Charles Hansen said:
Hi Charles,
Thanks for answering. However, with all due respect, you did not answer my simple questions. You may think by your experience that these questions and answers are silly, but others may not. Let them decide, based on your answers.
For this test of the relays, can you just tell us what, if any, controls you had in place?
I would hope and assume that you are more diligent about listening comparison controls than the average equipment reviewer (where there are virtually no controls, as far as I can tell).
Mind you, I am not trying to open an ABX can of worms here. There is a very big and reasonable spectrum of listening comparison protocal between totally uncontrolled and something like ABX. I am not a fan of ABX, either. It confuses me. Where do your listening tests lie in that spectrum? That is all I am asking.
Thanks,
Bob
Controls, my aunt Fanny! Some here act as if serious audio designers, like Charles or 'little old me' are either money grubbing charlatans, or dimwitted fools.
Let me describe an early listening test that I was in, back in 1967. I made the predecessor to my first comp diff amp (more like a video amp in topology) and we ran a listening comparison between it and a highly regarded triode tube amp. We matched the outputs by differencing it across an oscilloscope to adjust out any amplitude differences and merely switched the two outputs. We separately measured the two amps at all levels up to clipping, measured the damping factor and frequency response. That was over 40 years ago.
In later listening tests, we have been even more careful with amplitude, polarity, and frequency response. What else could we want. We are NOT in this business to fool ourselves or others. IF my first Levinson JC-2 was enough, AND it would be for most of our critics, I would have stopped there. There are so many more things to do than to just design amps and preamps.
Let me describe an early listening test that I was in, back in 1967. I made the predecessor to my first comp diff amp (more like a video amp in topology) and we ran a listening comparison between it and a highly regarded triode tube amp. We matched the outputs by differencing it across an oscilloscope to adjust out any amplitude differences and merely switched the two outputs. We separately measured the two amps at all levels up to clipping, measured the damping factor and frequency response. That was over 40 years ago.
In later listening tests, we have been even more careful with amplitude, polarity, and frequency response. What else could we want. We are NOT in this business to fool ourselves or others. IF my first Levinson JC-2 was enough, AND it would be for most of our critics, I would have stopped there. There are so many more things to do than to just design amps and preamps.
Re: searching for a perv in a dark room
Peanuts.
=> SDS S-series, sealed and hydrogen filled, guaranteed 10mOhm at 100mA/10V for +10 million operations.
Four contacts in (anti) parallel (S4 type) equals 2.5mOhm
Try 1uV thermoelectric potential : sheet , bottom of page 2.
(i recall that the Self brother couldn't measure significant distortion figures either, Ch13 of his power amp bookie)
dimitri said:
Peanuts.
=> SDS S-series, sealed and hydrogen filled, guaranteed 10mOhm at 100mA/10V for +10 million operations.
Four contacts in (anti) parallel (S4 type) equals 2.5mOhm
Try 1uV thermoelectric potential : sheet , bottom of page 2.
(i recall that the Self brother couldn't measure significant distortion figures either, Ch13 of his power amp bookie)
john curl said:
Why don't you just let Charles answer the question, rather than trying to distract with zero-information cynicsm. Charles is a big boy who makes very fine equipment, and he does not need the likes of you running inetrference for him. I frankly would like to learn something from Charles.
Bob
john curl said:
When are the experts going to appreciate that audio is a rather peculiar branch of electronics, and one where the only ultimate arbiter is how the results actually sound?
Provided that any such observed phenomena is repeatable, consistent and reversible, and is also appreciated by others who have carried-out similar trials, one doesn't need any controls.
Whilst measurements do undoubtedly assist in assessing certain parameters, ultimately in this rather unique electronics discipline, all that matters is what the effects are (good or bad) which we *hear*, because this is the only reason for the existence of audio reproduction in the first place.
When carrying out such trials which I also have for around 40 yrs now, it is very rare that the results are not immediately quite apparent when listened to on our own familiar systems and with equally well-known recorded material, provided that the overall systems are of appropriately high resolution.
In the comparisons which Charlie has referred to, hearing additional details like *for the first time* a very feint licking of the performer's lips prior to singing the first words of a recorded song which has been heard dozens of times previously without hearing this effect before, is verification enough to indicate that there is less 'masking' of such fine details. The inevitable conclusion is that the latest switch, or whatever, is simply less (or in other instances 'more') harmful to these very low-level passing signals.
Furthermore, and vitally important here, when reverting to the previous component or whatever, this same sound is *no longer heard at all* when listening at precisely the same amplitude (easy in my case for some years with 1dB steps on one of my attenuators) and with all other *obvious* factors being the same.
When this clearly-observed change can be repeated back and forth dozens of times and also on different occasions with absolute consistency (and quite independently in other parts of the world without any collusion), there cannot sensibly be even any unknown factors (like temporary unwanted changes in mains condition, local RFI, condition of ones ears, or whatever else) giving rise to what is heard, other than the deliberately-modified part in question.
Most people who have experienced these effects are not complete idiots, and I certainly don't wish to waste time nor money on fruitless pursuits and I detest it if I ever (very unusually lately, but it happened in the early days) discover that I have even unwittingly been deceiving myself, which is still possible no matter how well-controlled any trials are subject to.
Some of my bespoke work was used by a high-end UK speaker manufacturer at UK's most prestigeous audio show at Heathrow a couple of years back, simply because he wished to show-off his wares to the best effect. There was nothing to prevent him from using anyone else's electronic equipment to partner his latest speakers, but having heard the sonic results of my work, he adopted mine.
Every visitor to that demo-room who commented remarked on the unusual clarity, purity, resolution, etc., etc., of the sounds and this was mainly due to the years of careful 'optimising' of my circuits, not only through measurements, but most significantly through the careful choices of components selected by painstaking and extensive listening trials.
Regards,
Bob Cordell said:
The control was the "bypass" connector. Remember, we made up a batch of connectors. Each was a pair of rhodium-plated RCA connctors, one male and one female, with the grounds directly soldered together to make one large back-to-back connector.
Then we soldered short lengths (approximately 2") of Cardas hook-up wire to the center terminal of each connector. For the "bypass" reference, we simply soldered the two wires together. For each DUT, we would solder the DUT (relay or switch to be tested) to the wires. In this way the reference "bypass" had the exact same number of connections, solder joints, and wires as the DUT's. The only difference was whether there was a DUT in the signal path or not.
The arrangement was to plug the reference "bypass" directly into the input connectors of the power amp. Then the cable from the preamp would plug into the reference "bypass". For the testing protocol we would listen to about a half dozen tracks of various types until we felt we had a good fix on the sound of the system. The volume control of the preamp is a stepped attenuator, so it is trivial to ensure that the volume is always the same if we changed it for different tracks.
Then we would plug the DUT in place of the reference "bypass" and listen to the same songs at the same volume. If there was any question about what we were hearing, we would then simply replace the DUT with the reference "bypass". It would only take about 15 seconds to change between the DUT and the reference "bypass".
The person changing the DUT's obviously knew which one he was inserting. The other people listening (typically one and two others) sometimes were told what the DUT was and sometimes not. We tried it both ways many times and it never made any difference.
Essentially the experiment would be like putting 10 different slices of cheese on a plate and comparing their taste to a "reference" cheese. You would taste one cheese and see if it tasted the same as the "reference" or not. At any time, you could take a bite of the "reference" to get a baseline.
Listening to the different DUT's was as easy as tasting different types of cheeses. They all had marked "flavors", EXCEPT for the Shallco switch. With that one we had to go back and forth several times. There was a *very* small difference compared to the reference "bypass", but it was so small that it was hard to describe and I highly doubt that I could identify it if somebody inserted it into the system without my knowledge. In contrast, all of the other switches and relays we tried made an easily identifiable change in the sound of the system.
Bobken said:
OMG, yes!!!
Like the time we spent about a week listening to different axial-leaded film capacitors. We had a PCB with 12 capacitors on it, and we would change all 12 at once to one brand or another.
After a week we found out (completely by accident) that the sound of the capacitor was markedly different if it was mounted tight down against the PCB instead of up in the air on longer (solid core) leads. So we had to re-do all of the previous listening tests.
(Since we were only trying the capacitors out, we originally thought we would keep the factory's long lead length during the listening test. Then if we ever wanted to use the capacitor for something else, we could. But this put the capacitor up in the air, off of the PCB. We had to cut all the leads so that the capacitors were held tight against the PCB. It made for a more expensive test, as the capacitors were essentially useless for anything else after they were tested properly.)
Charles Hansen said:
Been there and done it all in the early days, partly to my shame Charlie, and I am still learning something new most days.
I have boxes and boxes of supposedly wonderful-sounding parts collected over the years which I will never use again, but that is all part of development trials, unfortunately.
What used to catch me out for few years a long while ago was listening immediately after soldering something new. Whatever the reasons for this (and I have my own theories) it always adversely affects the sound for a while, sometimes 30 mins or so, but frequently for longer.
The same confusing problem doesn't exist when socketing or clipping parts together cold, so it is definitely related to the soldering operation, and it can make quite a difference for a while.
No prizes for the first poster to suggest that I am simply getting used to this imaginary 'difference', or whatever.
Regards,
Charles Hansen said:
Charles,
Thanks, I really do appreciate your detailed answer. It sounds like you really worked hard to not fool yourselves. This is the kind of test I would really like to have participated in.
I have conducted listening tests as well, and I do appreciate how hard it is to do them even close to "right" (indeed, it can be argued there is no "right", but maybe some approaches that are less wrong).
I take it that the DUTs were all constantly in their "ON" position, so there were no identifiable noises made by them.
There are a few things about such listening tests that I am always curious about, and maybe you have an opinion on them.
The first is: how "golden-eared" does the listener have to be to identify the kinds of differences you heard?
The second is: how familiar with the system does the listener have to be to identify the kinds of differences you heard? For example, if someone walked in off the street and only had the half-hour or so of familiarization you described, would that be enough?
The third is: How high must the quality of the overall system be to identify the kinds of differences you heard? I'm sure you listened on a very good system. Would those differences have been quite identifiable in a system whose total cost was 1/4 to 1/2 that of the system you used (still probably a quite good system).
The last is the following: let's say you did such a listening test 10 days in a row, but on each day you listened to only one of two DUTs. Perhaps the switch and one of the middle-range relays from your testing. During the night, without your knowledge, you assitant secretly swapped or did not swap the DUT. With what degeee of success do you think you could identify the DUT each day?
Thanks again,
Bob
- Status
- Not open for further replies.