The width of the pulse is the analog, and the 1 and 0 involved in "on and off" in PWM cannot code a discrete value of the waveform absent the width component.
So it tracks based upon the analog width component, not the discrete on and off component.
So it tracks based upon the analog width component, not the discrete on and off component.
I understand that. It's the 1 and 0 component I'm interested in. It's digital because it's a pulse.
But I don't really care to convince anyone of my terminology, except to say jitter is an important element in any pulse-current waveform. Timing is critical---err, more than that, it's essential.
Now this is funny .... 😉
Class-D amplifier - Wikipedia, the free encyclopedia
Quote:
"The term "class D" is sometimes misunderstood as meaning a "digital" amplifier."
Class-D amplifier - Wikipedia, the free encyclopedia
Quote:
"The term "class D" is sometimes misunderstood as meaning a "digital" amplifier."
The terminology is irrelevant, Stig. Here's what I'm saying: the timing of the pulses is important, essential.
I'm not saying it does because I don't care about semantic games. What's important here is that the amp renders the input waveform a PWM version of that waveform where the timing of those pulses is essential to the amp's operational fidelity.
The best term I can find to describe variation in the timing of those pulses is jitter. Have you a better?
The best term I can find to describe variation in the timing of those pulses is jitter. Have you a better?
There is an abstraction, a CODING that makes digital what it is. The pulse width, an analog attribute, is what defines the signal output, a repeating 0 1 0 1 0 1 0 1 train ad inifinitum cannot contain an abstract variable coding, since the discrete on/off cycle is in a constant order.
A change in the timing of the PWM carrier will result in a non-reversible change in the analog audio output.
Jitter on a digital signal will not. The audio data remains the same, and can be successfully reclocked.
Jitter on a digital signal will not. The audio data remains the same, and can be successfully reclocked.
Yes, a non-reversible change in the analogue output. So get the timing right, yes?
Nothing is perfect. They got it right to the degree that their published specs show - and people listening to the amps can hear. Good enough for me. If you want to call a spade a forklift in your own terminology, feel free - but don't expect anyone else to agree or understand.
"Right to the degree"? That's vague fudge language that avoids real communication.
For instance, do those purported listeners know how a "righter" amp sounds to be able to say "right," which is a comparative term?
Here's my point. If comparator timing is imperfect, it can be improved. Same with forklifts. I hate badly timed forklifts.
Last edited:
I hate forklifts that violate the laws of quantum physics, for example those trying to occupy the same space at the same time.
Edit: The mods should be here soon and clean up.
I can hear them coming. But I don't want to lose my point in the semantic detour. Comparator (forklift) timing is imperfect, but how imperfect is it? Would a dedicated shunt, fed by dedicated series reg, fed by a transformer not clear up some ground bounce and other power supply artifacts affecting the timing of that delicate process? I think yes. Would that be audible? I think yes. Those are my hunches. I've yet to see evidence or reasoning that would suggest otherwise.
- Status
- Not open for further replies.