By the way, I use SY's bypass test myself - and I bypass the volume control while I'm at it. I found this out the hard way when twiddling around with various linestages, and making the dismaying discovery that the stupid volume controls were audible as well. They all sounded different, and none sounded like the output of the Monarchy DAC (Burr-Brown PCM63K converters) going straight into the power amp with zero attenuation. Not a happy moment. For a while there, I was importing the DACT products in the USA, which had less coloration than most others. But when I heard transformer volume controls, well, that was the way forward, despite unsettling measurements at the extreme HF (above 50 kHz).
All of this based on direct A/B bypass tests from the DAC into the power amp - the DACT was close but not identical, and most of the TVC's were inaudible.
The part of the A/B protocol I find most interesting is the second transition, from B -> A, when doing an A -> B -> A comparison. The A -> B transition is usually not that interesting, sounding just different (if there is a difference, that is). But the quick return back the original sheds light on whether the difference was an improvement or a degradation.
I don't like the fool-the-listener randomized guessing game. In my experience, that merely causes stress, listening fatigue, decreased acuity, and a disengagement from emotional responses (due to test stress). Decreased acuity completely defeats the point of any kind of subjective comparison - if you're going to stress out the listener, might as well just give up and listen to MP3's on an iPod.
Gary Pimm showed me the most acute subjective comparisons, with his A -> B -> A protocol and rapid, noiseless switching at the direct control of the listener. This protocol can rapidly zero-in on subjective results of circuit changes:
1) no result
2) results so small it doesn't matter
3) differences that are clearly audible but are merely different with no subjective preference
4) differences that result in clear and obvious subjective preference
5) and the most comical of all, differences that are so big you hear them before you return to your seat, or are audible from another room and elicit an unsolicited comment of "what was THAT?"
Gary, John Atwood, and I are in agreement the only stuff worth pursuing is category 5, and occasionally and if it's really easy, category 4. The rest of the stuff is a waste of time, the familiar terrain of audiophile tweaking. Who cares about that except the sad tweaking crowd.
So if DHT's elicit a category 5, well guys, go for it!!! But keep it clean with direct comparisons when possible.
Regarding the previous post about DHT's vs IDHT's, it's merely a compilation of conversations I've had with John Atwood and Gary Pimm about how these gizmos work. John A has a huge collection of original RCA manufacturing notes, and these specifically discuss arcane issues like emission-enhancing residues that end up on the grid as a result of manufacturing processes. Electrons from the space-charge cloud do bounce off the cathode, and transient thermal cooling does occur. All of these effects are real.
Anyone that's worked in manufacturing (John, Gary, and myself) knows that physics trumps measurements, for the simple reason that black-box measurements are a very incomplete description of how real devices work, and more importantly, how they fail. Failure mechanisms shed a lot of light on how the internal physical systems are working. If you know the underlying physics and manufacturing processes, you can operate devices well outside factory specs - and also know which factory specs are in fact dangerously misleading. If you know the device physics and manufacturing processes, you can get both better performance and a much longer operating lifetime.
But in turn, measurements always trump models. Models are basically a joke, things that only exist in a non-physical mathematical fantasy world. Yes, they more or less describe the bare-minimum parameters of the device, and are handy for back-of-the-envelope calculations, but you can get ugly surprises when the real device fails unexpectedly, or goes into strange undocumented modes. That's why manufacturers always provide a fair amount of slop in the nominal specifications, so lazy engineers don't get in trouble with application-cookbook designs. You'd be surprised how many consumer devices have circuits that are right out of the device manufacturers application handbook.
Listening? I'd say it correlates most closely with physics, less closely with the usual set of standard measurements, and not much at all with models.
All of this based on direct A/B bypass tests from the DAC into the power amp - the DACT was close but not identical, and most of the TVC's were inaudible.
The part of the A/B protocol I find most interesting is the second transition, from B -> A, when doing an A -> B -> A comparison. The A -> B transition is usually not that interesting, sounding just different (if there is a difference, that is). But the quick return back the original sheds light on whether the difference was an improvement or a degradation.
I don't like the fool-the-listener randomized guessing game. In my experience, that merely causes stress, listening fatigue, decreased acuity, and a disengagement from emotional responses (due to test stress). Decreased acuity completely defeats the point of any kind of subjective comparison - if you're going to stress out the listener, might as well just give up and listen to MP3's on an iPod.
Gary Pimm showed me the most acute subjective comparisons, with his A -> B -> A protocol and rapid, noiseless switching at the direct control of the listener. This protocol can rapidly zero-in on subjective results of circuit changes:
1) no result
2) results so small it doesn't matter
3) differences that are clearly audible but are merely different with no subjective preference
4) differences that result in clear and obvious subjective preference
5) and the most comical of all, differences that are so big you hear them before you return to your seat, or are audible from another room and elicit an unsolicited comment of "what was THAT?"
Gary, John Atwood, and I are in agreement the only stuff worth pursuing is category 5, and occasionally and if it's really easy, category 4. The rest of the stuff is a waste of time, the familiar terrain of audiophile tweaking. Who cares about that except the sad tweaking crowd.
So if DHT's elicit a category 5, well guys, go for it!!! But keep it clean with direct comparisons when possible.
Regarding the previous post about DHT's vs IDHT's, it's merely a compilation of conversations I've had with John Atwood and Gary Pimm about how these gizmos work. John A has a huge collection of original RCA manufacturing notes, and these specifically discuss arcane issues like emission-enhancing residues that end up on the grid as a result of manufacturing processes. Electrons from the space-charge cloud do bounce off the cathode, and transient thermal cooling does occur. All of these effects are real.
Anyone that's worked in manufacturing (John, Gary, and myself) knows that physics trumps measurements, for the simple reason that black-box measurements are a very incomplete description of how real devices work, and more importantly, how they fail. Failure mechanisms shed a lot of light on how the internal physical systems are working. If you know the underlying physics and manufacturing processes, you can operate devices well outside factory specs - and also know which factory specs are in fact dangerously misleading. If you know the device physics and manufacturing processes, you can get both better performance and a much longer operating lifetime.
But in turn, measurements always trump models. Models are basically a joke, things that only exist in a non-physical mathematical fantasy world. Yes, they more or less describe the bare-minimum parameters of the device, and are handy for back-of-the-envelope calculations, but you can get ugly surprises when the real device fails unexpectedly, or goes into strange undocumented modes. That's why manufacturers always provide a fair amount of slop in the nominal specifications, so lazy engineers don't get in trouble with application-cookbook designs. You'd be surprised how many consumer devices have circuits that are right out of the device manufacturers application handbook.
Listening? I'd say it correlates most closely with physics, less closely with the usual set of standard measurements, and not much at all with models.
Lynn Olson said:
Thanks Lynn. Do you think that there is still potential to develop some new measurement approach that could correlate more with what audio components are actually made for? I.e. Listening. Or simply the domains you mentioned are saturated for knowledge in audio by now, and simply will never unify?
Lynn Olson's model for audio testing
What Lynn describing in medical research is a cross over study, and if both the subject and the controller are blinded, it is a double blind cross over study.
I would extend it further to A->B->A->B->A->B
If you randomize the sequence, we have a double blind randomized cross over study.
Each time there is a change, the reviewer should write down the magnitude of the change using the scale 5 as described by Lynn.
I suspect the test retest reliability will be quite high, though I am not sure about the inter-observer difference.
Can we ask a psychologist or a statistician to set up a test bed and recruit 30 odd guys in the next occasion like ETF to validate the above test?
Then we can have a good protocol for testing in audio, at least to report the magnitude of changes, which we can then correlate with measurements.
Sorry for being too off topic.
What Lynn describing in medical research is a cross over study, and if both the subject and the controller are blinded, it is a double blind cross over study.
I would extend it further to A->B->A->B->A->B
If you randomize the sequence, we have a double blind randomized cross over study.
Each time there is a change, the reviewer should write down the magnitude of the change using the scale 5 as described by Lynn.
I suspect the test retest reliability will be quite high, though I am not sure about the inter-observer difference.
Can we ask a psychologist or a statistician to set up a test bed and recruit 30 odd guys in the next occasion like ETF to validate the above test?
Then we can have a good protocol for testing in audio, at least to report the magnitude of changes, which we can then correlate with measurements.
Sorry for being too off topic.
Measurement is not the problem, and least not the physical end of it. It's the software used to display it. Our ear/brain is designed to interpret sounds, our eye/brain is designed to interpret visual images. The problem comes in trying to "hear" with our eyes. Think about trying to see with our ears. This problem is not easily solved for something as complex as timbre.
Lynn, nice to see you here, and I enjoy your many contributions. I have to say in this case, that the space charge connection with timbre would not be my first choice to investigate. Why look in the next town over when what you seek may be across the street?
Let's look at this issue of timbre (or characteristic sound, if you will). If, as Andy suggest, the simple expedient of adding a DHT linestage to a chain with many amplification stages and sub stages, makes a critical difference in the character of the sound, it's highly likely that something is being added. Substituting a purer' amplification at that stage only would not lead to the results he has observed. I do take some issue with sighted tests, as suggestion is a universally accepted principle in cognitive science. Nevertheless, I'm willing to accept the general nature of his claims regarding tonal differences.
As I stated earlier, the very notion that the "pure" DHT is adding something (not distortion as we usually think of it) raises hackles on some, as if it insulted the gods of music. But it shouldn't. How would you like to do your listening in an anechoic chamber? Flat, lifeless, dull? But site the speakers properly in a good room and the music becomes more alive. Well, has the room taken away something, or has it added something? It hasn't distorted the sound in the usual sense, but it has added reflection and interactions that our ears/brain is used to hearing in conjunction with musical performances. In a sense, it makes what we hear more "real". The point is, that what's been added is something quite close - in character- to what occurs in a performance venue. It's certainly closer than we'd hear in very dead room. So doesn't that make it more accurate, in a cognitive sense?
Now let's look at the physical bits that help us hear a recording. First, let me say that pointless ill thought out analogies exasperate me - shiny capacitors sound bright, oil caps sound liquid, thin wire sounds brittle, yada, yada. However, in our case, there are some very real analogies. Andy, you are most impressed with the lifelike rendering of string instruments, which all have - well, strings, of course. Now strings all have a characteristic sound. They sound like strings. They physically vibrate in very characteristic ways, as opposed to bells, for instance. Ever thump on a DHT. The ones I've tried all ring in a pretty musical way. They have strings for filaments - thin strings. If vibration of those strings modulates the output, and of course it does, then we'd expect that added little bit to be in character, or consistent in timbre, with the instruments we recorded, particularly strings. Simple to test. Whack on the few DHT's and IDHT's and see which ring most, and more importantly, sound most like music. Heck maybe thoriated filaments sound better (less damped) than similar coated filaments. This may not, and probably doesn't explain things entirely, if at all, but it sure is the first place I'd look.
What else do stringed instruments have? They all have sound board of some sort, and usually wood. What does wood sound like when you strike it? Well, it sounds like wood. Everyone can agree on that. But can you measure that? Try picking that out of the graph on a consistent basis. What do speakers have, but diaphragms. Let's take a page or three out of Lynn's book. If you tap on a paper cone, does it sound different that a metal cone? You betcha. Does a well damped cone sound different from a light paper cone? Do I have to answer that, senator? Physically, a well damped cone will add less to the signal than a more "lively" cone, much as our anechoic chamber adds nothing to the speaker output. But does it sound more or less like the instrument it's supposed to reproduce? So which is more accurate, in the cognitive sense? Which will make a sound more like a stringed instrument? Hmm, how about a brass ensemble? Maybe metal cones and horns would be more realistic than paper cones and direct radiators.
Sheldon
Lynn, nice to see you here, and I enjoy your many contributions. I have to say in this case, that the space charge connection with timbre would not be my first choice to investigate. Why look in the next town over when what you seek may be across the street?
Let's look at this issue of timbre (or characteristic sound, if you will). If, as Andy suggest, the simple expedient of adding a DHT linestage to a chain with many amplification stages and sub stages, makes a critical difference in the character of the sound, it's highly likely that something is being added. Substituting a purer' amplification at that stage only would not lead to the results he has observed. I do take some issue with sighted tests, as suggestion is a universally accepted principle in cognitive science. Nevertheless, I'm willing to accept the general nature of his claims regarding tonal differences.
As I stated earlier, the very notion that the "pure" DHT is adding something (not distortion as we usually think of it) raises hackles on some, as if it insulted the gods of music. But it shouldn't. How would you like to do your listening in an anechoic chamber? Flat, lifeless, dull? But site the speakers properly in a good room and the music becomes more alive. Well, has the room taken away something, or has it added something? It hasn't distorted the sound in the usual sense, but it has added reflection and interactions that our ears/brain is used to hearing in conjunction with musical performances. In a sense, it makes what we hear more "real". The point is, that what's been added is something quite close - in character- to what occurs in a performance venue. It's certainly closer than we'd hear in very dead room. So doesn't that make it more accurate, in a cognitive sense?
Now let's look at the physical bits that help us hear a recording. First, let me say that pointless ill thought out analogies exasperate me - shiny capacitors sound bright, oil caps sound liquid, thin wire sounds brittle, yada, yada. However, in our case, there are some very real analogies. Andy, you are most impressed with the lifelike rendering of string instruments, which all have - well, strings, of course. Now strings all have a characteristic sound. They sound like strings. They physically vibrate in very characteristic ways, as opposed to bells, for instance. Ever thump on a DHT. The ones I've tried all ring in a pretty musical way. They have strings for filaments - thin strings. If vibration of those strings modulates the output, and of course it does, then we'd expect that added little bit to be in character, or consistent in timbre, with the instruments we recorded, particularly strings. Simple to test. Whack on the few DHT's and IDHT's and see which ring most, and more importantly, sound most like music. Heck maybe thoriated filaments sound better (less damped) than similar coated filaments. This may not, and probably doesn't explain things entirely, if at all, but it sure is the first place I'd look.
What else do stringed instruments have? They all have sound board of some sort, and usually wood. What does wood sound like when you strike it? Well, it sounds like wood. Everyone can agree on that. But can you measure that? Try picking that out of the graph on a consistent basis. What do speakers have, but diaphragms. Let's take a page or three out of Lynn's book. If you tap on a paper cone, does it sound different that a metal cone? You betcha. Does a well damped cone sound different from a light paper cone? Do I have to answer that, senator? Physically, a well damped cone will add less to the signal than a more "lively" cone, much as our anechoic chamber adds nothing to the speaker output. But does it sound more or less like the instrument it's supposed to reproduce? So which is more accurate, in the cognitive sense? Which will make a sound more like a stringed instrument? Hmm, how about a brass ensemble? Maybe metal cones and horns would be more realistic than paper cones and direct radiators.
Sheldon
rdf said:
The system might be, but the electronics aren't. If I build, for example, EQ into my preamp to compensate for frequency response deficiencies in my speakers, I will absolutely get a more accurate system. But the preamp isn't "accurate"; it's an EQ box. Nothing wrong with that, and indeed I use EQ in my system (see my article for NHT M3.3 mods). But the EQ box is not an accurate preamp, and if I stuck it in your system or tried comparing it to preamps that are flat or have different EQ built in, it would not be a happy thing.
Re: Re: Lynn Olson's model for audio testing
When we do subjective tests we should always be double blinded, otherwise the big factor of placebo will overwhelm every other factor, I can reassure that.
I work in the medical field and ABX studies are a very common tool in investigating effects of drugs, devices and even surgery. Evaluating symptoms like dizziness, headache and tinnitis are no different than audio. Asking a question of "Does this medicine make you feel less dizzy?" is similar to asking "Does this cable make you hear more depth of the sound stage?"
I cannot understand why guys in cable asylum never approved double blinded ABX test.
Being off topic again, I once had a patient whom I tried every many drugs to control his vertigo with no help. So I was fed up when I came back and I opened my desk drawer, where I found some old samples of a drug that never been proven to work in randomized control trials and nobody in my unit prescribed it. I gave it to the patient and send him away. He came back next time and thanked me, saying that the medicine was wonderful symptoms all gone. He said the medicine must be a wonderful drug, because I locked it up in the drawer, it must be a precious expensive medicine that I reserved for VIPs!
rdf said:
When we do subjective tests we should always be double blinded, otherwise the big factor of placebo will overwhelm every other factor, I can reassure that.
I work in the medical field and ABX studies are a very common tool in investigating effects of drugs, devices and even surgery. Evaluating symptoms like dizziness, headache and tinnitis are no different than audio. Asking a question of "Does this medicine make you feel less dizzy?" is similar to asking "Does this cable make you hear more depth of the sound stage?"
I cannot understand why guys in cable asylum never approved double blinded ABX test.
Being off topic again, I once had a patient whom I tried every many drugs to control his vertigo with no help. So I was fed up when I came back and I opened my desk drawer, where I found some old samples of a drug that never been proven to work in randomized control trials and nobody in my unit prescribed it. I gave it to the patient and send him away. He came back next time and thanked me, saying that the medicine was wonderful symptoms all gone. He said the medicine must be a wonderful drug, because I locked it up in the drawer, it must be a precious expensive medicine that I reserved for VIPs!
short cut for long test protocols
situation 1: If you start to feel fatigue and you get a bad temper after some listening than you have to tweak again
situation 2: If you feel that you should dance or jump around and you are getting cheerful than leave the stuff as it is. For classical listeners: if you suddenly discover that you are waiving your arms and directing a virtual orchestra you are in situation 2. Pop and blues listeners: playing an "air guitar" ? No, you are not getting nuts but you are in situation 2. No more tweaking to your amps.
situation 1: If you start to feel fatigue and you get a bad temper after some listening than you have to tweak again
situation 2: If you feel that you should dance or jump around and you are getting cheerful than leave the stuff as it is. For classical listeners: if you suddenly discover that you are waiving your arms and directing a virtual orchestra you are in situation 2. Pop and blues listeners: playing an "air guitar" ? No, you are not getting nuts but you are in situation 2. No more tweaking to your amps.
Re: Re: Re: Lynn Olson's model for audio testing
It's not the blind aspect, it's the 'identity' aspect. If the subject claims to hear an audible difference between two choices, the claim is either right or wrong. Asking to further determine if X is A or B adds complexity, difficulty and stress but no further useful information. And while there is a work around for subjects, it requires prior knowledge of the ABX test protocol itself and is of course an added hurdle bringing no value to the results.
rtsang said:
It's not the blind aspect, it's the 'identity' aspect. If the subject claims to hear an audible difference between two choices, the claim is either right or wrong. Asking to further determine if X is A or B adds complexity, difficulty and stress but no further useful information. And while there is a work around for subjects, it requires prior knowledge of the ABX test protocol itself and is of course an added hurdle bringing no value to the results.
If the subject has control of the switching in an ABX test he/she can convert each trial into a same or different one. No need to listen to both choices, just switch between A and X and guess "same" or "different." Completely equivalent.
I'm still waiting to see some evidence about all the "stress" and "confusion" claims, which don't seem to affect any other sensory work- including wine tasting.
I'm still waiting to see some evidence about all the "stress" and "confusion" claims, which don't seem to affect any other sensory work- including wine tasting.
We seem to have gone off onto ABX testing, which is perfectly fine with me, but I'd just like to slip in a reply to Sheldon.
Damn, I'm getting to like the way he thinks here. As I stated, I have difficulty with the idea that you put one "pure" DHT in the chain and the sound gets "realistic". It's just that bit implausible. I certainly buy the idea of enhanced realism. I agree that thoriated filaments sound better to my ears. I like the sound of the microphonic DHTs, and when you ping them you can hear it vibrate. Or to put it another way as someone else said "show me a microphonic DHT and I'll show you a good sounding tube". I've been through exactly the same "use the same materials" thinking as he suggests - I built a natural wood enclosure once which was - guess what - super realistic on cellos and basses. And my metal cone speakers were great for trumpets.....Well, you gotta laugh.... All these ideas are in harmony with the idea of enhancing sound "in the same direction" as ideal reproduction - some sort of enhanced super-real reproduction. It's all plausible - has some kind of face validity in measurement terms.
So does this seem to be what I'm hearing (or what we're hearing, since I'm far from alone here)? I'm listening through Apogee speakers, though. So all aluminium - no cabinet. Does this go a stage further - are ribbons the speaker equivalent of thoriated filaments? Tiny thin material that vibrates? Is this wholly a phenomenon of damping, then, in the final analysis? DHTs are less damped, ribbons are less damped? What do I hear with IDHTs - you could say they sound deader, i.e. more damped. That's plausible.
Hmmm..... got me thinking here. This is too simple. DHTS putting resonance back into the reproduction chain.....
andy
Damn, I'm getting to like the way he thinks here. As I stated, I have difficulty with the idea that you put one "pure" DHT in the chain and the sound gets "realistic". It's just that bit implausible. I certainly buy the idea of enhanced realism. I agree that thoriated filaments sound better to my ears. I like the sound of the microphonic DHTs, and when you ping them you can hear it vibrate. Or to put it another way as someone else said "show me a microphonic DHT and I'll show you a good sounding tube". I've been through exactly the same "use the same materials" thinking as he suggests - I built a natural wood enclosure once which was - guess what - super realistic on cellos and basses. And my metal cone speakers were great for trumpets.....Well, you gotta laugh.... All these ideas are in harmony with the idea of enhancing sound "in the same direction" as ideal reproduction - some sort of enhanced super-real reproduction. It's all plausible - has some kind of face validity in measurement terms.
So does this seem to be what I'm hearing (or what we're hearing, since I'm far from alone here)? I'm listening through Apogee speakers, though. So all aluminium - no cabinet. Does this go a stage further - are ribbons the speaker equivalent of thoriated filaments? Tiny thin material that vibrates? Is this wholly a phenomenon of damping, then, in the final analysis? DHTs are less damped, ribbons are less damped? What do I hear with IDHTs - you could say they sound deader, i.e. more damped. That's plausible.
Hmmm..... got me thinking here. This is too simple. DHTS putting resonance back into the reproduction chain.....
andy
Re: Re: Re: Re: Lynn Olson's model for audio testing
Completely agree.
I think the proper protocol for testing is that the only thing the subject needs to do is to give a score, being it a pain score, dizziness score or sound stage width score, high frequency extension score or whatsoever.
Identity in the test bears no meaning. We are here it determine whether A is better than B in providing a deeper sound stage, or better high frequency extension or better pain control.
The subject should not know the devices under test and deliberately not to identify the gears.
If we do the ABX test in 30 subjects or more, knowing the standard deviation and the mean score of the improvements/changes, we can then perform simple statistical test and then determine whether the change in the subjective sense under test is statistically significant.
For example. For our DHT vs IDHT preamp test, recruit 30 subjects.
Brief them on Lynn's scale of change.
Blind the subject under test and the controller. Don't even tell them we are testing DHT vs IDHT. Do the ABX test. Record the score of each subject.
Now we have 30 tests of ABX results. Feed the data to a computer and select the appropriate test from SPSS.
Bang, we have a test result. May well be there is no statistical difference between the change of perceived sound of DHT vs IDHT. Now that does not mean that we disprove the thing. Our measure may be wrong. But at least on Lynn's scale, there is not difference beyond random chance.
In biomedical research, we tried to isolate 1 variable to test. Adding a identification will complicate the test, making the statistics very complicated.
For SY's test, the same and different answer will only generate answer that whether there is a perceivable difference in A and B.
What I try to propose is a test using DHT and IDHT preamp, double blinded cross over study with a valid sample size of no less than 30 subjects and using Lynn's scale of audibility to measure the difference.
We may end up with a conclusion of "the study showed that subjects perceived a difference in Lynn's scale of 4.1 (95% confidence interval of 3.4-4.8) of the sound and DHT preamp vs IDHT preamp.
Well, may be I am too far gone. Time to sleep.
rdf said:
Completely agree.
I think the proper protocol for testing is that the only thing the subject needs to do is to give a score, being it a pain score, dizziness score or sound stage width score, high frequency extension score or whatsoever.
Identity in the test bears no meaning. We are here it determine whether A is better than B in providing a deeper sound stage, or better high frequency extension or better pain control.
The subject should not know the devices under test and deliberately not to identify the gears.
If we do the ABX test in 30 subjects or more, knowing the standard deviation and the mean score of the improvements/changes, we can then perform simple statistical test and then determine whether the change in the subjective sense under test is statistically significant.
For example. For our DHT vs IDHT preamp test, recruit 30 subjects.
Brief them on Lynn's scale of change.
Blind the subject under test and the controller. Don't even tell them we are testing DHT vs IDHT. Do the ABX test. Record the score of each subject.
Now we have 30 tests of ABX results. Feed the data to a computer and select the appropriate test from SPSS.
Bang, we have a test result. May well be there is no statistical difference between the change of perceived sound of DHT vs IDHT. Now that does not mean that we disprove the thing. Our measure may be wrong. But at least on Lynn's scale, there is not difference beyond random chance.
In biomedical research, we tried to isolate 1 variable to test. Adding a identification will complicate the test, making the statistics very complicated.
For SY's test, the same and different answer will only generate answer that whether there is a perceivable difference in A and B.
What I try to propose is a test using DHT and IDHT preamp, double blinded cross over study with a valid sample size of no less than 30 subjects and using Lynn's scale of audibility to measure the difference.
We may end up with a conclusion of "the study showed that subjects perceived a difference in Lynn's scale of 4.1 (95% confidence interval of 3.4-4.8) of the sound and DHT preamp vs IDHT preamp.
Well, may be I am too far gone. Time to sleep.
Sheldon said:
andyjevans said:
Lynn has many times pointed out in his ''Beyond The Ariel'' thread that speaker cones will always reveal the material they are made from, especially in the decay phase of notes, even if no particular ringing problems are present. Maybe that, in conjunction with your experiences, ties better his saying that subjective tone relates more with physics than with other domains.
Seems to me there is an easier, though perhaps more expensive, way to see what sort of damage your preamp does, which is to build 10 of them, dial in unity gain with a resistor divider on the input of each, and daisy chain them. Every first year photography student has done a similar experiment with a copystand or a photocopier and, at least in that context, it is quite revealing.
andyjevans said:
The general observation also goes with your description of the music being "richer". Sure, harmonic distortion of low even orders is said to do some of this, but I think the resonance added here is more subtle in level and, probably more importantly, more sustained. It's not going to be strictly a function of the fundamental.
I think that ribbons are actually very well damped. The air resistance compared to the mass of the ribbon is very large. So in this case, I'd guess that ribbons are not adding this characteristic. Rather, they are noted for being low in distortion, at least at low levels.
I don't think that just any resonance will do. It probably has to be resonance that has similar overtone structure so that it fits right in to our perceptual window. BTW, Earl Geddes has several times suggested that some reverb added to the bass can give a more realistic presentation of venue size.
Sheldon
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.