ELEKTOR September article - Page 5 - diyAudio
Go Back   Home > Forums > Commercial Sector > Vendor Forums > Elektor

Elektor News and offers from Elektor Magazine

Reply
 
Thread Tools Search this Thread
Old 17th September 2007, 09:22 AM   #41
diyAudio Member
 
ashok's Avatar
 
Join Date: Jun 2002
Location: 3RS
".....a computer with sound card, a pair of speakers or headphones and some THD analysis software or equipment can do the same test and see the shocking results for themselves....."

This kind of test probably will not replicate the same kind of spectrum as the harmonics of an unclipped waveform would. You can pass off as undistorted an unclipped sine wave with harmonics , though the sound would be 'different ' with different levels of odd and even harmonics. Even harmonics making it sound warmer and odd harmonics making it sound 'harder' .
A clipped waveform would have very high levels of higher oder harmonics making it unpleasant.
SE tube amps prove the point. Relatively high even order harmonics but still sounds great.

The point I was trying to make is that a clipped waveform is not the proper way to determine at what level harmonic content becomes audible.
__________________
AM
  Reply With Quote
Old 17th September 2007, 12:20 PM   #42
Pierre is offline Pierre  France
Banned
 
Join Date: Nov 2004
Location: Paris
Ashok,

So according to your thoughts, testing distortion perception the way Eva proposes, you are clipping the waveform, and even in that case, which is producing a more unpleasant form of distortion (due to the distribution of harmonics), you require very high THD numbers to be able to detect it.

So with non-clipped sound, you would need even higher THD to notice!
  Reply With Quote
Old 17th September 2007, 01:11 PM   #43
diyAudio Member
 
Join Date: May 2002
Location: Switzerland
Quote:
The point I was trying to make is that a clipped waveform is not the proper way to determine at what level harmonic content becomes audible.
Exactly

Distortion is often only aurally detected with quite high figures if we just listen to the change of timbre. Timbre change is the only property of harmonic distortion that is aurally perceivable with a sinusoid for instance and this can be masked quite strongly by speaker distortion.
But distortion mechanisms like crossover distortion will mask fine details in music because it has the most influence at low levels.
Class-d amps usually suffer less from the latter than class-AB.

Furthermore speaker distortion is usually only 2nd and 3rd order which is less perceivable than higher order distortion (that is often generated by electronics). I assume that there are two resons for this: 1.) they are more harmonically related (in the musical sense) than the higher order ones and 2.) they are more likely to be suppressed by our hearing's masking mechanisms.

Regards

Charles
  Reply With Quote
Old 17th September 2007, 01:46 PM   #44
diyAudio Member
 
Join Date: Aug 2004
Location: Uppsala,Sweden
In a system many things matter. Loudspeaker distortion could be kept at low levels, but drivers providing this are often expensive,
e.g. Mundorfīs big AMT units (about 800 Euro/each). Our hearing is less accurate in the bass, so distortion here may be unnoticed if it doesnīt interfere with mid-frequencies; the area in which our hearing is usually best developed (probably for evolutionary reasons).

So what are the criteria for a good amp? No one seems to agree in all the aspects, but if I may guess for a moment?

Constant output impedance in the audio area.

Little phase shift. It seems our hearing is sensitive to time domain faults. And of course you can again say and claim in general that the loudspeakers are much worse than amps in this case, but this is what I believe in.

Small distortion figures when it comes to tranisent signals. This means that a large bandwidth may be necessary. The amp should be able to handle all signals that passes through it, and an input filter that limits upper frequency response is preferable to internal slew induced distortion, or even distortion caused by the output filter. Short randomly repeated tone bursts are probably a better way to assess amps than traditional methods.

How does a amp "recover" from being driven hard, even if clipping has not been the case. In other words, is there a thing which could be called "memory distortion"? Does an ampīs responds depend of what happened just a fraction of a second earlier? And of course recovery from clipping is important as well!

A nice small and rapidly falling spectrum (for higher overtones) when it comes to steady tone measurements of harmonic distortion.

This is from the HIFI Akademie (German) power amp, text is in German but itīs the spectra that matters.
http://hifiakademie.de/?id=1.2.2&si=...NS4zNS4yMzV8IA

BTW, crossover distortion is usually independent of how hard the amp is driven (in non-clipping conditions at least). Thus the sensitivity of the speakers may count if this form of distortion is perceived or masked.
--------------------------------------------------------------------------------
The benefit of the F/E method I have mentioned is that one of the drawbacks of conventional blind testing; that the test persons tend to concentrate on recognizing the objects under test instead of telling how the objects sound: i.e. how good they are, this limitation falls to the ground. Quite the contrary, in this sort of test the purpose of it is to see if and in case how many times and by which test persons an object can be identified.

A drawback may be that e.g. an amp may be identified by a clearly bad quality and another by a less negative but still obvious quality. Both amps are identified to the same extent, but one is inferior. The best amp is the one that canīt be identified at all, but in reality this seldom happens thus a small degree of identifications is what usually counts for the verdict "a good amp".
__________________
Without wonderment,no life.
  Reply With Quote
Old 17th September 2007, 02:40 PM   #45
Eva is offline Eva  Spain
diyAudio Member
 
Eva's Avatar
 
Join Date: Oct 2003
Location: Near the sea
Send a message via MSN to Eva
Quote:
Originally posted by ashok
".....a computer with sound card, a pair of speakers or headphones and some THD analysis software or equipment can do the same test and see the shocking results for themselves....."

This kind of test probably will not replicate the same kind of spectrum as the harmonics of an unclipped waveform would. You can pass off as undistorted an unclipped sine wave with harmonics , though the sound would be 'different ' with different levels of odd and even harmonics. Even harmonics making it sound warmer and odd harmonics making it sound 'harder' .
A clipped waveform would have very high levels of higher oder harmonics making it unpleasant.
SE tube amps prove the point. Relatively high even order harmonics but still sounds great.

The point I was trying to make is that a clipped waveform is not the proper way to determine at what level harmonic content becomes audible.
I think that you are just repeating what you have read from other people about THD without any personal experimentation. Consider these facts:

- Usual clipping does not produce more higher harmonics than lower harmonics, if you sort the harmonics by amplitude you always get 3rd > 5rd > 7rd etc... no matter how much clipping is applied... unrtil you get an square wave. I have done the test while looking at the spectrum analyzer.

- This story about warmer sounding and cooler sounding that everybody is telling is a plain myth. You can actually hear the separate tones that result when you clip a given tone. I can't detect a progressively clipped 1Khz tone until the resulting dominating 3Khz tone becomes loud enough to be heard, and if I increase clipping further I can clearly identify 3Khz.

- Higher harmonics resulting from clipping are easily get rid of by turning trebble down, leaving only 3rd and 5th
__________________
I use to feel like the small child in The Emperor's New Clothes tale
  Reply With Quote
Old 17th September 2007, 09:03 PM   #46
diyAudio Member
 
Join Date: Jul 2004
Location: Seattle USA
Quote:
Originally posted by Eva


I think that you are just repeating what you have read from other people about THD without any personal experimentation. Consider these facts:

- Usual clipping does not produce more higher harmonics than lower harmonics, if you sort the harmonics by amplitude you always get 3rd > 5rd > 7rd etc... no matter how much clipping is applied... unrtil you get an square wave. I have done the test while looking at the spectrum analyzer.

- This story about warmer sounding and cooler sounding that everybody is telling is a plain myth. You can actually hear the separate tones that result when you clip a given tone. I can't detect a progressively clipped 1Khz tone until the resulting dominating 3Khz tone becomes loud enough to be heard, and if I increase clipping further I can clearly identify 3Khz.

- Higher harmonics resulting from clipping are easily get rid of by turning trebble down, leaving only 3rd and 5th

Hi Eva
This is absolutely spot on. Every point.

I pointed out this site in another thread:
http://www.st-andrews.ac.uk/~jcgl/Sc...ing/page3.html

I postulate that really the "bad" sound from music clipping is due to the input signal. To which, as you point out, turning down the treble gets rid of.

Next argument on "warmth" I often see posted are the quality of the distortion. The old even verses odd harmonic. This usually ends up in repeat some misconceptions about devices used to make an amp, but as Dr. Berners at UA audio points out "absence of even harmonics has often been (wrongly) attributed to solid-state technology"

PS - Here's a ear training site with some THD examples.
http://www.pcabx.com/technical/nonlinear/index.htm
I generally think it's a bad idea to teach yourself to notice small distortion elements that you normally wouldn't. It's like training yourself to hear the resonance of your ear canal.

PSS - I think we're all in good company with ColdAmp and NCD posting to this forum. (Hypex in other threads) It's pretty incredible the quality of these modules and the support they give.
  Reply With Quote
Old 19th September 2007, 01:41 PM   #47
diyAudio Member
 
Join Date: Aug 2004
Location: Uppsala,Sweden
How come that so much irrelevant information about the performance of amps have been the norm for so long?

Harmonic distortion under steady state conditions doesnīt say much but itīs likely that our hearing is more sensitive to distortion created under transient conditions. Some like Vandersteen make this claim and say they have empirical data to conform the claim. (The instrument which exhibit the least overtones are said to be the flute and some old analog synths may also be a rather easy signal to amplify).
An ampīs ability to respond when driven into e.g. 8 Ohms in parallel with a capacitor may say something about the stability but I donīt believe this to be enough!

Power amps are supposed to deliver current into often reactive loads so the stability of amps and the function of feedback systems receiving delayed signals may be of interest. This also incorporates the ability of the amps to handle back-EMF from the speakers.

An amps distortion to the mains may be of interest as well. And when it comes to Class D amps EMI is of importance as its occurrence can affect sources and preamps.

I think there was a time many years ago when there were a genuine interest in better and more relevant audio measurements.
For some reason this seems to have declined. Why?
__________________
Without wonderment,no life.
  Reply With Quote
Old 19th September 2007, 02:41 PM   #48
soongsc is offline soongsc  Taiwan
diyAudio Member
 
soongsc's Avatar
 
Join Date: Mar 2005
Location: Taiwan
Quote:
Originally posted by The golden mean
...
I think there was a time many years ago when there were a genuine interest in better and more relevant audio measurements.
For some reason this seems to have declined. Why?
I miss those times too. It seems making money has taken the front seat to making better amps for most businesses.
__________________
Hear the real thing!
  Reply With Quote
Old 19th September 2007, 02:53 PM   #49
diyAudio Moderator Emeritus
 
Bas Horneman's Avatar
 
Join Date: Apr 2002
Location: The Netherlands
Blog Entries: 18
Quote:
I think there was a time many years ago when there were a genuine interest in better and more relevant audio measurements.
There must be many reasons. But one reason I think simply has to do with the amount of research that that would take. Who would pay for that?

Certainly most big players won't...there is zero requirement for that kind of research (not from the engineering department or marketing department).

Most people are convinced that a home theatre receiver with cd or dvd player is already perfect sound forever. So we've come to the end of history as far as most people on earth are concerned.

So the research for better measurements would be for a couple of freaks like us in the fringe. Hopefully most of us will continue to use our ears to be the final judge in any case.
  Reply With Quote
Old 19th September 2007, 04:23 PM   #50
soongsc is offline soongsc  Taiwan
diyAudio Member
 
soongsc's Avatar
 
Join Date: Mar 2005
Location: Taiwan
Quote:
Originally posted by Bas Horneman

...
So the research for better measurements would be for a couple of freaks like us in the fringe. Hopefully most of us will continue to use our ears to be the final judge in any case.
I just watched a documentary on the history of hacking. It also talked about the freaks that started out on little computers. Who knows what we will end up with.


__________________
Hear the real thing!
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
We described a 500W amplifier in the August, September & October 1997 issues of SILIC jamesfeline Solid State 7 26th January 2011 05:25 AM
Elektor Nov 93 - Power Mosfet Matcher Article Zero Cool Elektor 10 12th May 2008 11:37 PM
Elektor September 2007 issue.... mobileusa Elektor 0 14th March 2008 04:36 AM
Elektor DRM article rpapps Everything Else 9 9th November 2005 11:55 PM


New To Site? Need Help?

All times are GMT. The time now is 09:58 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright Đ1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2