obviously possibly different, depending on how non-flat the frequency response is. If we're talking about +/- 0.1db 20-20000hz it's probably going to be very hard to tell the difference. If however one amp is +0.4 db from 100-1000hz and then +0.2dbs from 2500-5000hz it might possibly be audible. Can you point me to a place where anyone here has suggested that amps that have significantly variant freq response might not be audibly distinct? Nobody really cares much about those amps because ftmp we here at ASR would view those amps as "broken."
One question, do you consider that 2 amplifiers with a different frequency response curve sound the same or different?
Although I'm not sure what the point of this thread is, I'll play a long for a just a little bit:
One would think ASR types would consider such an amp broken,
but they often don't. Why? This sort of thing is not always easy to catch in testing. I'm not about to read 52 pages of comments for an amp I don't care about at all, but the Wiim Amp clearly has this issue, and it is noted in the review as "load dependency". Since it
appeared to impact 20kHz, no one cared. But just how problematic the load dependency is did not get fully captured in basic tests here because ASR does not use a complex load. To date this amp has a 85%+ "Fine" or "Great" user rating. I just slapped a "Poor" on it because it really does suck that bad by any objective measure where "potentially audible issues" is the standard. It's miles worse than the boat of of amps with 78dB SINAD that get regularly panned. Unfortunately, it can be hard to tease this out unless you get really in depth on the testing,
and understand what measurements mean. EAC got more in depth, and found significant response deviations on complex loads of up to 1dB. A lot? No, but it is enough that you're almost
guaranteed to hear it on
certain program material. Is it broken? Yup. Did it get panned here? Nope. Not as hard as it should have. And I'll bet you LOTS of people bought it. Even when the EAC review came up, most of the comments there were about output power issues. Not the load dependency. And this is forgivable, since the load dependency even at (potentially) 1dB will still be hard for a lot of people to detect.
So, what's the point of bringing this up (other than my weeks long theme of picking on people for being obsessed with measurements but often not understanding them nearly as well as they ought for people obsessing over measurements)? Well, let's go back to OP's question about the Carver amp. My recollection (and I think someone mentioned) that the bad SINAD was mostly distortion. The SNR is actually not that bad. So there's not going to be anything
obviously amiss. On plenty of program material, over plenty of speakers, I would venture a guess that 60% of listeners,
at least, would fail to differentiate the amp when used within its limits. But there's some load dependency there, too, so like the Wiim amp, it's going to have frequency response deviations. Now, you take a trained listener with proverbial golden ears and hook both amps up to a set of extremely low distortion speakers? They'll probably nail the Carver it every time. Whether the distortion will be objectionable is an entirely different matter.
Now let's add another layer: Play some symphonic material with 20dB dynamic range. Turn up the volume. The Wiim is easily driven past its modest limits. The Carver is still cooking. The Carver even with all its distortion will be a far superior amp to the Wiim or any other low powered toy amp, and audibly so. Now toss a decent Class AB amp in there. Depending on the speakers, you might be able to pick it out from the Carver. In fact, you probably will with careful listening since the Carver just has
that much distortion. Will it sound better? I have no idea. But it will sound more accurate. But now go and throw almost any other quality AB amp into the mix with similar output power. Your little project just got
much harder.
The question is, from what levels of difference can we differentiate? 1, 10. 100, 1000, 10000 times worse?
In the end, OP END_WAY, you're trying to distill a question with a complex answer down into some simple Yes or No answer. It's not that simple. At what level can you differentiate? Totally depends on the acuity of the listener, the quality of the speakers, potentially the load presented by the speakers, the volume level selected, the program material, you name it. At what level is "good enough"? Assuming flat response and sufficient power not to clip or vary into the load at all impedances and phase angles, probably way down at like .06% THD from 20Hz to 20kHz and 65db or 70dB unweighted SNR (if not worse) at 1W, on insensitive speakers. In other words, a lot "worse" amps than most people here probably think they need. Only the odd man out is going to be tell much of a difference past that. Once you figure that out, you realize that proper amp "audibility" testing can likely be distilled down to a power test at 20Hz and 1kHz, a power swept 1kHz and 20kHz tone, SNR at 1W, and a complex load test. Done. Most amps will pass and only a handful won't. The worst test will be 20kHz, which really doesn't promise anything audible anyway. Just hints at it.
EDIT: If you want to be really sure, go for .01% or less at 20kHz. That mostly ensures extremely competent engineering because that is difficult. What static testing does not do is guarantee there are not dynamic distortions. Hypothetically, you could be an evil pr*ck and engineer an amp that would look great to an Audio Precision and still sound like absolute garbage due to some sort of unspecific dynamic distortion that the AP does not catch.
They’ll drive anyone into despair!
