Perhaps some recall Jakob's comment to the effect that if only as much money were used for auditory research as the food industry spends on its research, then most of the disagreements here would likely be settled.
It may be easy to imagine an assignment to research something as a task that shouldn't be too hard to do, but people systematically tend to underestimate how long a project should take and how much resources will be needed to get it done. In particular, it might not be so easy to figure out why some piece of equipment sounds the way it does aside from existing measurements. First, one should probably be equipped to do all existing measurements and prove that they don't account for the effect to be researched. Then some underlying cause must be found and it could turn out there are multiple underlying causes. It may be necessary to device and prove the effectiveness of multiple tests. Then there is the matter of credibly demonstrating audibility in humans.
Even Earl Geddes said a preliminary study of what he considered the unstudied 5% or so of the population for auditory perception would probably cost several tens of thousands of dollars. He said that new tests would have to be developed, etc. And a preliminary study is not something designed to provide proof, it would be to provide an estimate how much more time and money would be needed.
It doesn't sound like simple homework for retired people to me.
You can judge your audio system all day and night as you want. No one will challenge you for that. But what about posting things online?
This has what to do with sound replying electronics?
When investigating why two DACs sound different, one must first confirm that they do indeed sound different when objectively compared. Have you done that?
You obviously don't spend any time around audiophiles, or audiophile products.
The key to understanding why electronics seems as important as it is, appears to derive from the use of Global Negative Feedback. It is problematic.
Now, where did I get this revelation?
It so happens that back in 1971-72, I was visiting some audio friends in Santa Barbara, one person you might recognize was Bascom King, a fellow audio designer. We asked each other the same question: WHY is electronics so critical when loudspeakers have so much more MEASURED distortion? We decided then to visit the Richard Heyser at his house, and we made the journey to his place, to ask him about this question. Now, at this time Richard Heyser, then with JPL, was one of the most acknowledged experts in audio design, and had written several papers at the time for the AES.
When we asked him, he said it was GLOBAL NEGATIVE FEEDBACK in electronics that made the difference. He also said that more that 10 years before, he had accidentally found this out, by testing an open loop design, that he had made for JPL (to go to the moon) on his K-horn loudspeaker, just for the heck of it, and was astounded what he heard. Richard Heyser actually had given a paper about his rather unique amplifier at the AES and it can still be found in the archives. Actually it is astounding that this amplifier of his (made for another purpose) even worked at all with the K-horn, but it did and many people at the time (around 1960) tried it and even one guy took Richard's circuit and built it to sell for a time, because it sounded so good.
The KEY once again, is NO GLOBAL NEGATIVE FEEDBACK in the amplifier. Trust me, this boggled both Bascom King (a guy with and engineering degree and lots of audio design experience at the time) and me (working on designing amps for the previous 4 years), because it was such a radical concept.
Of course, around that time, Matti Otala had suggested to reduce negative feedback in order to increase open loop bandwidth, and that might have made the biggest correction.
Richard Heyser had not looked that far, just his experience with his open loop amp and the commercial stuff out there. This is where I learned to use negative feedback carefully, instead of with abandon, like so many others.
Now, where did I get this revelation?
It so happens that back in 1971-72, I was visiting some audio friends in Santa Barbara, one person you might recognize was Bascom King, a fellow audio designer. We asked each other the same question: WHY is electronics so critical when loudspeakers have so much more MEASURED distortion? We decided then to visit the Richard Heyser at his house, and we made the journey to his place, to ask him about this question. Now, at this time Richard Heyser, then with JPL, was one of the most acknowledged experts in audio design, and had written several papers at the time for the AES.
When we asked him, he said it was GLOBAL NEGATIVE FEEDBACK in electronics that made the difference. He also said that more that 10 years before, he had accidentally found this out, by testing an open loop design, that he had made for JPL (to go to the moon) on his K-horn loudspeaker, just for the heck of it, and was astounded what he heard. Richard Heyser actually had given a paper about his rather unique amplifier at the AES and it can still be found in the archives. Actually it is astounding that this amplifier of his (made for another purpose) even worked at all with the K-horn, but it did and many people at the time (around 1960) tried it and even one guy took Richard's circuit and built it to sell for a time, because it sounded so good.
The KEY once again, is NO GLOBAL NEGATIVE FEEDBACK in the amplifier. Trust me, this boggled both Bascom King (a guy with and engineering degree and lots of audio design experience at the time) and me (working on designing amps for the previous 4 years), because it was such a radical concept.
Of course, around that time, Matti Otala had suggested to reduce negative feedback in order to increase open loop bandwidth, and that might have made the biggest correction.
Richard Heyser had not looked that far, just his experience with his open loop amp and the commercial stuff out there. This is where I learned to use negative feedback carefully, instead of with abandon, like so many others.
Last edited:
Define measures well... My criteria might be a bit different than yours. A THD plot doesn't tell me anything unless something has distortion that well above the decimal place. I means why do we assume that something that is the difference of a hundredth to a thousands THD measurement is going to be more audible than all the other parts objectionist claim have no effect??? That's like discussing how much foreign matter is in a paint supply for a wall, 1 or 2 human hairs, and forgetting that there are nearly endless color options.
You should just go read the closed JCBT II thread... Someone has already stated everything you say. You're years behind. You're proposing that a DBT be done. We all know it will end in pointlessness. Now in the old thread we've discussed why numerous times, that it might not work as a test. At this point people that think like you are a huge hold up in the industry. The only ones advancing anything are the ones butchering good measurements - and that's very limited because the gains are small before everything falls apart.
It's like this... Everyone is a bunch of monkeys. The objectionist monkeys are busy defending their territory from the subjectionist monkeys. But they're so offended they continually attack the sujectionist monkeys because their morals are offended everytime they see them. But the subjectionist are busy ripping their land to pieces in search of something new - they mutilate it all day long, provoking objectionst attacks. But the irony is both territories border each other and everywhere else around them is open land. Both sides are afraid to venture into the open land in fear that someone from the other side will come into their existing land. They'd rather remain petty and focused on their border than expand into new land; to make damn sure the other doesn't encroach on existing land.
One must've been hi-fi DAC and the other must've been lo-fi DAC.
Define hi-fi and lo-fi. Hi-fi is a bad word among audiophiles. But since it seems you've just come around to all of this you're not aware of that... Lo-fi is bad word, too. And no, the answer is no. You must have missed some of the best engineers in the world discussing that the DAC3 sounds different than the other same measuring, same chip, based DACs.... just a few pages before these ones. Oh wait, you don't believe anything anyone says. Go make your own company and tell us about all the rave reviews and sales you get. Thx.
Not exactly. As Earl Geddes has pointed out, speaker distortion is generally low order because of their mechanical nature. Amplifier distortion can be high order which humans are much more sensitive to and are more likely find objectionable. The idea that significant low order distortion can mask much lower levels of high order distortion is not necessarily correct.
^^^^
It's true.
But I'd say the type of global feedback matters, too. There aren't many designs using it that are particularly good. (I've heard it sound good, too) But even so, the production of music basically makes it extremely problematic because the media we have to listen to isn't made to sound good with lots of global feedback in general. Even if it's more accurate, people don't like the sound. Also a huge problem is what it does to speakers... it often makes the separate drivers sound like separate speakers. People choose 1st order crossovers to avoid this issue, even when they don't have high global feedback amps.
Recommended Components: 2019 Edition Power Amps | Stereophile.com
Class A amplifiers with HUGE amounts of GNFB in there as well as some with very little. And yet all recommended.
Class A amplifiers with HUGE amounts of GNFB in there as well as some with very little. And yet all recommended.
Neither was my reply rhetorical or trying to be funny, as you may have thought. You are best positioned to answer your own question.
Let us accept this hypothesis. (That I do not as such).
What is the difference between local and global feedback, appart a question of phase at HF ?
What the hell with GFB when slew rate is > 1000V/µs and phase margin big enough ?
I tend to prefer the so called current feedback topology. The main difference with so called voltage feedback is long tailed pair add a transistor in the feedback loop with (usually, for thermal stability) a high impedance feedback bridge and the base parasitic cap. Now, one thing to compare should be an inverting amplifier using the - input: nothing in between. Signal and feedback are added with only passive network.
The problem, now, is to keep the feedback resistance low enough, so this amp should be driven at very low impedance.
Those are open minded questions, not controversy.
Actually I agree with your overall assessment. I was merely pointing out a widespread illogical practice of citing low distortion figure of the electronics as the distortion of the sound perceived by a listener.
Measurement by PMA showed 0.5 - 1% distortion for a decent quality speaker at ~88 dB(?), nowhere near extremely low.
How would spending time with them makes any difference on this matter?
I only mean a "distortion of sound" should refer to a measure on distortion of the sound itself instead of anything else.
Pointlessness does not fit in general; a DBT is just part of the toolbox, if used correctly it provides some usefull insights.
Unfortunately it is often used in a way that does not allow to draw further conclusions from the results and that in fact is often pointless.
Sorry, but my question was:
"I am interested to learn what kind of "proof" or evidence you´d consider as sufficient even if the evidence contraditcs your prior belief."
As it is about what you´d consider as sufficient, you are the only person able to answer it.
Many, many devices measure so low as to be way, way below the expected audibility. Loudspeakers definitely dominate by orders of magnitude. You need to show an amazing level of sensitivity to high order distortion, or a new mechanism altogether.
You left out third monkey, sellers, a.k.a. shills.To see what it's like to sell audio electronics? No need to because I can already see that on this forum.
When Richard Heyser first discovered problems with negative feedback, tube electronics was almost entirely dominant, and virtually everything solid state was Germanium. Generally, power amps back in the early 1970's, more than 10 years after Heyser's discovery, and when he talked to us, had a slew rate of 10-30VuS. Open loop bandwidth, who knows? Even Otala had not published his paper on amps developed at Philips Research Labs, showing us an alternate approach to amp design.
- Status
- Not open for further replies.