E) The effects we're trying to measure can't be resolved accurately with the test equipment we are using. Numbers don't tell the story. Input and output must be compaired and any difference quantified ( % ) to get a handle here.
A lot of D) also, but all of the above are true.
Later
Bruce
A lot of D) also, but all of the above are true.
Later
Bruce
Amp measurement parameters
John,
(I changed the subject. This has digressed to a different topic, don't you think?) 😉
Generally, I agree. Measurements should correspond to what one hears. I wonder though how far one goes with measurement to say now its impossible to hear a difference?
For example, I usually replace my op-amps in CD players to BIFET. Both measure in the thousanths when it comes to distortions, but BIFET op-amps sound better than their BJT cousins. Why???
As you ask, are we failing to measure important parameters?
At any rate, all these different things make for a fun hobby, don't they?
Gabe
John,
(I changed the subject. This has digressed to a different topic, don't you think?) 😉
Generally, I agree. Measurements should correspond to what one hears. I wonder though how far one goes with measurement to say now its impossible to hear a difference?
For example, I usually replace my op-amps in CD players to BIFET. Both measure in the thousanths when it comes to distortions, but BIFET op-amps sound better than their BJT cousins. Why???
As you ask, are we failing to measure important parameters?
At any rate, all these different things make for a fun hobby, don't they?
Gabe
SY,
I did the experiment with two different amplifiers. A single ended 300B without NFB, and a push-pull with NFB (didn't expect any difference there). I measure frequency response (not all of us can afford the $30,000+ HP audio analyzers), all other things (audio source, speakers, etc) being equal. No noticable change in frequency response limits (both responses were from 15-30KHz -1dB).
Gabe
I did the experiment with two different amplifiers. A single ended 300B without NFB, and a push-pull with NFB (didn't expect any difference there). I measure frequency response (not all of us can afford the $30,000+ HP audio analyzers), all other things (audio source, speakers, etc) being equal. No noticable change in frequency response limits (both responses were from 15-30KHz -1dB).
Gabe
Sorry to all for the topic drift....
Yes. What I'm trying to get my head round, is the separation of science from mysticism.
When I can measure something, it is generally repeatable, and it's possible to build and improve further.
When I cannot, then I'm lost in a sea of maybe's. Dreams merge with reality, the foundations don't exist. What works today won't tomorrow.
Cheers,
Gabevee,
Yes. What I'm trying to get my head round, is the separation of science from mysticism.
When I can measure something, it is generally repeatable, and it's possible to build and improve further.
When I cannot, then I'm lost in a sea of maybe's. Dreams merge with reality, the foundations don't exist. What works today won't tomorrow.
Cheers,
I'm just curious; can anyone give me an example of two pieces of electronics that have *verifiable* sonic differences that aren't amenable to measurement?
I can't, but the silver wire brigade say it sounds better, the more you have of it.
Yet at audio frequencies there would appear to be no measurable difference.
Before people jump down my throat, I'm not rubbishing the use of silver, what I want is scientific proof.
NB I'm talking of use as conductors, not contacts.
I agree with what has been said about measurements so far.
It's my opinion that we need to look into the more subtle areas to find ways to identify these subtle effects. Directly compair the input and output signals. My feeling is that group delay and rise time issues will show up these subtle changes made when swapping parts. Can I prove any of this? No, my equipment isn't good enough. I can see hints of these things however by placing the input over the output and gain matching the two. If the traces seperate, something changed. More often it's a smear in time, not amplitude that I see. To me when I see less of this on the o-scope the sound improves too.
The signals we use to test with could be part of the problem. Try using pulses, triangle and square waves to show up some of these things. Sign waves just don't do it.
Later
Bruce
It's my opinion that we need to look into the more subtle areas to find ways to identify these subtle effects. Directly compair the input and output signals. My feeling is that group delay and rise time issues will show up these subtle changes made when swapping parts. Can I prove any of this? No, my equipment isn't good enough. I can see hints of these things however by placing the input over the output and gain matching the two. If the traces seperate, something changed. More often it's a smear in time, not amplitude that I see. To me when I see less of this on the o-scope the sound improves too.
The signals we use to test with could be part of the problem. Try using pulses, triangle and square waves to show up some of these things. Sign waves just don't do it.
Later
Bruce
John, that's why I ask the question. One guy says that silver is better, another says it's terrible. One guy says it's audible, but offers no verification, another says it's inaudible, but can't prove a negative. So, in my confused little brain, I want to start with the first basic question.
Bruce, I can see some comonality in our experience:
Yes, I repair a lot of broadcast kit too. But actually, the requirements of an audio buffer are far less stringent (except for s/n) than even a standard definition Video Distribution Amplifier, let alone an HDTV one. Yet it's all so easily done in video. And was in the days of tubes in video too..
1 and 10 T pulses will sort 'em out😎
.
Yes, I repair a lot of broadcast kit too. But actually, the requirements of an audio buffer are far less stringent (except for s/n) than even a standard definition Video Distribution Amplifier, let alone an HDTV one. Yet it's all so easily done in video. And was in the days of tubes in video too..
1 and 10 T pulses will sort 'em out😎
Exactly my point (in November):
Peter Daniel said:
(I detected a certain "heavyness"; or is it paranoia on my part?
I answered:
John;
I agree, If this were video it would be a no brainer, I just think that maybe this is where we need to look for the kind of very subtle effects we're talking about here. Does that sound reasonable to you?
Later
Bruce
I agree, If this were video it would be a no brainer, I just think that maybe this is where we need to look for the kind of very subtle effects we're talking about here. Does that sound reasonable to you?
Later
Bruce
Bruce,
Frankly speaking; If I thought our measurement methods were producing rational results, I'd be in the SS forum, not this one😉
Absolutely. But I can't see the woods for the trees at the moment, and much of the "audio folklore" doesn't help🙄
Frankly speaking; If I thought our measurement methods were producing rational results, I'd be in the SS forum, not this one😉
dhaen said:
John, I don't see any problem with that dual perception. Tube amps have measurable performance differences, at AUDIBLE frequencies, they have vastly different construction, parts values, and different methods of interacting with the outside world than SS amps do. Build a SS amp with output transformers, and you will get a whole lot of the "tube sound". Make it class A, you'll get more, etc.
I hear a big difference between my Pioneer receiver, and my SET! Everyone seems to forget that as they are roasting me alive for saying a $15 paper-in-oil is a marketing success story.
Joel said:
However they are PP and class B
And they are going to go soon. eBay buyers beware 😉
Well Joel, I actually have the misfortune to be using a pair of Quad 50E's in my workshop at present. Actually I hate them🙁
However they are PP and class B
And they are going to go soon. eBay buyers beware 😉
Just a thought here....
I would think that if there were apparent differences here that they would be measureable on a distortion analyzer. Wouldn't the apparent gain in audibility/performance manifest itself by showing a lower distortion rate? If there is a lower harmonic distortion it should be noticeable on a good distortion analyzer.
I have rebuilt several hundred or so SAE 2400's and Ampzilla Solid state power amps just to mention a few. The SAE's have metalfilm resistors throughout. I have just for kicks measured the distortion on my Sound Technology 1710A and 1700B analyzers only to find small differences of around .002 between the metalfilm and common carbon resistors after doing a rebuild.
Show me somewhere in print where the human ear can hear a difference this small. I certainly couldn't hear a difference.
I could be totally wrong here.... As all my experience is with SS and none is with tubes.
Joe
I would think that if there were apparent differences here that they would be measureable on a distortion analyzer. Wouldn't the apparent gain in audibility/performance manifest itself by showing a lower distortion rate? If there is a lower harmonic distortion it should be noticeable on a good distortion analyzer.
I have rebuilt several hundred or so SAE 2400's and Ampzilla Solid state power amps just to mention a few. The SAE's have metalfilm resistors throughout. I have just for kicks measured the distortion on my Sound Technology 1710A and 1700B analyzers only to find small differences of around .002 between the metalfilm and common carbon resistors after doing a rebuild.
Show me somewhere in print where the human ear can hear a difference this small. I certainly couldn't hear a difference.
I could be totally wrong here.... As all my experience is with SS and none is with tubes.
Joe
I think Joe brings up a good point, and one which John alluded to. The differences people have described hearing between caps, and resistors, are obvious to them. They say they can quite clearly hear a difference. Well, I ask you, if it's a large enough change in response/distortion/phase shift, etc, that would HAVE TO manifest itself in a visible way on either a scope or a dist analyzer or both - wouldn't it?
If you are claiming to hear "more bass" with a paper cap, show me a freq response chart that shows less of a low end roll off. If you claim the sound is "too bright" with polypropylene, where is the increase in measurable high end response?
Now, people bring up the famous steve bench website... does a passive component show effects other than pure capacitance or resistance - of course. Some resistors have inductance, etc. Do they show non-linearities? Yes. Can you find minute traces of them in measurements - yes! Can you HEAR them? Well, I just don't think so. If I lay 2 pieces of wire parallel to each other, 100 miles apart, do they have a capacitive effect? Yup, and we can even calculate it on paper - but who cares? So, where does it end?
I think a lot of the stuff the silver-wire crowd talks about would be perfectly true, and perfectly valid in an RF circuit. A paper cap may indeed be lossy in the MHz range. But my home stereo is NOT an RF amplifier, so why would I design it like one?
If you are claiming to hear "more bass" with a paper cap, show me a freq response chart that shows less of a low end roll off. If you claim the sound is "too bright" with polypropylene, where is the increase in measurable high end response?
Now, people bring up the famous steve bench website... does a passive component show effects other than pure capacitance or resistance - of course. Some resistors have inductance, etc. Do they show non-linearities? Yes. Can you find minute traces of them in measurements - yes! Can you HEAR them? Well, I just don't think so. If I lay 2 pieces of wire parallel to each other, 100 miles apart, do they have a capacitive effect? Yup, and we can even calculate it on paper - but who cares? So, where does it end?
I think a lot of the stuff the silver-wire crowd talks about would be perfectly true, and perfectly valid in an RF circuit. A paper cap may indeed be lossy in the MHz range. But my home stereo is NOT an RF amplifier, so why would I design it like one?
I might take some issue with that. If I read you right, you believe that any audible difference MUST be reflected in a measurement difference, presumably using existing measurement techniques.
Now, you MAY be right, but isn't it at least POSSIBLE that there could be some audible phenomenon that is escaping the usual existing measurements? I'm not saying there is one (and my question about verifiable audible differences that can't be conventionally measured remains unanswered), but I'm not sure how you can go so far as to say there can't even BE any as-yet-unmeasured audibly significant phenomenon.
Instruments may be sufficient, but maybe not. The rule of thumb that the output of an electronic device should *sound* like the input when evaluated by a competent set of ears and brain is still a more valid one, I think.
Now, you MAY be right, but isn't it at least POSSIBLE that there could be some audible phenomenon that is escaping the usual existing measurements? I'm not saying there is one (and my question about verifiable audible differences that can't be conventionally measured remains unanswered), but I'm not sure how you can go so far as to say there can't even BE any as-yet-unmeasured audibly significant phenomenon.
Instruments may be sufficient, but maybe not. The rule of thumb that the output of an electronic device should *sound* like the input when evaluated by a competent set of ears and brain is still a more valid one, I think.
Burned, Sy, Joel,
Back in the late 80's, I think actually about 1990, there was a "shootout" between two amplifiers that measured virtually the same. A top of the line solid state and top of the line tube amplifier. They were chosen for their virtually exact specifications. 500 watts, <0.01 % distortions, even the individual harmonic content was very close.
The only actually measured difference was a slight (like +0.2dB) increase in the audio range from 200-400 in the tube amp, and 400-600 in the solid state amp. Otherwise all else was virtually exact (something like 0.5% differences).
The subjective differences is what intrigued me. I had before then also noticed how the sound of SS, while one can make a very incredibly good sounding SS amp, even with off the shelf parts if done right, tended to have a slight boost in the 500 hertz range, regardless of all the techniques to keep things flat. Tube amps OTOH always sounded like they had a boost in the 200-300 Hz range.
This is what the listening test also revealed. This is what the one doing the test heard, and explained it by the 0.2dB boost in the particular range. So yes, there is a definite correlation to a what is heard and the measured response.
However, he also lamented a fact of SS amps at the time. Upon listening to the amp throughout the volume range, he noticed a point that was just right... the "sweet" spot. The tube amp he said, "made the music just sound more and more live as volume was increased."
The point is, our hearing is sensitive enough to notice differences in music with only very slight changes in amplitude within a range. So a capacitor with slight nonlinearities will cause a slight change in the response curve. Not enough to call it non "flat", but enough to be able to say "I like the sound of the amp with this one, but not with that one" due to the way capacitors work.
Another interesting thing: Up until I found the schematic to the old Magnavox amp I love, I had been listening to a variety of P-P circuits using a variety of phase splitters. While the overall frequency response did not change (25-30K) the audible difference was interesting. The concertina phase splitter sounded limited. Warm and pleasant, but high end was kind of dull. One two varieties of paraphase sounded a bit brighter in the high end, and punchier, but too forward for me. The Magnavox circuit made the biggest difference of all. Bass seemed boosted and deeper, and highs seemed much more extended... but the response was still exactly the same! Go figure. All had the same FR range, but had different sonic signatures.
Oh well. I would like to get a complete audio analyzer, which may come soon. Then I can tell you all why these different phase splitter types sound so dramatically different. But for now, it is a mystery. I haven't read anything, except perhaps the reason why an LTP type phase splitter would sound different... cancellation of even order harmonics. But that should have nothing to do with FR contour.
So yes, there is a correlation between what is measured and what is heard to be sure. But it is those things that, all else being equal, we hear that doesn't seem to be measurable that is what is most intrigueing.
Fun stuff!
Gabe
Back in the late 80's, I think actually about 1990, there was a "shootout" between two amplifiers that measured virtually the same. A top of the line solid state and top of the line tube amplifier. They were chosen for their virtually exact specifications. 500 watts, <0.01 % distortions, even the individual harmonic content was very close.
The only actually measured difference was a slight (like +0.2dB) increase in the audio range from 200-400 in the tube amp, and 400-600 in the solid state amp. Otherwise all else was virtually exact (something like 0.5% differences).
The subjective differences is what intrigued me. I had before then also noticed how the sound of SS, while one can make a very incredibly good sounding SS amp, even with off the shelf parts if done right, tended to have a slight boost in the 500 hertz range, regardless of all the techniques to keep things flat. Tube amps OTOH always sounded like they had a boost in the 200-300 Hz range.
This is what the listening test also revealed. This is what the one doing the test heard, and explained it by the 0.2dB boost in the particular range. So yes, there is a definite correlation to a what is heard and the measured response.
However, he also lamented a fact of SS amps at the time. Upon listening to the amp throughout the volume range, he noticed a point that was just right... the "sweet" spot. The tube amp he said, "made the music just sound more and more live as volume was increased."
The point is, our hearing is sensitive enough to notice differences in music with only very slight changes in amplitude within a range. So a capacitor with slight nonlinearities will cause a slight change in the response curve. Not enough to call it non "flat", but enough to be able to say "I like the sound of the amp with this one, but not with that one" due to the way capacitors work.
Another interesting thing: Up until I found the schematic to the old Magnavox amp I love, I had been listening to a variety of P-P circuits using a variety of phase splitters. While the overall frequency response did not change (25-30K) the audible difference was interesting. The concertina phase splitter sounded limited. Warm and pleasant, but high end was kind of dull. One two varieties of paraphase sounded a bit brighter in the high end, and punchier, but too forward for me. The Magnavox circuit made the biggest difference of all. Bass seemed boosted and deeper, and highs seemed much more extended... but the response was still exactly the same! Go figure. All had the same FR range, but had different sonic signatures.
Oh well. I would like to get a complete audio analyzer, which may come soon. Then I can tell you all why these different phase splitter types sound so dramatically different. But for now, it is a mystery. I haven't read anything, except perhaps the reason why an LTP type phase splitter would sound different... cancellation of even order harmonics. But that should have nothing to do with FR contour.
So yes, there is a correlation between what is measured and what is heard to be sure. But it is those things that, all else being equal, we hear that doesn't seem to be measurable that is what is most intrigueing.
Fun stuff!
Gabe
Gabe, how was the listening test set up? The same way as your test of the audibility of a cap substitution?
- Status
- Not open for further replies.