Apparently you missed the 40 years mark.
A small observation :
- the term 'science' is most frequently used by non-scientists
- the more often the use, the lesser the chance it's a scientist
- if it is a non-scientist who uses the word, the higher the odds it's an excuse to hide behind for lack of a veritable argument.
DBT has been just about the only subject you've posted about in the last 9 years.
Would you like me to tell you how many times you've mentioned science, or a variation of the word, in the course of close to a decade ?
It is a poor response when a student asks "why are you teaching something which is different from the textbook?". In this case, the student ought to tell the prof to go and read the textbook although this will not make him popular with the prof.bear said:
It is a good response when the 'student' wants the prof to read the textbook for him, tell him exactly which parts he needs to learn parrot fashion, and then drop big hints about what questions will be in the exam.
My understanding is that the broad parameters for high fidelity sound reproduction were established in the 1950s, and have not changed much since then. If I wanted to know the actual numbers I would use Google. The reason they have not changed much since then is that good experiments seem to confirm that the original tests gave the correct results. Perhaps the biggest change has been the realisation that huge 'distortions' of the music signal can be slipped past most listeners if done correctly, so we now have perceptual coding. In addition, it has always been clear that many listeners don't like high fidelity sound reproduction, but prefer something different.
Surely this was done (at least by the Germans) in the 1970s? DIN standards for hi-fi? These days, of course, people prefer personal freedom and market forces so would not wish a standards committee to tell them what to buy. Standards now stick to boring things like safety, interoperability, EMC.
It is a modified version of "if you have never played with X then you have no right to apply circuit theory to X". Or "my system cost a lot more than yours, so therefore I have better hearing than you". Or "I have sold a lot of items or written lots of articles, therefore I know more than you". Or "I have been playing with electronic experiments for many years, therefore I understand circuit theory better than someone who has read a textbook - by the way, that Fourier fellow talks a lot of nonsense, doesn't he?".billshurv said:
Lowest THD does not necessarily correspond to lowest audible distortion - but I assume you know that? Increasing THD as the bias is changed is likely to vary the distortion spectrum, which may be audible or may not be audible depending on how much distortion there is and what the spectrum is. Why are you asking these questions: seeking knowledge or seeking an argument?bear said:
Same THD; sound different? Of course, unless that THD is sufficiently low. Same distortion spectrum, same frequency response, same clipping behaviour, same output impedance, same noise/hum level; sound different? Probably not.
So you can only hear differences when, by some means (such as level differences), you can be sure which device you are listening to? In other words, you only hear differences when you expect to hear differences. When you can't tell whether you should hear a difference you don't hear a difference. To me that sounds rather like "they sound the same, I can't tell them apart, except by some means other than what they sound like (such as how loud they are, or what they look like, or how much they cost)".john curl said:
"Better" can be confused with 'prefer', but even then it is useful to know whether someone can actually tell the difference between what he professes to prefer and what he does not. If I say I prefer chocolate icecream to vanilla icecream it might be useful to check whether I can tell the difference when I can't see the icecream; otherwise the truth might be that I prefer brown icecream to yellow icecream.Jakob2 said:
And that John is an honest account of what we are trying to tell. The preamps sound so much alike that your ears only cannot determine any differences so your brain is at a loss, and consequently you are confused.
Your brain can tell you which is playing at any one time when your eyes can see it, or a buddy tells it, or whatever additional non-auditory clues are present.
Jan
Its ok to not be able to tell similar spec'd amps or pre-amps apart. It is not a crime.
People buy equipment because it
1. They see an advert, a picture or a friend tells them about this awesome gear they saw. We could call this 'nascent product lust' or NPL for short.
2. They go and find the product. It sounds good - I'm sure a JC designed 400W Parasound amp is awesome paired with some B&W 802's
3. They like the way it looks. Polished aluminium, chrome, high quality printing, weight and feel - no different to an auto! It looks like an 'engineered' product.
4. As they hand over their money, the 'pride of ownership' thing kicks in: "**** it man, I actually have one of these things" is what they are thinking. Then there's the excitement that follows.
5. For the next two weeks, they get up early every morning, go down into the living room to look at it, stroke the dials, feel the heft, enjoy the quality of the thing, congratulate themselves on their purchase.
So, when you design a piece of audio gear, or a car, a chopper motorbike or anything that is going to sell for a few thousand $, this 1-5 above is what you are aiming to do for the customer.
ABX and DBT tests are a small part of it.
People buy equipment because it
1. They see an advert, a picture or a friend tells them about this awesome gear they saw. We could call this 'nascent product lust' or NPL for short.
2. They go and find the product. It sounds good - I'm sure a JC designed 400W Parasound amp is awesome paired with some B&W 802's
3. They like the way it looks. Polished aluminium, chrome, high quality printing, weight and feel - no different to an auto! It looks like an 'engineered' product.
4. As they hand over their money, the 'pride of ownership' thing kicks in: "**** it man, I actually have one of these things" is what they are thinking. Then there's the excitement that follows.
5. For the next two weeks, they get up early every morning, go down into the living room to look at it, stroke the dials, feel the heft, enjoy the quality of the thing, congratulate themselves on their purchase.
So, when you design a piece of audio gear, or a car, a chopper motorbike or anything that is going to sell for a few thousand $, this 1-5 above is what you are aiming to do for the customer.
ABX and DBT tests are a small part of it.
Between the real electronics experts here isn't there a way to create an AB or ABX control system that is noiseless, so no clues to which selection is chosen? Do you have to use a mechanical relay to do this, or is there a noise free, at the level of audibility, mechanical relay that exists?
I can see how an AB test would be easier to identify a difference than an ABX test but do understand how a random factor(X) reduces the probability of a simple 50% guess in determinations.
What would the minimum number of test subjects be to have a statistically significant and reliable bell curve that would show a trend or number of outliers, (true Golden Ears) in the general population? Statistics was not my best subject!
I can see how an AB test would be easier to identify a difference than an ABX test but do understand how a random factor(X) reduces the probability of a simple 50% guess in determinations.
What would the minimum number of test subjects be to have a statistically significant and reliable bell curve that would show a trend or number of outliers, (true Golden Ears) in the general population? Statistics was not my best subject!
Not the only possibility, different residual DC offsets at input/output can cause a pop when flipped around..
Indeed, if two items sound the same to you, that's just that. Not terribly important in the grand scheme of things.
And using ABX/DBT for a purchase decision isn't terribly useful either.
Blind testing is an important tool during development when you want to find out what the effect is of circuit changes, part substitution, matching, that sort of thing.
Edit: Andrew, see my sig line ��
Jan
Last edited:
Fade up and fade down a masking noise before and after each switching operation. This was to be part of the switcher I was working on for Harman, and I believe they do use something comparable now with the speaker mover. Note that this is a completely independent sound source, not something supplied by the devices under test.
I was thinking the DUT's, this would not always be a problem depending what you are switching in and out just something to watch for.
Former: probably a good thing, unless you're saying it negatively affects the short-term memory of the prior test.
Latter: asymmetric DC offset in the DUTs, most likely, but may creep in from the switching hardware itself. Probably benign unless the DUT is seriously messed up, but I haven't spent much time listening to relay clicks under different DC biases.
Beats me. Remember that the devices under test [edit: that is, for the Harman system] are loudspeakers, and the masking noise is to mask the speaker mover's noises. Also each speaker has an independent power amplifier---there is no switching except for line-level inputs. In my original design I used photoconductors to smoothly transition between line level signals with preset gains, which would be completely bypassed at the end of the switching operation by mercury-wetted reed relays. But this system was never completed---it was going to require more money and that had to come from somewhere, possibly a secretary being let go, which was unpalatable. There was also the concern that at the end of it all the monster wouldn't work. In their defense I really didn't jump on the project as quickly as I should have, instead starting to design new products for the consumer group. The two principal people interested in the switcher were not equipped to evaluate my designs on paper, which didn't help.
In the case of electronic components the slow transition between selections via the photoconductors would probably allow for the absence of the masking noises. As far as I know there is little motivation for testing source components, preamps, and amplifiers per se, although Olive occasionally remarks that he will have to do something someday, despite the strongly-held belief that results will be null ones.
Last edited:
- Status
- Not open for further replies.