Rigid inner walls were not a requirement, just the current state of the development. If it doesn't work, then obviously changes are allowed.
Or, keep the rigid walls but drill lots of holes in them ("between the cones", if you must, heheh) and put the cancelling panels just outside of them.
Last edited:
Amp Challenge
Oh boy. Hi Demian, I'm one of those engineers working with Richard Clark (President and head engineer of AMI, LLC), having been hired by him in 1986 to design his cassette replication facility and associated mastering suites. I became his Dir. of Engineering from the early 1990s until 2013, and I'm still working with him (and yes he is highly opinionated). He used to use me as the "good guy" foil to his "bad guy" act in the challenges, since I often disagreed with him (and still do) on sonic issues.
At this point, I'm sure some on this list are now throwing me in the "all amps sound the same" camp, but please wait: Like most prejudices, this one is often applied mindlessly too. The WHOLE POINT of the exercise that Richard and I performed over and over again, and what was originally intended only for solid state mobile amps was:
"When operated under clipping the differences between amps is below the threshold of human hearing."
After maybe thousands of people challenged in their own cars and facilities and couldn't hear a difference reliably, we began to be challenged by home audiophile amp manufacturers. I considered myself an ardent audiophile, and Richard, for all his experience could probably have been considered one as well, though he would never admit it.
When tube amps were put up for testing, we realized with their frequency dependent output impedance they could sound different. Therefore the only addition to the test regime was to analyze output impedance vs. frequency of the challenger, and make a simple RC network to make the reference amp's output impedance match. Hundreds, perhaps thousands of people failed to hear a difference.
So I took away some conclusions which have been thoroughly thrashed to death here:
1) ABX testing has some problematic issues, with the psychological state of mind playing a strong role in the ability to both concentrate and evaluate clearly. The longer I did the ABX testing (years), the more I was able to hear differences.
2) The magnitude of mutation impressed upon the audio by just about all amplifiers when used below clipping is extremely small and for most systems below most people's ability to detect. Note I intentionally used "most" in several places in that statement.
3) Just about all self-proclaimed "audio professionals" do not have well educated hearing and regularly miss sonic problems, and conversely imagine sonic differences where there are none. I have had both of these flaws proved to me in the past as well. Being brutally honest and rigorous with yourself is the only way to not back yourself into a problematic state of denial of your own making. I have seen well-regarded audiophiles make ignorant and easily disproved statements, yet they refuse to apply logic to the situation and recant. Like the well-known idiot who tried to sell us directional cables, and when performing a test I lied to him about the direction of the arrows, yet he stuck to his claims of audible change...when there had been no change in the hookup!
Demian's point relates to the fact that an audiophile merely swapping amps at home is not doing a controlled test. Therefore their opinion of one amp over another is just that, not any sort of fact. And hey, it's their money and ears, so why should anyone tell them not to go with one amp over another? I prefer an amp which does not have a frequency dependent output impedance, but that amp may be symbiotic with a certain speaker in someone's home. The advantage of my approach is that I can change speakers with impunity and not worry as much about interactions.
Flame suit on,
Howie
Howard Hoyt
CE - WXYC-FM 89.3
UNC Chapel Hill, NC
www.wxyc.org
Oh boy. Hi Demian, I'm one of those engineers working with Richard Clark (President and head engineer of AMI, LLC), having been hired by him in 1986 to design his cassette replication facility and associated mastering suites. I became his Dir. of Engineering from the early 1990s until 2013, and I'm still working with him (and yes he is highly opinionated). He used to use me as the "good guy" foil to his "bad guy" act in the challenges, since I often disagreed with him (and still do) on sonic issues.
At this point, I'm sure some on this list are now throwing me in the "all amps sound the same" camp, but please wait: Like most prejudices, this one is often applied mindlessly too. The WHOLE POINT of the exercise that Richard and I performed over and over again, and what was originally intended only for solid state mobile amps was:
"When operated under clipping the differences between amps is below the threshold of human hearing."
After maybe thousands of people challenged in their own cars and facilities and couldn't hear a difference reliably, we began to be challenged by home audiophile amp manufacturers. I considered myself an ardent audiophile, and Richard, for all his experience could probably have been considered one as well, though he would never admit it.
When tube amps were put up for testing, we realized with their frequency dependent output impedance they could sound different. Therefore the only addition to the test regime was to analyze output impedance vs. frequency of the challenger, and make a simple RC network to make the reference amp's output impedance match. Hundreds, perhaps thousands of people failed to hear a difference.
So I took away some conclusions which have been thoroughly thrashed to death here:
1) ABX testing has some problematic issues, with the psychological state of mind playing a strong role in the ability to both concentrate and evaluate clearly. The longer I did the ABX testing (years), the more I was able to hear differences.
2) The magnitude of mutation impressed upon the audio by just about all amplifiers when used below clipping is extremely small and for most systems below most people's ability to detect. Note I intentionally used "most" in several places in that statement.
3) Just about all self-proclaimed "audio professionals" do not have well educated hearing and regularly miss sonic problems, and conversely imagine sonic differences where there are none. I have had both of these flaws proved to me in the past as well. Being brutally honest and rigorous with yourself is the only way to not back yourself into a problematic state of denial of your own making. I have seen well-regarded audiophiles make ignorant and easily disproved statements, yet they refuse to apply logic to the situation and recant. Like the well-known idiot who tried to sell us directional cables, and when performing a test I lied to him about the direction of the arrows, yet he stuck to his claims of audible change...when there had been no change in the hookup!
Demian's point relates to the fact that an audiophile merely swapping amps at home is not doing a controlled test. Therefore their opinion of one amp over another is just that, not any sort of fact. And hey, it's their money and ears, so why should anyone tell them not to go with one amp over another? I prefer an amp which does not have a frequency dependent output impedance, but that amp may be symbiotic with a certain speaker in someone's home. The advantage of my approach is that I can change speakers with impunity and not worry as much about interactions.
Flame suit on,
Howie
Howard Hoyt
CE - WXYC-FM 89.3
UNC Chapel Hill, NC
www.wxyc.org
One way of doing it is to take a bass speaker that goes really low, and measure it anechoically. You only have to do this once. Subsequently you measure the same speaker in your small anechoic room, and presto, here you have your compensation file. It is of course not a 100% solution. It will only take care of compensation for compression (room gain).
If that's allowed (Ed denied this), then there's quite a few solutions, Yours would probably work. Turning the walls into broadband LF resonators (a few ways to do that, mass-compliance elements being obvious) would also work. You can also have the box be acoustically transparent at LF and measure in a large space.
The key is to not impose the artificial restriction of all sound having to be absorbed inside the box and not audible outside the box.
Bingo.
The three approaches I see:
Infinite baffle stuffed with fiberglass
Active cancellation (OK one wall is not rigid)
Over sized plane wave tube (Bottom is open and stuffed with fiberglass)
Now once you look at active cancellation the problem becomes almost trivial. You just use two of the same model driver coupled face to face and driven out of phase. One is your reference and the other is the device under test.
Sometimes diyaudio doesn't send notification emails for hours . . . or days.
So a 2 hour delay has no significance.
Believe me, For me this thread has never been so interesting - you have my full attention.
I see some problems with this approach. In order not to disturb the measurements, the walls would need to have the same compliance as air. Plus, if the walls would be transparent to LF, this would go both ways. In other words, the large space you measure in should be anechoic or sound would reflect back into the box.
Since the compliance of absorbing material is generally quite different than air, I don't see how you come to this conclusion. If you wanted the compliance to be the same, that can be arranged, but the key is to have some damped mass.
I don't see how this would work either.
I can't picture this infinite baffle stuffed with fiberglass. You would want an anechoic chamber to see what a speaker does in 3D-space at various angles. How would this work?
Active cancellation with a same model driver coupled face to face and driven out of phase would create almost perfect cancellation, so what are you going to measure and how? Plus, the driver being measured would change its behaviour if it does not receive the acoustic load it would normally receive from the air it radiates in.
An oversize plane wave tube would produce a different acoustic load as compared to free air radiation, thus influencing measurements.
Hence, I still think the only way to approach a large anechoic room with a smaller one is by creating a compensation file, using a known woofer, much like you do with microphones. And still, it would not be perfect.
If you want to do anechoic measurements without a room, rent a crane and hang the DUT with a microphone brace high up in the air in a silent spot. At 90 feet, you can measure down to 10 Hz or so.
Absorbing material doesn't comply with sound pressure. If it did, you would call it a flexible wall. A flexible wall can never have the same compliance as air, it will always be lower, if only because on the other end of the wall there is also air.
A fiberglass spike has pretty much the same compliance as air, although the fiberglass takes up some volume, so in a closed enclosure, compliance would go down a bit.
How would this damped mass react to an impuls measurement? It would look as a wall. It may have the same compliance as air under a static force, but it will behave differently under a dynamic one. Soundwaves in the air cannot tranfer their kynetic energy efficiently to something that is much denser, so for the largest part they will be reflected.
Last edited:
Scott's ckt from post
http://www.diyaudio.com/forums/anal...ch-preamplifier-part-ii-4431.html#post3669202
http://www.diyaudio.com/forums/anal...ch-preamplifier-part-ii-4431.html#post3669202
I have no clue of semi. physics, so why is that? No way of Dielectric isolation?
Last edited:
Not impossible but not economical, i.e. highend audio is not a large market. There might be a small market for cryo amps (FET's should work at 77K). We have tried it but there is no compelling reason to run the extra layers.
Last edited:
No indeed, you are talking about compression, which is completely predictable, so you can callibrate Ed's "puzzler" using a speaker with known properties as I suggested before.
Any other solution will not work because it either reflects back sound or would offer a different acoustic load to the speaker being tested as compared to it's free air behaviour, so measurements would become useless.
I hope someone proves me wrong because I have been dreaming about having my own anechoic room for ages. If it could have been done, it would have been done.
Lab curiosity for now to evaluate the possibility. As you might expect they are very non-complimentary. I would be interested to see if LTSPICE results match mine, that compensation is possibly still a little squirrely at low closed-loop gains. You could also glue Dick's buffer on instead of the two FET's and see if it still works.
Last edited:
- Status
- Not open for further replies.