Lobbying for a Class D Forum

[Bruno Putzeys]

Trouble is, the speed of reaction of the moderators is as if they have to pay for it... In that case one switching power forum may be easier to get than two (one for amps, another for supplies)

On the other hand, if the luxury is available, two forums would certainly be more practical.

[JaroMi]

I also vote for a "class D" forum.

[thoriated]

IMHO, to enter the audiophile ranks, a class D amplifier will need to switch at 500khz or more. Why? To preserve the interchannel phase subleties that are required for good imaging, where measurements have shown listener sensitivity down to the 2 microsecond range.

I believe analogies can be drawn here with high sample rate PCM and DSD where substantial improvements in imaging and inner detail are noticed up to a sample rate of at least 384Ks/s. What is even more interesting is that the transient characteristics of the anti aliasing filters (particularly the relative amount of pre- and post- ringing (less is better)) may also affect the perception of the reproduced image 'size' and 'depth' and even bass frequency 'weight'.

[classd4sure]

Originally posted by thoriated
IMHO, to enter the audiophile ranks, a class D amplifier will need to switch at 500khz or more. Why? To preserve the interchannel phase subleties that are required for good imaging, where measurements have shown listener sensitivity down to the 2 microsecond range.

I believe analogies can be drawn here with high sample rate PCM and DSD where substantial improvements in imaging and inner detail are noticed up to a sample rate of at least 384Ks/s. What is even more interesting is that the transient characteristics of the anti aliasing filters (particularly the relative amount of pre- and post- ringing (less is better)) may also affect the perception of the reproduced image 'size' and 'depth' and even bass frequency 'weight'.

So, I guess you're for another class d forum? What you say is interesting enough, though, I can't comment on it. Try posting it in the "distortions for classd" thread please.

[thoriated]

So, I guess you're for another class d forum?

Sure. This technology might even have 'silk purse' potential if properly developed

[Bruno Putzeys]

Originally posted by thoriated
IMHO, to enter the audiophile ranks, a class D amplifier will need to switch at 500khz or more. Why? To preserve the interchannel phase subleties that are required for good imaging, where measurements have shown listener sensitivity down to the 2 microsecond range.

I suggest you go and study the sampling theorem (for the first time [] or again[], check what applies).

http://graphics.cs.ucdavis.edu/~okre...gTheory101.pdf

The timing resolution of 44.1kHz/16bit PCM is around 90ps. NINETY PICOSECONDS. At higher word lengths, it becomes even better (24 bits: 350 femtoseconds).

Look at it like this. Suppose you have a transmission channel with only 20kHz of bandwidth. Does it only have 25us timing accuracy? Of course not! Phase information counts too, you know...

I hope that after reading the above document you will understand that a properly executed sampled transmission channel is fully, entirely, completely and totally equal to a brickwall lowpass filter, and this goes for the time domain as much as for the frequency domain (owing to the complete duality among those two).
Note "properly executed". 90% of the sonic difference between high-rez audio formats and 44.1kHz comes down to technical shortcuts taken in the execution. The remaining 10% is due to the "20kHz hard limit" of the human ear being something of an oversimplification.

Ergo: you don't need 250kHz bandwidth or 500kHz sampling to reproduce signals (be it steady state or transient) down to a resolution of a meagre 2us.

I wished people who didn't understand sampling (or quantisation) theory would be capable of detecting this themselves, instead of having to irritate other people first.

Come to think of it, I wished the same would go for all those peope who make pseudoscientific claims about anything.

[el burto]

mmm I made a self oscilating class d amplifier, based on Bruno Putzeys' patent US6046636, switching on 200 kHz and it performed gooooooood
And no, it wasn't only for woofer applications, it reproduced nice sound from 20 Hz to 20 kHz
And it was a SODA, so what would be the result if it was such a sweet UCD
But yea, this is going off topic so i better be quiet.

Greetz

Bart

[Bruno Putzeys]

Eek! I'm getting "fans" close to home now... Hope I don't get stalkers...

[el burto]

Nah, i'm only a class d beginner.
Like probably everyone in his last year of the "secondary education", i had to make a thesis, or whatever you wanna call it.
That was a year ago, i didn't know what i wanted to make so after all i said "yea, then i will make an amp because i had some 20 years old speakers".
Like usual in our school, that would be a AB.
But then there was a teacher who totally dislikes audio but used class d amps for controlling motors and he said, Bart try to make a good amp, a class d and not one that runs hot all the time.
But thanks to the school system i din't know what class d amps were because they stopped in school after the class c
So i started searching information and people told me, kid try to search information about the circuits of Bruno Putzeys because they are "simple" and good.
So i found some nice info and got help of other people who made SODA's.
I also had the possibility to see a schematic of a student who had worked together with you (thats what people told me).
So i made one simple SODA and couldn't believe the result

BTW: i'm not a "dangerous fan" but someone who is interested in good class d amps like probably much more here

Greetz

Bart

[thoriated]

It would appear that Bruno Putzeys will have fundamental philosophical differences with the authors of the following:

Effects in High Sample Rate Audio Material


The authors write:

Recording engineers, and many musicians - particularly in the classical area - are becoming aware that material recorded and edited using these higher sample rates have some attractive qualities. Current theory on how human hearing works has so far been unable to explain the basis for these qualities, but they are none the less easy to demonstrate.

He should note in particular the correlation between the anti-aliasing filter responses used and pre- and post- impulse characteristics as well at at which frequencies their effects apply and their subjective effects, in conjunction with the effective sampling rate, which strongly affect inner detail, soundstage size and depth and even, apparently, subjective bass weight. Then, he may assert that none of these factors pertain to Class D again if he wishes.

The fact is that redbook 16 bit linear PCM is deficient in, among other things, being able to present anything but the most threadbare stereo image and that is largely because of its poor interchannel time domain resolution (90ps = 1 bit resolution for redbook which is about 40 db (7 bits) shy of minimally reasonable audio quality for the purpose of lateral discrimination). And, of course, this is the optimal case, which is substantially better than what the end result will be in practice with an actual audio progam, which is not recorded at an average level equal to the maximum available dynamic range of the format on my side of the planet.

Many Class D amps have poor enough output waveform timing stability that the substantial cycle to cycle jitter at the midrange and above is easily observable on an oscilloscope and also apparently creates phase lock problems with some audio test equipment. I'd call that a 'problem' that is likely to present symptoms similar to that described above for redbook.