple THD numbers does not tell it all. Jitter cause a lot of crap (the way I understand it, and I do not claim to understand this completely if these show up in simple THD+N numbers) which is clearly audible.
To clarify: I believe the worst degrading distortion in a CDP/SACD is the one that comes from a clock with bad precision. With a precision clock installed the sound gets many steps better. That does not mean that if the spec says 0.001% THD+N, it´s all done and perfect. By avoiding the opamps that is in 99% in CDP´s analog sections it´s possible to better the sound still with several steps, indicating a THD+N number of 0.001% (for example) is NOT good enough depending on what kind of distortion and circuitry.
"Then again, I understand that the difference is only heard with certain speakers? "
I don´t know what you try to say with your post, but in my experience various distortions becomes much more easy to hear on high resolution speakers, do you suggest otherwise. Is that something you have a problem understanding? I have never said that this or that is only possible to hear on this or that speaker. What I have said is that if someone wants to hear the differences I am talking about, it is wise to go and listen to a high resolving speaker as crappy ones may (or may not) mask many audible differences.
"Wow, I hope all those amp designers out there read your posts, otherwise they may really turn out crappy amps!"
I don´t understand what you talk about, what are you trying to say with your post?
/Peter
No, I can´t tap dance.
"now you take an 180 degree saying ah,"
Absolutely not, if you think that maybe you can clarify your post?
"THD really doesn't tell it, its the jitter really (in CDPs)"
What I say and mean, is that even with distortion levels as low as in good CDP´c (like 0.001% or so) there is a clear audible difference. Simply means that simple THD numbers does not tell it all. Jitter cause a lot of crap (the way I understand it, and I do not claim to understand this completely if these show up in simple THD+N numbers) which is clearly audible.
To clarify: I believe the worst degrading distortion in a CDP/SACD is the one that comes from a clock with bad precision. With a precision clock installed the sound gets many steps better. That does not mean that if the spec says 0.001% THD+N, it´s all done and perfect. By avoiding the opamps that is in 99% in CDP´s analog sections it´s possible to better the sound still with several steps, indicating a THD+N number of 0.001% (for example) is NOT good enough depending on what kind of distortion and circuitry.
"Then again, I understand that the difference is only heard with certain speakers? "
I don´t know what you try to say with your post, but in my experience various distortions becomes much more easy to hear on high resolution speakers, do you suggest otherwise. Is that something you have a problem understanding? I have never said that this or that is only possible to hear on this or that speaker. What I have said is that if someone wants to hear the differences I am talking about, it is wise to go and listen to a high resolving speaker as crappy ones may (or may not) mask many audible differences.
"Wow, I hope all those amp designers out there read your posts, otherwise they may really turn out crappy amps!"
I don´t understand what you talk about, what are you trying to say with your post?
/Peter
Still In The Name Of Science.
Bratislav, Nw_avphile and Mikek,
In the name of peer review, please publish here your EXACT null-testing methods, in order that they be scrutinised and be shown to be free from experimental errors.
Until this is done, your assertions cannot be held to be true.
This is not criticism, rather it is for the purpose of ensuring that your findings are correct, or otherwise.
So, out with it.
Eric.
Bratislav, Nw_avphile and Mikek,
In the name of peer review, please publish here your EXACT null-testing methods, in order that they be scrutinised and be shown to be free from experimental errors.
Until this is done, your assertions cannot be held to be true.
This is not criticism, rather it is for the purpose of ensuring that your findings are correct, or otherwise.
So, out with it.
Eric.
Re: Still In The Name Of Science.
Sorry to be late Eric.
I have no intention to write in anyone else's name.
This is my own view of how to objectively compare distortion residuals of different amplifiers.
Pick a set of speakers that you intend to listen to. It helps if they are well matched.
Connect one channel of the first amp to one speaker.
Connect the same channel of the second amp to the other speaker. Either connect the same output from the preamp to both of the amps in test, or use mono signal.
Now both amps will play the same music on the same real life load.
Most amps will have similar gain, and no too pedantic level matching is required. In case of Krells as they have rather low gain it may be good to trim the input to the other amp in test (feel free to use megabuck resistor divider).
Play some music.
Connect channel A (or X on most oscilloscopes) to the input of the first speaker. Connect channel B (or Y) to the second speaker.
Click on "X-Y". Adjust gain on either of the channels until the signal on screen is minimized (it won't be necessary if you used aforementioned precision divider at the input of more sensitive amp). If you want to be hyper pedantic, use test signal (available on most CROs) to adjust the gain between amps to be *exactly* the same. Now all you see is the DIFFERENCES in distortion residuals between two amps in the test.
Note that this test will tell you nothing about how much distortion either of the amps produces. It will only tell you how much DIFFERENCE in distortion components is there between two amps in test.
But this is all of the information we'll ever need - as this is the all the difference between the amps there is.
Yes, I'm aware that we test only one channel of either amp this way. So what - repeat the procedure with other channel.
I'm also aware that speaker load will not be exactly the same. So what - this makes difference actually seem WORSE than in reality.
If you want to be hyper pedantic again, use laboratory quality power resistors and add a high quality percision capacitors parallel to it, or inductors, or both, whatever (simulated "real life" load).
Whatever you do, you will find that difference in distortion residuals between well designed amps operating well within their limits is MINUTE. So small in fact that if you played signal of the same magnitude you would be hard pressed to hear anything up close to the speaker.
If you consider that there will be MUCH louder signal that your ear is dealing with at the same time (our "good" component), it would take some wild imagination to beileve that you could actually hear it.
You can in fact record the difference and inject it afterwards at random while music is playing. But that would constitute another blind test, so that doesn't count I guess
Bratislav
PS when making above test one has to be very pedantic with grounding of equipment. Any ground loop, or if impedance between different channel sondes' and speakers is not exactly the same, will invidate test.
mrfeedback said:
Sorry to be late Eric.
I have no intention to write in anyone else's name.
This is my own view of how to objectively compare distortion residuals of different amplifiers.
Pick a set of speakers that you intend to listen to. It helps if they are well matched.
Connect one channel of the first amp to one speaker.
Connect the same channel of the second amp to the other speaker. Either connect the same output from the preamp to both of the amps in test, or use mono signal.
Now both amps will play the same music on the same real life load.
Most amps will have similar gain, and no too pedantic level matching is required. In case of Krells as they have rather low gain it may be good to trim the input to the other amp in test (feel free to use megabuck resistor divider).
Play some music.
Connect channel A (or X on most oscilloscopes) to the input of the first speaker. Connect channel B (or Y) to the second speaker.
Click on "X-Y". Adjust gain on either of the channels until the signal on screen is minimized (it won't be necessary if you used aforementioned precision divider at the input of more sensitive amp). If you want to be hyper pedantic, use test signal (available on most CROs) to adjust the gain between amps to be *exactly* the same. Now all you see is the DIFFERENCES in distortion residuals between two amps in the test.
Note that this test will tell you nothing about how much distortion either of the amps produces. It will only tell you how much DIFFERENCE in distortion components is there between two amps in test.
But this is all of the information we'll ever need - as this is the all the difference between the amps there is.
Yes, I'm aware that we test only one channel of either amp this way. So what - repeat the procedure with other channel.
I'm also aware that speaker load will not be exactly the same. So what - this makes difference actually seem WORSE than in reality.
If you want to be hyper pedantic again, use laboratory quality power resistors and add a high quality percision capacitors parallel to it, or inductors, or both, whatever (simulated "real life" load).
Whatever you do, you will find that difference in distortion residuals between well designed amps operating well within their limits is MINUTE. So small in fact that if you played signal of the same magnitude you would be hard pressed to hear anything up close to the speaker.
If you consider that there will be MUCH louder signal that your ear is dealing with at the same time (our "good" component), it would take some wild imagination to beileve that you could actually hear it.
You can in fact record the difference and inject it afterwards at random while music is playing. But that would constitute another blind test, so that doesn't count I guess
Bratislav
PS when making above test one has to be very pedantic with grounding of equipment. Any ground loop, or if impedance between different channel sondes' and speakers is not exactly the same, will invidate test.
Tube_Dude test...
Actualy i use a evolution of the Hafler method:
If the amp at test have a pot at the input fine...if not you must put one even temporalely...
Put 2volt sinus or music from a CD and load the amp with the loudspeakers of your choice
Put the ground of the osciloscope proof at the + of the loudspeaker and the other side of the osciloscope prove at the input of the amp before the pot...
Imagine that you are doing the proof with 2Volts sinus...
at the minimum of the volume you will see 2 volts at the osciloscope as the output is at zero
Now increase the volume...the voltage in the osciloscope begin to decrease and you will find a minimum where the output voltage is the same as the input voltage...that residual signal is the overall distortion of the amp...and this is true for all distortion typs invented or not yet invented...
So easy ...so simple...and so true
You have another explanation of this test in:
Post #8 in
http://www.diyaudio.com/forums/showthread.php?s=&threadid=13415&perpage=15&pagenumber=1
You can use also this test:
Actualy i use a evolution of the Hafler method:
If the amp at test have a pot at the input fine...if not you must put one even temporalely...
Put 2volt sinus or music from a CD and load the amp with the loudspeakers of your choice
Put the ground of the osciloscope proof at the + of the loudspeaker and the other side of the osciloscope prove at the input of the amp before the pot...
Imagine that you are doing the proof with 2Volts sinus...
at the minimum of the volume you will see 2 volts at the osciloscope as the output is at zero
Now increase the volume...the voltage in the osciloscope begin to decrease and you will find a minimum where the output voltage is the same as the input voltage...that residual signal is the overall distortion of the amp...and this is true for all distortion typs invented or not yet invented...
So easy ...so simple...and so true
You have another explanation of this test in:
Post #8 in
http://www.diyaudio.com/forums/showthread.php?s=&threadid=13415&perpage=15&pagenumber=1
mikek said:...i reckon not even God Himself can convince pan that his audible cables, and different sounding shakti stones...amps. etc are well known, and scientifically accounted for tricks played on him by his brain....
Where did you get the shakti stones from?
I never talked about such and I certainly do NOT believe in those.
Have not tried them and have no intention to do so either.
About the audible differences between cables and amps, that is certainly not tricks in my brain. These are real, easy to hear differences that all people that have listened to the same components also have heard.
The tricks are in your brain, failing to understand what high performance audio is about.
Have a nice Sunday and happy listening!
/Peter
Half Valid Testing.
This is quite a big 'so what' actually.
When a stereo amplifier is driving two speakers in a room, a lot of things happen that do not happen in your test method.
Firstly, the stereo amplifier power supply is only half loaded, when driving only one channel, and PSRR issues are much lessened.
Secondly, speakers act as generators when subjected to in room sounds, and this will cause further PSRR and crosstalk issues.
Your test method is not fully representative of real world listening conditions, and therefore invalid - sorry.
A more correct method I expect would be to capture (to hard drive) both channel output waveforms for both amplifiers, and then compare them.
This would require a high sampling rate/high resoloution soundcard to achieve the definition required.
Eric.
This is quite a big 'so what' actually.
When a stereo amplifier is driving two speakers in a room, a lot of things happen that do not happen in your test method.
Firstly, the stereo amplifier power supply is only half loaded, when driving only one channel, and PSRR issues are much lessened.
Secondly, speakers act as generators when subjected to in room sounds, and this will cause further PSRR and crosstalk issues.
Your test method is not fully representative of real world listening conditions, and therefore invalid - sorry.
A more correct method I expect would be to capture (to hard drive) both channel output waveforms for both amplifiers, and then compare them.
This would require a high sampling rate/high resoloution soundcard to achieve the definition required.
Eric.
I Agree With You.
Hi Peter,
Yes you can hear these differences, and so can I, and so can a heap of other listeners.
For various reasons other claim that they cannot hear these levels of differences, and therefore claim that these differences are non-existant.
Then they propose tests to validate their claims, but these tests are themselves invalid so we are all back to square one.
In absence of a suitably finely resolving test methodology, we are left with our ears as the judging tool, and as is to be expected there are variations in listening ability.
This still does not however give credence to the 'psychological factors' stance as proposed by the purely objectivist followers.
Eric / - Keep using your ears.
PS - Ever heard of a VPI Brick ?.
Hi Peter,
Yes you can hear these differences, and so can I, and so can a heap of other listeners.
For various reasons other claim that they cannot hear these levels of differences, and therefore claim that these differences are non-existant.
Then they propose tests to validate their claims, but these tests are themselves invalid so we are all back to square one.
In absence of a suitably finely resolving test methodology, we are left with our ears as the judging tool, and as is to be expected there are variations in listening ability.
This still does not however give credence to the 'psychological factors' stance as proposed by the purely objectivist followers.
Eric / - Keep using your ears.
PS - Ever heard of a VPI Brick ?.
Nope, Still Not Valid, Try Again.
Hi Jorge,
Your test compares the signal apearing at the output terminal with that apearing at the input terminal, and yes is a measure of the transfer function of the amplifier under test only.
However, it does not reveal changes at the input terminal due to loading of the previous stage.
To improve your test, your reference signal needs to be further upstream - at the cdp audio output stage, and upstream of the output buildout resistor, and not at the amplifier input terminal.
Further, your test is at 2V level, and this is not high enough for the reasons given in my above posting.
Eric.
Tube_Dude said:Hi Eric!!
But you can do my test with the two chanels loaded by the loudspeakers...and in real live conditions...and even playing music...not with a sinusoide signal...
If this is not valid!!!!!!
Hi Jorge,
Your test compares the signal apearing at the output terminal with that apearing at the input terminal, and yes is a measure of the transfer function of the amplifier under test only.
However, it does not reveal changes at the input terminal due to loading of the previous stage.
To improve your test, your reference signal needs to be further upstream - at the cdp audio output stage, and upstream of the output buildout resistor, and not at the amplifier input terminal.
Further, your test is at 2V level, and this is not high enough for the reasons given in my above posting.
Eric.
Re: Bricks? Building Glass Houses
Hi Frank,
Mention was make of Shakti stones.
I have heard the effects of VPI Brick and Mpingos, and whilst the audible effects are fine, they are probably difficult to measure.
My point is that whilst difficult to measure, the effects are clearly audible on a finely resolving audio system.
Unfortunately, sales pitch of Mpingos etc etc are cloaked in gobbldegook and doubletalk, but this does not deny the existence of sonic effects.
It seems that many around here automatically discount that these sorts of devices can have ANY effect because of the loose marketing terms, especially those who have not tried them for themselves.
Eric / - Always open minded when it comes to audio.
fdegrove said:Hi,
Yup, used a similar offering from the "Hi-Fi News Accessories" called flux dumpers if I recall it correctly...till they fell apart that is.
What's your point here Eric-san?
Ciao,
Hi Frank,
Mention was make of Shakti stones.
I have heard the effects of VPI Brick and Mpingos, and whilst the audible effects are fine, they are probably difficult to measure.
My point is that whilst difficult to measure, the effects are clearly audible on a finely resolving audio system.
Unfortunately, sales pitch of Mpingos etc etc are cloaked in gobbldegook and doubletalk, but this does not deny the existence of sonic effects.
It seems that many around here automatically discount that these sorts of devices can have ANY effect because of the loose marketing terms, especially those who have not tried them for themselves.
Eric / - Always open minded when it comes to audio.
I have heard the effects of VPI Brick
Hi,
While I can offer an explanation based on mere physics for the VPI bricks, doing the same for a lot of other mystical gadgets is going to be a though cookie though.
Nonetheless, and I emphasise this, when a difference is heard, repeatedly so, and not just by one individual than something IS going on...just waiting for an explanation perhaps?
Can someone please pick up my mind? It just fell out.
Cheers,
Hi,
While I can offer an explanation based on mere physics for the VPI bricks, doing the same for a lot of other mystical gadgets is going to be a though cookie though.
Nonetheless, and I emphasise this, when a difference is heard, repeatedly so, and not just by one individual than something IS going on...just waiting for an explanation perhaps?
Can someone please pick up my mind? It just fell out.
Cheers,
- Status
- Not open for further replies.