Now, back to loudspeaker 'problems'. Loudspeakers have been made for more than a century in some form, starting with horns, then headphones, and finally direct radiators. However, the basic physics is relatively not understood by most, about how the devices work.
The direct radiator is an ACCELERATION controlled device, on purpose. Its parameters of mass and compliance have to make a LOW frequency resonance in order for it to operate over a reasonable range of operation. This effects its actual phase response and it is difficult to get a decent square wave or similar response over a wide range of operation because of this.
The horn loudspeaker works with a dominant VELOCITY response, due to the extreme loading on the driver by the horn, and can be relatively ideal over a limited range, BUT it is difficult to use the horn near or at its fundamental cutoff. You get lots of excess phase shift from the horn's problems and it is difficult to interface with a direct radiator as a result. Amateurs should note this, not all of you can or will buy a JBL or a Meyer speaker system, so you could be disappointed if you are not aware of the problem.
I know this from experience when I attempted to mate an Emilar 800 Hz horn with a 15" Gauss loudspeaker about 40 years ago. I had to try all kinds of Xovers, before I could get it 'right'.
Actually the 'best' speaker device that I know of for good transient response is the MANGER transducer. It looks like a direct radiator, but it acts more like a horn, being velocity controlled, rather than acceleration controlled. It's problems include harmonic distortion, cost, and little low frequency response. This MANGER principle is so good, that a small version, to allow extended high frequency operation was used to test the finest microphones in the world by Herr Manger and the results published in an AES article by Manger. This is pre-print 2336 given in 1986 in Montreux, Switzerland. Well, enough for now.
The direct radiator is an ACCELERATION controlled device, on purpose. Its parameters of mass and compliance have to make a LOW frequency resonance in order for it to operate over a reasonable range of operation. This effects its actual phase response and it is difficult to get a decent square wave or similar response over a wide range of operation because of this.
The horn loudspeaker works with a dominant VELOCITY response, due to the extreme loading on the driver by the horn, and can be relatively ideal over a limited range, BUT it is difficult to use the horn near or at its fundamental cutoff. You get lots of excess phase shift from the horn's problems and it is difficult to interface with a direct radiator as a result. Amateurs should note this, not all of you can or will buy a JBL or a Meyer speaker system, so you could be disappointed if you are not aware of the problem.
I know this from experience when I attempted to mate an Emilar 800 Hz horn with a 15" Gauss loudspeaker about 40 years ago. I had to try all kinds of Xovers, before I could get it 'right'.
Actually the 'best' speaker device that I know of for good transient response is the MANGER transducer. It looks like a direct radiator, but it acts more like a horn, being velocity controlled, rather than acceleration controlled. It's problems include harmonic distortion, cost, and little low frequency response. This MANGER principle is so good, that a small version, to allow extended high frequency operation was used to test the finest microphones in the world by Herr Manger and the results published in an AES article by Manger. This is pre-print 2336 given in 1986 in Montreux, Switzerland. Well, enough for now.
I attended a recording session that was being recorded by Peter McGrath when Dave Wilson was still living in the SF/bay Area in Calif. USA. Peter had brought along a portable digital recorder to back-up the analog tape recorder of Dave Wilson's. After a session Peter compared the two recordings on the monitor speakers -- WATT/Puppies. he was so excited that he asked myself and Sallie Reynolds (Editor of Absolute Sound) to listen. Peter said they sounded almost identical. And this was many years ago... now today digital is a lot better. We all agreed there was no noticeable difference in the two recording formats. Maybe if on a more familiar system to me, or better speakerrs, we could hear or detect something different but it was really very very close. So good that it started me looking in the digital direction. Now, all is digital. I gave away all my LP and TT and pre-pre/RIAA system long ago. I am still trying to get it better but now it isnt so much in the electronics as the speakers. Especially dynamic drivers need a lot of help/work to make more accurate music reproduction.
THx-RNMarsh
Last edited:
Well, before Dave died, he was still using my analog recorder to show off his latest speakers (the 2/3 million dollar ones) to TAS reviewers. IF digital was that perfect, I'm sure he would have used it for that demo.
I have heard direct comparisons, myself, fairly recently, between Peter McGrath's digital recordings and quality analog recordings, using $50,000 Wilson speakers through a Constellation demo, and while the digital was quite good, the analog was better. (still)
In my experience, some can live more easily with digital, and others cling to analog (like me) due to how they hear the sound coming from the loudspeakers.
I remember one guy, who worked at HP, about 35 years ago, finding even the first digital CD's were virtually perfect sounding. Even Dave Wilson teased him about some new improvements in CD's and this engineer said that then they would be even more perfect! We thought that was funny at the time. Later, Dave Wilson was having some trouble with a digital time delay network used in transferring his analog master tape to a vinyl recording. He decided it was important enough to create a test record, where he left the digital in one side of a vinyl record and without on the other side. He actually made a number of these vinyl test records. I had one myself, about 30 years ago. Unfortunately, it was lost in a firestorm. Anybody out there have one?
I have heard direct comparisons, myself, fairly recently, between Peter McGrath's digital recordings and quality analog recordings, using $50,000 Wilson speakers through a Constellation demo, and while the digital was quite good, the analog was better. (still)
In my experience, some can live more easily with digital, and others cling to analog (like me) due to how they hear the sound coming from the loudspeakers.
I remember one guy, who worked at HP, about 35 years ago, finding even the first digital CD's were virtually perfect sounding. Even Dave Wilson teased him about some new improvements in CD's and this engineer said that then they would be even more perfect! We thought that was funny at the time. Later, Dave Wilson was having some trouble with a digital time delay network used in transferring his analog master tape to a vinyl recording. He decided it was important enough to create a test record, where he left the digital in one side of a vinyl record and without on the other side. He actually made a number of these vinyl test records. I had one myself, about 30 years ago. Unfortunately, it was lost in a firestorm. Anybody out there have one?
Last edited:
My experience of the Manger transducer is a little different and the directional properties were strange to be sure but I downloaded the paper to see what was in it. The claim of a 5 uS rise time is exceptional and hard to verify. The scope photos were interesting. The spark discharge shows the doublet he talks about. Is that true with modern measurements as well?
Manger drivers are quite rare here so few will have any experience with them. I'm not sure how a voice coil in magnet connected to a diaphragm would work differently from others but would like to know.
It is possible to design equally good if not even better (CMA - CFA) with cheap small BJTs, and not only because those jfets are unobtainable. I designed preamps with both types and the listening does not show that this one with Toshiba jfets is better. The same is valid for power amps. I am old but try to live in present.
Damir
Well, as an old guy, I'll stick to complementary jfets for input stages at least. Most of my clients and even my competitors are using the same devices. That includes Nelson Pass, Ayre, and PS audio as direct competitors, and Parasound, Constellation, and Audio Illusions as clients. We will have a wonderful time competing with each other.
Dodad, I then hope you can convince Richard's client to build more of your design in order to compete with the rest of us.
Now, this does not exclude bipolar devices entirely, I am now reading "Analog IC design: the current-mode approach" for some new ideas, especially for 2'nd stages. I'm pretty sure that Ayre is far ahead in that department. Perhaps, dodad, you are as well.
Dodad, I then hope you can convince Richard's client to build more of your design in order to compete with the rest of us.
Now, this does not exclude bipolar devices entirely, I am now reading "Analog IC design: the current-mode approach" for some new ideas, especially for 2'nd stages. I'm pretty sure that Ayre is far ahead in that department. Perhaps, dodad, you are as well.
I'm sorry but this is entirely wrong.
Loudspeaker drivers, no matter which working principle and acoustic loading, are minimum-phase devices. Phase only depends on frequency response and vice versa, they are strictly coupled. No excess phase seen anywhere (well, there is a tiny amount in certain larger cone drivers where the acoustic center moves back when the outer sections of the cone start to decouple).
Last edited:
Today, it is. Take any decent speaker that has been optimized for every single aspect except for total phase response (because if you don't, striving for transient-perfect response with analog means *will* most often compromise things that are much more important for speaker performance, no way out).
To rectify time domain, you need to know the excess phase response, that is, the actual allpass response introduced by the acoustic crossover functions, preferably in analytical form (one can use curve fitting to get analytical data from measurements, though). Derive the impulse response of that allpass, invert it in time and use it as a convolution kernel the source signal is processed with. Done. Attacking global phase issues is done the same way than attacking FR problems in-situ: process the source signal. If your playback chain starts in digital, this is a perfect solution at absolutely zero cost and almost no side effects (the FIR phase-unwrapping has some latency and eats a bit of digital headroom).
See http://www.grimmaudio.com/site/assets/files/1088/speakers.pdf , section 2.4, for a practical example, integrated into the speaker. The XO functions are min-phase (IIR) thus the sum is not linear-phase, and the excess phase is corrected by the simplest possible FIR unwrapping kernel derived from analytical data (means no risk of artifacts because the kernel is exceptionally "clean").
Demian, as far as the Manger transducer is concerned, I have never had one to play with, even though I first met Herr Manger as early as 1975. Listening to one of his speaker systems was a disappointment though, for some reason. It did some things very well, but not well enough compared to many other speakers that I had heard.
However, Herr Manger made a microphone calibrator from a smaller version of his transducer and here are some of the results. He measured just about every popular condenser microphone in his paper, as well. His calibrator is a lot easier to follow than any spark gap, etc.
However, Herr Manger made a microphone calibrator from a smaller version of his transducer and here are some of the results. He measured just about every popular condenser microphone in his paper, as well. His calibrator is a lot easier to follow than any spark gap, etc.
Attachments
Yes - thanks!!! Something for Johns upcoming measurements!? And the affiliated How to Choose a Loudspeaker -- What the Science Shows - AVS Forum | Home Theater Discussions And Reviews
//
//
Last edited:
Good link!
But based on recent thoughts on step response dominant importance, I am not sure if there is a right audience here in this thread.
There is a misunderstanding between us, here, I think.
In my mind, the question was not what "sound" one can prefer, but to find between vinyl, analog tape and digital copies the closer one to an original.
Made a lot of listening comparisons with producers and musicians in various studios, at the beginning of digital. And the result was without appeal. Digital was the winner.
Vinyl ? Oh, Lord. It is like an old Polaroid in photography. One could even say, like a hyper realistic painting.
It's full of charm, but immediately recognizable.
Analog tape ? Closer, but this hiss and obvious distortion at peak levels ...
Let say, like a landscape and his photography. That some can enjoy because it gives both a feeling of "power" and a feeling of foundation to the sound construction.
Digital ? Could I afford a comparison with 3D glasses and this strange discomfort they provide ?
Then, I discovered that the analog part of the DAC was the most responsible of the "sound" color of various digital devices.
Let me add that, with time, and on an absolute personal and subjective point of view, my firsts impressions of discomfort with digital had disappeared with time and habit.
I conclude that our brains become accustomed to the character of the systems of reproduction, as they become familiar. Like we forget the color of some sun glasses after a short time.
Anyway, nowadays, digital has killed analog in all industry. Vinyl and tape are dead horses. So, the only reasonable position should be to work on digital gears in order to improve what *you* don't like. And I believe it is on the analog part of it that you can put your dough for the best.
How to convince you of such an evidence ?
Just compare the measurements between the best tape recorder you can find and any digital one, even the one in your smartphone.;-)
Last edited:
Yes, that and the noise nature of the clocking.
Clocking noise nature is trivially easily controlled and can cause sound from clean/accurate through to warm/involving etc through to 'running from the room' bad/nasty.
Agreed that we can become 'accustomed' to particular sound signatures, however the nature of the sound signature can long term lead to contempt if not 'right', and there are strong reasons for preferences and/or dislikes.
Does anybody know the typical source/provenance of the 'carbon black' used to colour vinyl records ?.
Dan.
Last edited:
Hi T...
Its amazing how similar our experiences are.... I asked the designer of the BenchMark DAC what was major problem for affecting the sound and he said it was the analog portion of the circuitry.
I also have a big Parasound amp to compare the newest/best CFA designs at the factory listening room in Bangkok.
Its pretty well known that many of the older reviewers also like a familiar sound from the past.... higher distortion. Some designers are now adding or leaving the distortion a bit audibly high deliberately as "an entertainment" value to the sound as so many "like" its familiar affect on the sound and sound field..
I long long ago left behind magazine reviewers of what sounds nicer or prettier or better than another. I stopped subscribing to them. After years of reading them and listening, you get to know their biases and can almost predict what one would say. So, the magazine reviews - as important as it is to mfr sales - is not relevant to me for the most part. I can listen as well as they can.
I understand JC side too... wider BW, faster Tr etc is better sounding than brick wall filters at 20KHz. A lot of that has since after CD era been addressed to the point where much more obvoious issues - like distortion within audio BW . Which is where legacy LP system fails miserably and obviously.
THx-RNMarsh
Last edited:
It not only depends on the media but the pre-amp, amplifier and speakers in the system. I have heard some terrible record players but also some great ones too.
As for sound I love the sound of a Marshall amp on full power but its not a faithful reproduction of what is attached, its high distorted.
- Status
- Not open for further replies.