For perspective, lots of interesting thoughts, ideas and opinions here. Based on what has been presented, confounded with my own opinions, is that there is not any single good objective indicator to predict perceived sound. More than that, there appear to be many potential objective indicators that are or were implicated in sound quality. As was demonstrated in a previous post a scheme can be devised to show how many small (values) contributors can be summed to be a poor result. Problem is, most of these indicators are not orthogonal and we have not the slightest idea of how they combine (sum), realistically many will interact and there will be many indicators (I suspect the # of indicators will be roughly proportional to the number of contributors). Bottom line is that if AFOM is to happen, quite a few subjective choices/decisions will be needed and that means tradeoffs and compromises. FWIW
Bill
Bill
(1) They have stated they don't care about "perceived" sound because that's subjective.
(2) They also claim they can define a set of repeatable tests based strictly on some measured criteria ( aka: objective ) that is wholly sufficient to measure the "quality" of an amplifier.
(3) They fail to see that (2) and (1) are in conflict since the selection of measurement criteria and decision not to include listening criteria is strictly subjective and an act of faith.
(4) They get upset when you tell them this and propose alternative or expanded measurement criteria. Peer Review Comments are not welcome, even though they asked for comments.
Not a scientific endeavor.
Meaningless.
We are wasting our time discussing this... their minds are set.
The Julian Hirsch, David Ranada, ASR school of thinking.
Well the main problem with perceived sound is that it involves a loudspeaker+room or headphones, not just the amp. The job of the amp is to amplifier an electrical signal to produce another electrical signal, plain and simple.
As well as your opinion. But I have not seen one single engineering post of yours - circuits or measurements. You guys talk to much and work very little in audio. To debate is not same as to do something. Anyone can debate.
I think this thread somehow summarises the whole trade. Very interesting. I would say that there can be made valuable additions to the general state of the nation. But I still remain in the view that it needs to be strictly tech. And even then, there will be aspects of limiting measurements in range and level + trying to define a quantifiable scale, that will involve "subjective" aspects - but as engineers we need to see this clearly and navigate professionally to create a balanced and usable outcome - and one that we can describe in terms of why and how it shall/can be used.
//
//
Okay, but its not just about measurements and circuits. Its not directly about circuits at all. And its not about certain measurements that have been posted, those particular measurements were off-topic, just like a lot of the discussion. Its about black box measurements with a soundcard and a scope using existing tests. Its about picking a number of such tests and assigning scores to them with the aim of discouraging the common over-focus (WYSIATI) on Distortion/THD/THD+N/IMD/SINAD, and encouraging a more balanced focus. Period.
Last edited:
You have to take precautions for this. CFA has the excess VAS current problem, but the rail sticking precautions luckily cure that, so it’s two birds with one stone. There are a few techniques here covering VAS sticky rail that do thd job of limiting excess VAS currents but it’s not exhaustive
https://hifisonix.com/wp-content/uploads/2019/02/Anti-Saturation-Diodes.pdf
To clarify my OPINION. I try to be objective and believe in the pursuit of objective metrics in design and validation. Unfortunately, I find that many engineering or applied science/mathematics problems cannot be solved without assumptions/opinions/omissions and those assumptions are not always right. As an extreme, galloping girdy gets designed ( a wonderful example of instability). Also, I find that when multiple, knowledgeable people get into a polarized debate, they are often all correct ... within their assumptions. Effectively they are solving different problems).
I think this thread has not dismissed perceived sound. The thread has pointed out the difficulty in knowing what it is and the difficulty in using subjective data. I like the Bob Carver/Reviewer experiment and the follow up experiment that showed the drift of the reference.
I, also like data that has physical units or a well defined and maintained standard.
Dimensionless data is difficult to use in making trade-offs. Granted that making trade-offs is not the goal of AFOM.
Bill
I think this thread has not dismissed perceived sound. The thread has pointed out the difficulty in knowing what it is and the difficulty in using subjective data. I like the Bob Carver/Reviewer experiment and the follow up experiment that showed the drift of the reference.
I, also like data that has physical units or a well defined and maintained standard.
Dimensionless data is difficult to use in making trade-offs. Granted that making trade-offs is not the goal of AFOM.
Bill
It really isn’t worth trying to discuss this is you misrepresent the initiative in this way. Ignored.
Bonsai - The clamp diode does not affect the charge stored in the Miller compensation capacitor. The CFA's two halves have asymmetrical drive currents into the VASes/integrators - one side's current can go very high but the other side's current cannot go below zero. This asymmetry shows up during slewing.
PMA mitigated that problem by using shunt capacitors instead of Miller capacitors. His amplifier's pole is at 20KHz. That should result in lower excess VAS current.
Ed
PMA mitigated that problem by using shunt capacitors instead of Miller capacitors. His amplifier's pole is at 20KHz. That should result in lower excess VAS current.
Ed
Ed, CFA’s don’t slew. The clamp diodes prevent the VAS current from going much above the standing VAS current. Without them, a 10 mA VAS standing current easily goes in excess of 100mA. They also very effectively prevent rail sticking (Baker clamp). Current limiting the VAS as Ostripper and dadod do also achieves the same thing.. See the kx2 amp for more info on this.
Maybe start a new thread on this?
Maybe start a new thread on this?
Bonsai - The input currents to the two VASes/integrators remain asymmetric. Instead of slewing, one integrator will supply base current to the other through the Miller capacitor. Both VAS'es collector currents will increase. This may not be a problem as long as the VAS can handle the current.
Ed
Ed
