|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
96.25.74.167
In Reply to: RE: metrics of amplifier quality/value? posted by zdog on September 01, 2014 at 07:40:01
"In another life, I am a data analyst, so my instinct still tells me..."
That it's really the corrolation coefficients that matter? Probably. But how the heck do you establish it?
Many years ago I got the notion that I needed a more powerful amplifier, perhaps one with a lower output inpedance but knew that I wanted to stick with the general style I had: transistorized push-pull. (In hindsight, maybe I just wanted to buy something new..)
Since I didn't care what it looked like or cost (within reason) I was ready to roll, I had my requirements: >120 W/ch. SS. The power requirement was figuring it needed to be at least 3dB more than my existing unit. That was long enough ago that there still several dealers within a 50 mile radius so I could drag stuff home.
The bottom line: They all sounded differently. Some moreso than others but I wasn't really expecting any sonic shifts beyond just more ummph in the transients. It was an ear-opener and in general I only listened rather than measured. However one unit was doing something that was so eerie that I hauled home some gear and sorted it out. Even when no music was playing it was making me nervous just being turned on. That turned out to be a problem with very narrow transients from the reverse-recovery of the rectifiers making their way into the output via the C-E capacitance in the bugs and being reproduced by my film tweeters.
I'm still using the amplifier I bought now about thirty years later so it was probably worth the effort. But the specifications and topology really aren't adequate to characterize the performance of home audio gear for fussy listeners. I think they could be improved quite a bit but there doesn't seem to be much interest.
Try reading some of morricab's posts while you are in the analytical frame of mind, he finds corrolation between the pattern of the distortion harmonics and the quality of the sound. In preamps at least I have found a correlation between the group delay and the SQ. The sooner the first inkling of signal shows up at the other end, the better. This may be an oddball way to see T.I.M.
IMHO the 'routine' specifications: THD, Frequency response, Power out at 1KHz, Output impedance, were more useful in the old tube era when they were written. We never characterize every possible parameter of anything, it would just be insane. A lot of things are just intrinsic in certain materials and topologies so the old Specs. look at the areas where tube designs tend to come up short and they also reflect what was commonly measurable then. Audio frequency spectrum analysers simply didn't exist. For that matter even scopes were rare.
If we really had the machine/man interface well defined your approach would have merit, as it is you will be looking for patterns in what is essentially noise.
Wish that it weren't so but that's the state of the art...
Rick
Follow Ups:
My background comes with a LOT of statistical process control, so I'm kind of data driven, myself.
If you get a chance, take a look at my post about REACTANCE and how amplifiers are seldom measured for performance INTO such loads.
That may help you close the gap on some of your musings. Also, One OTHER thing to toss into the mix, espcially when dealing with group delay and such would be BANDWIDTH. My new stuff is very wide bandwidth and sounds much better than the previous 'D' amp with the output Zobel getting in the way.
Too much is never enough
"look at my post about REACTANCE and how amplifiers are seldom measured for performance INTO such loads"
Yup, right on. And it isn't just power-amps either, preamps have the similar issues. They don't have to deal with mechanical reactance and resonance but on the other hand they usually face worse dielectric absorption problems and RF ringing issues. Having a lot of bandwidth just sort of takes the load off unless something downstream doesn't!
Rick
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: