(1) The room is not important once you agree upon a "reference room" and a reference recording in a reference hall. Somewhat similar how THX started.
(2) What sounds good for one listener is not important, what is important is what sounds realistic to a population of listeners. We are looking at comparing the sounds in the reference hall with the recorded sound in the hall when played via a reference system in a reference room to figure out what electronic/mechanical behavior is the most realistic and desireable... ie: "good sounding".
(3) The reference system is also not absoultely important either, turns out. I think we could have a reference system for each person... the trick is that we are trying to measure the DIFFERENCES between the live event and the recorded event...
(4) The listener is necessary because we want to figure out the psychoacoustics related to hearing and interpreting the sound
Note I'm going one further. Before understanding what is a good amplifier, I want to understand what makes a realistic "sounding" amplifier. Indeed, this goes for most circuits and transducers in the chain. At no time should be change more than one time in the reference "chain"...
We are attempting to shift amplifier design away from a mad dash to sub ppm distortion figures to one which makes sure other important issues are also covered.
We aren’t trying to tell you or anyone else what a good sounding amplifier looks like, or what you think sounds good. That’s a subjective judgement and that has been specifically excluded from the proposed assessment.
Can it be made any clearer?
We aren’t trying to tell you or anyone else what a good sounding amplifier looks like, or what you think sounds good. That’s a subjective judgement and that has been specifically excluded from the proposed assessment.
Can it be made any clearer?
You want to take black box measurements with a soundcard and an oscilloscope. That's it, right.
What is there to say about that. There are standard tests and existing ways to weight FFT harmonic spurs.
There are also pretty standard scope tests.
You just want to pick the best 10 or so out of that menu, and call it done?
What is there to say about that. There are standard tests and existing ways to weight FFT harmonic spurs.
There are also pretty standard scope tests.
You just want to pick the best 10 or so out of that menu, and call it done?
Unfortunately, IMHO, you are ignoring other aspects, aka "important issues" that also go into the determination of "quality".
Unless, you are changing the definition of "quality"...
You all asked for comments, I gave them to you... if you can't handle comments, then don't ask for them.
Who else has that goal? You're trying to do that. It's clear.
Why does lower distortion come at the expense of other factors? You've said that over and over. It doesn't. It MAY, sure. Is this for DIYers at all? Are you trying to affect the commercial market? Show me a few examples of commercial amps that have gone up in flames (or otherwise failed catastrophically) or had reliability issues because they had sub ppm distortion. Yes, I used "because" thus implying causality because you seem to think that designers will look at your scoring system and somehow choose to make "better amplifiers", correct?
Also... even though it seems to be set in stone... I will try one more time (then I promise to stop).
It's really hard to support taking continuous data and bracketing / bucketing / categorizing it. It's not only a pet peeve of mine, it really is considered bad practice.
If one of the goals is to direct engineers designers toward more practical choices.... or "informing" people of what the designers of the study / scoring think is better or "more important"; then see below example with a more practical situation and see if it aligns with your intent.
One example -
Below are the initial proposed criteria for THD+N
| 1 point for -65 dBV; add 1 point for every 10 dBV improvement up to -105 dBr |
Assuming a table is created and winds up something like I've put below with agreed rounding up or down or truncation ... or whatever...
What happens between -75 and -75.1 that's so magical that the amplifier gets a full extra "point"?
If an amplifier designer is actually going to use this scoring to determine their direction, is it wise to spend time and effort to improve THD+N from -105 to -105.1dB? If not... then consider changing to a different type of scoring. IMO, it is counter to the stated purpose vs. supporting it. Your previous examples given were akin to saying that goal was to not "reward" beyond -105dB. Fine, don't "reward" anyone at all. Just the facts, please.
Heck, depending on the work done, the 0.1 difference may be within test error. We don't know yet.
Don't say that the difference between 1 and 2 doesn't matter in the grand scheme of things; that's also counter to the stated purpose. If the difference between 1 and 2 matters, then so does the difference (however small) between the upper edge of bucket 1 and the lower edge of bucket 2.
It's not that I wholly disagree with the stated goals. I'm trying to provide some advice from experience to help guide toward more meaningful metrics / scoring within your stated goals. A score is generally intended to incent/discourage behavior or using your words "reward".
My general idea is that I'll take what I can get, and I love the overall concept. I'm also one to put my support behind the effort and let it go if my suggestions are not implemented.
If you've understood the point, and choose to not implement the suggestion, it's not an issue. I still think this is an extraordinary effort. Continual quibbling doesn't help it move forward.
I simply happen to think that if scored, a very well-thought out scoring mechanism is vital... and I'm not in love with the overall idea of this one. I'll happily take the raw data (if available) and use it for my own purposes w/o scoring. I hope the data will be somehow available vs. simply the 1-5 and/or total scores.
0 | > -65 |
1 | -65.1 to -75 |
2 | -75.1 to -85 |
3 | -85.1 to -95 |
4 | -95.1 to 105 |
5 | -105.1 to -...... |
Who are 'we' and did you ask them if they agree with your viewpoint?
Why do you think that people who challenge your method are in favor of sub ppm designs?
How did you even came up with the idea that you can read other peoples mind?
How can a subjective judgment be excluded from your proposal if your proposal is build on a subjective weighting system?
To me it's clear, the potential of objective measurements is heavy degraded by locking them into a subjective rating system.
@Bonsai - adding to #386
I hope it's clear that I do support the effort. I'm not a naysayer.
Additionally - As previously suggested.... if you must reward or score... consider scaling the continuous data accordingly vs. "buckets".
Consider using a simple multiple of the result to bring the value up or down as you see fit to align with your desired level of importance / relative to other metrics.
=result in dBV*somefactor + some other factor
You can do linear scaling ... or whatever you like it for relative importance within the measurement and between measurements. You can keep it linear within the measurement to possibly allow more direct comparisons between amplifiers. You can also use a geometric or exponential adjustment ... whatever you choose.
Another different goal is to have it "reversible". If for some reason you decide down the road to re-bracket the categories. It's easily done. A 7.51 can easily be returned to a -75.10. If the resolution is lost and a -75.1 becomes a "2", then it's much, much more work.
To me... it's all gain with no loss. I admit to possibly missing a compelling reason as to why it must all be whole numbers, but ...
My only admittedly VERY long-winded point is to keep the data continuous for what I think are good reasons. I hope you'll see those reasons as valid also, and consider it.
I hope it's clear that I do support the effort. I'm not a naysayer.
Additionally - As previously suggested.... if you must reward or score... consider scaling the continuous data accordingly vs. "buckets".
Consider using a simple multiple of the result to bring the value up or down as you see fit to align with your desired level of importance / relative to other metrics.
=result in dBV*somefactor + some other factor
You can do linear scaling ... or whatever you like it for relative importance within the measurement and between measurements. You can keep it linear within the measurement to possibly allow more direct comparisons between amplifiers. You can also use a geometric or exponential adjustment ... whatever you choose.
Another different goal is to have it "reversible". If for some reason you decide down the road to re-bracket the categories. It's easily done. A 7.51 can easily be returned to a -75.10. If the resolution is lost and a -75.1 becomes a "2", then it's much, much more work.
To me... it's all gain with no loss. I admit to possibly missing a compelling reason as to why it must all be whole numbers, but ...
My only admittedly VERY long-winded point is to keep the data continuous for what I think are good reasons. I hope you'll see those reasons as valid also, and consider it.
But you are arguing around what is audible when you try to sort out measurements and thresholds - this is a subjective judgment and right now it seems like you will miss your goal as I see it.
//
He argues against nonsensical race for lowest sub ppm THD limits as we can see it at ASR. I support his approach.
I merely pointed out the prohibitive costs of the kind of tests you were proposing and that the idea behind the assessment was an objective assessment of amplifier performance other than sub ppm distortion.
How are subjective assessments of quality going to help if everyone has a different idea of what sounds good?
I’m quite capable of taking feedback but forgive me if I pushback on some of this stuff.
Question: what objective assessments of amplifier performance other than ultra low distortion do you think are important. If you can provide that then you are contributing positively but allow others to challenge your view.
If you don’t want to offer anything constructive, or don’t like the idea, why participate?
A weighting system isn’t subjective if the weighting parameters are clear and agreed. That’s the reason for opening the thread. Comment on the concrete proposals if you want to participate and offer some input.
I am struggling to understand where the ‘how can you read other peoples mind’ thing comes in.
Last edited:
Problem is there are really only two ways: either just keep in the tech race
But the thread starter has already stated the reason for why there is a need for a new "figure" - why shouldn't a tube amp have a good rating... this is either subjective from a socio-echonomic perspective or an SQ one. So it is a fail already because it breaks the thread starters stated goal when he says: "That’s a subjective judgement and that has been specifically excluded from the proposed assessment."
It aint going well...
//
I agree that the one parameter race is not ideal (I do think it has probably improved the general level of amps since a few years compared if it hasn't been around at all) - but I do think that the strict technical way is probably the only way. What needs to be improved has been mentioned - add tests so that there is no way of letting a bad amp come through - and this should not be based on listening aspects but on analytical knowledge of what an amp is and how it behaves. Its just about amplifying a signal...
But the thread starter has already stated the reason for why there is a need for a new "figure" - why shouldn't a tube amp have a good rating... this is either subjective from a socio-echonomic perspective or an SQ one. So it is a fail already because it breaks the thread starters stated goal when he says: "That’s a subjective judgement and that has been specifically excluded from the proposed assessment."
It aint going well...
//
I’ve proposed some first cut tests in the document.
The whole debate seems to have run aground on the notion that standard distortion tests might not be measuring the right thing. IMD and THD were proposed and the question asked if there was a better distortion measure and could one be proposed.
If you’d like to make a proposal please go ahead. If you think IMD and THD are ok, that’s fine.
Wow!! I better not use any amp with ONLY 7th harmonic at 0.01% Goes & checks zillions of amp measurements and finds any amp with such EVIL 7th harmonic level has loads more 2nd 3rd 4th 5th 6th harmonics.
Whew. Da audio industry is saved and should survive this Millenia too 🙂
But seriously folks, you'd only get such predominantly high order THD with something like really crappy xover.
It came from you in post #284:
... "I don’t think zero global feedback amplifiers should be penalised. Ditto vacuum tube amps. A well engineered 0.5% tube amplifier is still a valid amplifier and should be assessed on its merits within what is possible with the chosen technology."
This is as I see it very subjective approach.
//
I cannot imagine a real amplifier that would have only base frequency and -80dBr 7th harmonic in the spectrum😉. The only way to do it is a summation of two sine waves.
John Curl did establish any occurrence of the 7th as an amplifier evil 🙂.