Home Propeller Head Plaza

Technical and scientific discussion of amps, cables and other topics.

Re: Another Sample Of An Audio Component That Doesn't Measure Well, But Sounds Great.

Hi Tubeguy

Thanks for the excellent discussion, sometimes its good to stop and think about why one has an opinion about something or, what is your opinion, now that you had a chance to ponder.

So far as distortion, I see it this way.
Harmonic distortion is caused by a (the sum of) nonlinearity somewhere and as a “wire with gain” or a loudspeaker, ideally has none.
Back when they first started measuring it, pretty much all the gear (tubes) had the same general kinds of distortion.
At this point in time, 10% was actually better than 20% THD when listening, so, “less distortion is a good thing”
As time went on, going from 5% to 1% was sometimes audible too so “less distortion is a good thing”.
During this same general time the bandwidth of amplifiers surpassed the magic 20 –20K and so what else can be improved?

After amplifiers reached the point of diminishing returns so far as “audible” improvement, much more effort started going into exotic looks and “less distortion is a good thing”.
Some of the gear now is down right silly looking (to me), as silly as the price tags too.

For those who “attack” the single ended amplifier because of harmonic distortion, refer them to some research work a friend has done on moderate amounts of distortion which might be “thought provoking”.

http://www.gedlee.com/downloads/THD_.pdf

http://www.gedlee.com/results.htm

So, let me ask you this, don’t most SET folks use high efficiency loudspeakers, yes?
Lets say you had a speaker that was 102dB 1W 1M, an efficient speaker but not the most efficient.
Now lets say you had another speaker that was 88dB 1W 1M.
You have a 5 Watt amp on the efficient speaker and a 125 Watt amp on the inefficient speaker.
Which one can produce the higher sound level?

The 5 Watt amp can because although the power and sensitivity cancel each other out (making them equal), the low efficiency speaker will also suffer much more dynamic compression because the Voice coils are heated by power.
ALL loudspeaker drivers begin to exhibit changes in response and dynamic compression at powers greater than about 1/8 to 1/10 rated power.

Now, look at the program material.
A “good” hifi demo recording (on CD) might have a peak to average level difference of 30 dB and the quietest parts (not including digital silence between tracks) might be 40 or 50 dB below the peak level.
If you had a peak level of 5 Watts, this puts the average level at –30dB or .005 Watts and the quietest parts at .0005 to .00005 Watts.
A “distortion analysis” relevant to music, would involve measuring the amplifier over its entire range.
Do they even measure amplifiers at what would be the “average” signal level?

Another point worth mentioning from Earls work is he found distortion was most detectable as the signal was near / crossing through the zero level.
Now, isn’t that right where crossover nonlinearity is on an AB or B output stage and exactly where a class A stage is most linear, why what do you know, yes it is isn’t it.
Doesn’t that also mean that when a tube amp “soft clips” or compresses, that this would be “less” audible than its “number” might suggest? Why yes to that too.
If one were especially sensitive to THAT (nonlinearity near zero), then, its presence may be audible while at a very low “number” couldn’t it?
I guess what I’m saying is there are lots of shades of gray and many variables here, a feeling of certainty here is only possible if your not seeing to many of these annoying little details.
These details are not relevant to selling product however.
Rather than science providing all the answers to questions, it often answers one in exchange, leaving you with another question or three to ponder.
Best,

Tom Danley
What are your current speakers?




This post is made possible by the generous support of people like you and our sponsors:
  Parts Connexion  


Follow Ups Full Thread
Follow Ups


You can not post to an archived thread.