SY said:
This is very perceptive, and a view I haven't seen before. It also helps me better to understand some of the things going on in your mind when you are trying to compare something in the past (an experience) with something in the present (the current perception).
The way you "remember" the past experience is very much influenced by what you are currently experiencing or what you are searching for. Memory isn't absolute; it brings up whatever is important in the current context.
So, if you have say a memory of a birthday party, and someone asks you if Mr X was present, your memory brings up the parts that will allow you to answer the question, like faces you remember or a seating arrangement. It will not bring up the music played at the party, although it may hover at the edges depending how strong the memory and how complete you try to remember the scene. If you have a hard time to remember Mr X's presence or not, you are likely to bring up more and more details looking for clues. This will also take more time, as the brain tries to reconstitute internal sensory mappings and connections corresponding to the remembered scene (because that's basically how memory recall works, you try to reconstitute the brain's state at that time so you can 're-live' the situation).
What all this means is that what you remember is not an absolute truth but biased by what you are looking for and your own current values. If you stated beforerhand that Mr X was present, and someone disagrees with you, you have a big chance that you actually "remember" seeing him, even if he wasn't there. (Like in the ambulance story I told before). I think the implications for subjective audio testing are clear to see.
This is all fairly well established if you study the field, and generally it is assumed that it is evolution's answer on dealing effectively and efficiently with the immense amount of info we are constantly bombarded with.
So, SY, yes, I 100% agree with your observation.
Jan Didden
I have two audio systems having technical values, 20 and 50. At the same time, I also have my mood indication factors: 0.1, 0.2, ..., 1.0. My mood however actually works in this way: (0.1)^2, (0.2)^2, ..., (1.0)^2.
One evening, I was driving my car having car audio of value, 20, and was listening to CD music. That time my mood indication was 0.9. So, total effect of the music was 20 x (0.9)^2 = 16.2. Hmmm...... it was very sweet!!!
Another night, I was at home having audio system of value, 50, and was listening to the same CD. But, that night my mood indication was 0.5. So, total effect of the music was 50 x (0.5)^2 = 12.5.
I could not get the same emotion as I had in the car.Of course, I personally desire both nice audio system and good mood, working together!!!
janneman: In my book, you do that by trying to neither add nor subtract from the musical signal while it travels trough the sytem. I guess we agree on that, the controversy is about how we check that we accomplish this.
Accurate reproduction? Do we agree on that? "What we have here is a failure to communicate", said in my best fake southern (US) accent, while wearing reflective aviator style sunglasses and holding my state issued prison guard rifle. And just what is it we are trying to communicate? Is it faithfull reproduction of a live performance? That might be feasible (almost) now. Here's my suggestion:
Model a human head and ears (mine, of course, since it's entirely representative of all mankind), and place accurate (absolutely accurate, of course) transducers where the tympanic membrane would be (well, you gotta pick somewhere) and place it in the venue (don't ask me annoying questions like; "which section, which row, what am I wearing?".) Now record a series of test sounds to cover the entire audible range (I'll let someone else determine what that is), set up the playback chain (electronics, speakers, room, furniture, drapes, temperature, humidity, etc.), put my modeled head in your favourite seating position and play the recording. Then go back and process the recording so that everything matches, including amplitude, phase, reverberation amplitude and delay (for all frequencies), decay, etc. (insert your favorite esoteric parameter here),. This would, of course, require some sort of surround stuff and could violate the prescription of "nothing added or nothing taken away". But as you see, the question is; "added or taken from what, exactly"?.
For studio recordings, I guess "accurate" would be to follow the same procedure but set up my proxy head in the mixing room in front of the monitors.
That's a definition of an accurate reproduction chain (well, my definition, anyway). Would you like it? It would sure be different from what you're listening to now. You wouldn't have the "detail" you get from most current live recordings.
If we don't like this, or it's impractical, we'll have to settle for the specific innaccuracies that we (I anyway) prefer most.
For the record, I agree with Janneman that we're probably there now with amplifiers (assuming, given Sy's caveat, that we are operating them always within their limits of accuracy). And assuming, of course that "there" has the narrow meaning of output exactly mimics input. And, assuming that we're driving ideal impedences (and whatever assumptions, as required). It strikes me as a perfectly reasonable design goal for the amp to be neutral and add our filigrees elswhere. That said, my favorite amps, in my system, are tubed (they have such a nice glow). As always, I reserve the right to practice hypocrisy, whenever it suits me.
Sheldon
Accurate reproduction? Do we agree on that? "What we have here is a failure to communicate", said in my best fake southern (US) accent, while wearing reflective aviator style sunglasses and holding my state issued prison guard rifle. And just what is it we are trying to communicate? Is it faithfull reproduction of a live performance? That might be feasible (almost) now. Here's my suggestion:
Model a human head and ears (mine, of course, since it's entirely representative of all mankind), and place accurate (absolutely accurate, of course) transducers where the tympanic membrane would be (well, you gotta pick somewhere) and place it in the venue (don't ask me annoying questions like; "which section, which row, what am I wearing?".) Now record a series of test sounds to cover the entire audible range (I'll let someone else determine what that is), set up the playback chain (electronics, speakers, room, furniture, drapes, temperature, humidity, etc.), put my modeled head in your favourite seating position and play the recording. Then go back and process the recording so that everything matches, including amplitude, phase, reverberation amplitude and delay (for all frequencies), decay, etc. (insert your favorite esoteric parameter here),. This would, of course, require some sort of surround stuff and could violate the prescription of "nothing added or nothing taken away". But as you see, the question is; "added or taken from what, exactly"?.
For studio recordings, I guess "accurate" would be to follow the same procedure but set up my proxy head in the mixing room in front of the monitors.
That's a definition of an accurate reproduction chain (well, my definition, anyway). Would you like it? It would sure be different from what you're listening to now. You wouldn't have the "detail" you get from most current live recordings.
If we don't like this, or it's impractical, we'll have to settle for the specific innaccuracies that we (I anyway) prefer most.
For the record, I agree with Janneman that we're probably there now with amplifiers (assuming, given Sy's caveat, that we are operating them always within their limits of accuracy). And assuming, of course that "there" has the narrow meaning of output exactly mimics input. And, assuming that we're driving ideal impedences (and whatever assumptions, as required). It strikes me as a perfectly reasonable design goal for the amp to be neutral and add our filigrees elswhere. That said, my favorite amps, in my system, are tubed (they have such a nice glow). As always, I reserve the right to practice hypocrisy, whenever it suits me.
Sheldon
This is exactly correct. And in the light of evolution, it makes sense- if you don't respond to a pattern, you can be eaten. There's a large evolutionary consequence in this. If you are fooled into reacting to a pattern that's not there, at worst you wasted a little bit of energy responding to a false alarm. So our brains are wonderful in pattern recognition but prone to a certain type of error.
As one of my professors once remarked: " Gettin' eaten is generally considered as a bad career move".
Jan Didden
Jan Didden
Sheldon said:
Sheldon,
Yes, this is all very sensible and I agree to most of it. But the subject of the current debate was whether you can use listeners to measure amp performance. In that context, I defined "faithfull reproduction" as neutral amplification of an input signal. I think that for the present discussion that would be acceptable.
Jan Didden
janneman said:
Agreed. The amp is one place in audio where we can (well, most of us anyway) probably agree on the definition and physically measure accuracy (whether or not our current measurements are up to the task, i.e., measuring the right things, I'll leave for others). The post was just a long sinded way of saying it gets messier from there in and from there out and in these debates we don't even have commonly accepted definitions.
Sheldon
"But the subject of the current debate was whether you can use listeners to measure amp performance. "
I recall a comment in an audio mag in the late 70's to early 80's regarding this with respect to speakers. The comment (as best I recall) was adressing the observation that while the big Japanese manufacturer's built pretty good active electronics, they were a bust when it came to loudspeakers. I was pretty much in agreement at the time since when shopping for new speakers, AR, KLH, KEF, Advent, etc (I would include Klipsch and Allison except they were beyond my budget) seemed to occupy a level of performance that Sony, Panasonic, Technics etc were not even close to.
Anyway, the comment in the magazine observed that a key difference in approach was that the Japanese companies used a consumer listening pannel. They paid consumers to sit in a room and comment on which speakers they liked, the the engineers would fiddle with something and have the pannel rate the sound again. In contrast, the US and British companies had Mr. Vilchur, Mr. Klipsch, Mr. Koss and so on who could care less what a consumer pannel thought because they themselves damn well **knew** what was right.
I'm not sure if this relates to amps, but if subjective listening can work for amp design, it may be critical to consider who is listening. In any case, I just can't envision a Pass, Curl, Carver or even a Self, Slone or Elliot letting a consumer pannel advise them on their amp design.
I recall a comment in an audio mag in the late 70's to early 80's regarding this with respect to speakers. The comment (as best I recall) was adressing the observation that while the big Japanese manufacturer's built pretty good active electronics, they were a bust when it came to loudspeakers. I was pretty much in agreement at the time since when shopping for new speakers, AR, KLH, KEF, Advent, etc (I would include Klipsch and Allison except they were beyond my budget) seemed to occupy a level of performance that Sony, Panasonic, Technics etc were not even close to.
Anyway, the comment in the magazine observed that a key difference in approach was that the Japanese companies used a consumer listening pannel. They paid consumers to sit in a room and comment on which speakers they liked, the the engineers would fiddle with something and have the pannel rate the sound again. In contrast, the US and British companies had Mr. Vilchur, Mr. Klipsch, Mr. Koss and so on who could care less what a consumer pannel thought because they themselves damn well **knew** what was right.
I'm not sure if this relates to amps, but if subjective listening can work for amp design, it may be critical to consider who is listening. In any case, I just can't envision a Pass, Curl, Carver or even a Self, Slone or Elliot letting a consumer pannel advise them on their amp design.
Sam9,
I just read exactly that in an issue of Multi Media Manufacturer (may/june 2004), from Larry Klein recounting his 45-year experience of audio and electronics. He felt that at the time Japanese manufacturers thought speakers needed to be designed to taste. They did large-scale scored listening tests using assembly line people, maintenance men, clerks, a "carefully selected listener cross section". Their resulting speakers sounded awfull.
Jan Didden
I just read exactly that in an issue of Multi Media Manufacturer (may/june 2004), from Larry Klein recounting his 45-year experience of audio and electronics. He felt that at the time Japanese manufacturers thought speakers needed to be designed to taste. They did large-scale scored listening tests using assembly line people, maintenance men, clerks, a "carefully selected listener cross section". Their resulting speakers sounded awfull.
Jan Didden
There seems little surprise that these are audible differences between loudspeakers. The measured performance of different loudspeakers vary greatly with harmonic distortion in the range of < 0.1% – >5 % and wide variations in frequency response. With in this context the differences between amplifiers is minute.
Stuart
Stuart
One thing puzzles me; If you can hear differences between amplifiers, and judge which sounds better, why is it so difficult to set up a loudspeaker crossover by ear alone?
You know what sounds "right", right?, so why do we need to measure for something that has a far greater influence on the sound than , say, the type of cap used in the feedback loop?
You know what sounds "right", right?, so why do we need to measure for something that has a far greater influence on the sound than , say, the type of cap used in the feedback loop?
johnnyx said:
...but can you? Why are you so sure about that? Do they actually sound different, or do you perceive one as being different. If you do, where is that perception based on? Sound, knowing the brand, seeing the amp, the story from your buddy about this great amp, the effort you had to make to get one of your own, all of the above? Why is it that blind tests mostly fail to substantiate these differences? I know, mostly questions, but they are necessary before getting to the answers.
Jan Didden
johnnyx said:One thing puzzles me; If you can hear differences between amplifiers, and judge which sounds better, why is it so difficult to set up a loudspeaker crossover by ear alone?
Because the amps are finished products.
The speaker, it seams that you are finishing, or "setting up".😀
Speakers are different.
5 cm out of the right place and they can sound bad.
Differences in the sound produced by different amps can be huge, even a serious blind test should reveal that.
With experienced listeners, otherwise the difference is there and people can't identify it.
BUT different sound just means better for one listener and not necessarilly so for the other one.
That's why audio is subjective.
A few years ago I bought a new graphics card for my PC. In the various hardware forums I frequented at the time there were plenty of people offering opinions on this card. One of the most notable ones was that 2D image quality was superior. On installing the card the image was indeed superior; it was brighter, more vibrant.
Some time later I was faced with the opportunity of comparing it with my old card side-by-side on identical monitors. What did I see? There was absolutely no perceivable difference between them! I would consider myself to be as objective as the next person, but still I fell foul of purely psychological effects of perception.
This effect makes it very hard for any normal subjective measurement to be taken at face value. It's not possible to compare two sound sources by listening to them both simultaneously of course, so the only way to eliminate it is with double-blind testing, which seems unfortunately rare.
Of course if something subjectively sounds bad despite good objective measurements then it is reasonable to consider the subjective measurements as very important, but if they aren't double-blind then I would consider them suspect.
Some time later I was faced with the opportunity of comparing it with my old card side-by-side on identical monitors. What did I see? There was absolutely no perceivable difference between them! I would consider myself to be as objective as the next person, but still I fell foul of purely psychological effects of perception.
This effect makes it very hard for any normal subjective measurement to be taken at face value. It's not possible to compare two sound sources by listening to them both simultaneously of course, so the only way to eliminate it is with double-blind testing, which seems unfortunately rare.
Of course if something subjectively sounds bad despite good objective measurements then it is reasonable to consider the subjective measurements as very important, but if they aren't double-blind then I would consider them suspect.
janneman said:
I was asking the question to other forum posters. I haven't compared the sound of different amplifiers, but I have compared CD players when I was buying a replacement. One sounded worse than my present one and one sounded better. I could only tell when there was only a few seconds between hearing each one, because differences get lost in the mists of time.
carlosfm said:
None of my amps are finished products😀
It's difficult because I have been building active systems, I asked the question because of the difficulty of setting up the crossover slopes, frequencies, etc by ear, so I have to use a CAD measuring system to help.
Yet this thread is about (as I understand it) differences in amplifiers which cannot be measured. I just find it strange.
johnnyx said:
Huh? Maybe next time you say: "Question for the other posters except janneman"😱
Saves us all some time...
Jan Didden
Two interesting points here- first, this sounds like a classic case of a demo being set up without proper level-matching. That may or may not be true, I'm making an assumption, but so many audio legends sart out with just that sort of uncontrolled variable.
Second, your point about delay is startlingly pertinent. Claims about long-term sensory tests have not stood up to the data- we seem, in all of our sensory inputs, not just music-listening, to be most sensitive to differences when we can make comparisons with a minimum of delay. This is just as true for wine-tasting, haptics, and color-matching as it is for music-listening. I mean, if you were asked to compare the weights of two objects that differed by a few percent, what's more accurate, holding one in your hand, then switching to the other, or holding one in you hand all day one day, then holding the other in your hand all day the next day?
janneman said:
I didn't mean to exclude you, Jan, I meant all others, you included, and not me. If that makes sense.I was careful that the CD player differences weren't due to a level difference. They were cheap players, using different DACs, and differed in distortion. This was audible when played side by side.
Try looking at one side of a "spot the difference " cartoon, the following day look at the other side. How many differences can you spot now?😀
How did you do this? Level matching to the degree necessary to prevent contamination of the test is VERY tricky. And what makes it insidious is that a tiny level difference (like 0.1 dB) will not be audible to a good ear as a level difference, but rather as a quality difference. A test disc, a multimeter, and the CD player volume control are NOT sufficient to do proper level-matching.
Your cartoon analogy is great. May I steal it?
- Status
- Not open for further replies.