Class A vs AB vs D amplifiers

[Blenderguy]

You won't hear the defects even if you listen for them if you don't know the type of artifacts to listen for. I did a lossy mp3 vs lossless ABX with a colleague en he got all of them right. He told me he just listens for artifacts and not the actual music.

I do the same. It took several years of training by splitting the audio spectrum in several bands in your head. When designing audio amps you know what circuit changes effect what and sometimes fixing one problem creates another. So I never really listen to the music and pick musical pieces that could highlight certain problems.
There have been several cases where I thought for sure it would sound better and it was not so. I like to be in this kind of control instead of just taking what someone else thinks is the right circuit.

 

[Jim Taylor]

I do the same. It took several years of training by splitting the audio spectrum in several bands in your head. When designing audio amps you know what circuit changes effect what and sometimes fixing one problem creates another. So I never really listen to the music and pick musical pieces that could highlight certain problems.
There have been several cases where I thought for sure it would sound better and it was not so. I like to be in this kind of control instead of just taking what someone else thinks is the right circuit.

When playing regular, off-the-shelf recordings, any amplifier design that lets you hear changes, or which sounds better or worse compared to another circuit, is either being used beyond its envelope, or else it's an incompetent design.

Although there is a greater possibility of hearing noise with laboratory signals (I support a level of -80 dB), the accepted threshold of hearing differences with music is higher ... about -60 dB. The reason for this is the ever-changing spectrum content and signal characteristics, which make it difficult for the brain to isolate signal characteristics.

Any modern design with flat response that can't get SINAD below -60 dB is pretty poor, and any modern design that is unstable into so-called "normal" speaker loads is also pretty poor. It's better to design with tests and measurements, made with instruments specifically designed for that task, rather than by ear. Otherwise, the results are apt to be caused by cognitive bias.

Jim

 

[Blenderguy]

I agree that signal generator and spectrum analyser / oscilloscope can zero it in. But the problem I see is that our audio measurement methods are still in the dark ages. Seriously measuring amplifiers using sinewaves is dinosaur methods. I would have expected newer tests to have come out that measure assymetrical pulses like what music really is. All new digital generators can produce arbitrary waveforms. Feed those into a device under test and compare output with reference. Add delay to direct signal to compensate for extra cable length and transit through the amp ( in this example). And maybe you getva clearer view of what the differences are.

 

[Doodski]

I agree that signal generator and spectrum analyser / oscilloscope can zero it in. But the problem I see is that our audio measurement methods are still in the dark ages. Seriously measuring amplifiers using sinewaves is dinosaur methods. I would have expected newer tests to have come out that measure assymetrical pulses like what music really is. All new digital generators can produce arbitrary waveforms. Feed those into a device under test and compare output with reference. Add delay to direct signal to compensate for extra cable length and transit through the amp ( in this example). And maybe you getva clearer view of what the differences are.

What kind of waveforms repeated or pulsed one time are you considering?

 

[NTK]

Quoting Doug Self again.

The Dan Rather Approved Douglas Self Site

"Sinewaves are steady-state signals that represent too easy a test for amplifiers, compared with the complexities of music."​

This is presumably meant to imply that sinewaves are in some way particularly easy for an amplifier to deal with, the implication being that anyone using a THD analyser must be hopelessly naive. Since sines and cosines have an unending series of non-zero differentials, "steady" hardly comes into it. I know of no evidence that sinewaves of randomly varying amplitude (for example) would provide a more searching test of amplifier competence.​

I believe this outlook is the result of anthropomorphic thinking about amplifiers; treating them as though they think about what they amplify. Twenty sinewaves of different frequencies may be conceptually complex to us, and the output of a symphony orchestra much more so, but to an amplifier both composite signals resolve to a single instantaneous voltage that must be increased in amplitude and presented at low impedance. The rate of change of this voltage has a maximum set by the frequency response and amplitude capability of the channel and is not generally greater for more complex signals; you do not get hgher slew rate with bigger orchestras. You must remember that an amplifier has no perspective on the signal arriving at its input, but literally takes it as it comes.​

 

[SIY]

The minimum now is 320, except for free streaming. That’s at the edge of being detectable in double blind tests. I doubt if any male over 40 can hear it.

I did a double blind comparison (sorting test format) when I was just over 50 and nailed it. And I am decidedly male.

Could I do it now? Likely not.

 

[Blenderguy]

The reason why I am saying this is after spending decades in the CATV industry, I saw the development of new measurement techniques as the number of channel loading increased. And you could not claim that there was no problem with the amplifiers that measure really good CTB and people were seeing thing (no pun intended). So then came crossmodulation and the last measurement to be developed was Composite second order.
So I take it with a grain of salt when people throw around numbers of "undetectable" distortion you can't possibly hear. When maybe your measurements are archaic. Has anyone else ever thought about developing different measurement techniques?

 

[SIY]

I like to be in this kind of control

Speaking of controls...

 

[SIY]

Has anyone else ever thought about developing different measurement techniques?

It might help for you to familiarize yourself with actual measurement methods.

And of course, after decades of people saying this stuff and decades of being asked to give ONE example of an audible difference that isn't trivially measurable, the response has been, "Ummmmmm.... I don't have to!"

 

[Blenderguy]

@NTK
There it is again. "Continous" random sinewaves. What if the signals are not continous? If anyone has ever taken higher math, you know that impulse responses are a special class of functions. Also the Fourier theorem cannot be applied to non repeating signals.

 

[Ron Texas]

A 25 Watt per channel Class A amp draws about 200 watts at idle. That's enough to chase me off.

 

[Doodski]

A 25 Watt per channel Class A amp draws about 200 watts at idle. That's enough to chase me off.

LoL. Yes, they are gobblers. I had a 75W@8R class A and it sucked about 7 amps from the wall AC mains socket at idle. I shut the system down when not in use and it was a room heater too.

 

[SIY]

Also the Fourier theorem cannot be applied to non repeating signals.

That is absolutely incorrect.

I teach a course on this- perhaps if you're anywhere near western NY, you might think about sitting in.

 

[Ron Texas]

LoL. Yes, they are gobblers. I had a 75W@8R class A and it sucked about 7 amps from the wall AC mains socket at idle. I shut the system down when not in use and it was a room heater too.

Where I live the air conditioning runs 7 months a year. Just switching from incandescent to LED lighting made a difference in how the room cools.

 

[Palladium]

Well I can't hear past 8kHz now. I used to have electrostatic speakers which were a hybrid but reversed in that they used a tweeter that cut in at 7kHz. I bet if you disconnected it I would not be able to tell...

I saw a Bilibili video, where it featured a blind A/B of an untouched anime track versus it but with the 12KHz shelved up by a massive 6dB. There were still a large % of viewers who still couldn't heard a difference despite the site audience being heavily youth-oriented.

 

[Petrushka]

I saw a Bilibili video, where it featured a blind A/B of an untouched anime track versus it but with the 12KHz shelved up by a massive 6dB. There were still a large % of viewers who still couldn't heard a difference despite the site audience being heavily youth-oriented.

Henry Kloss famously said there was no useful content on LPs above 12 kHz.

Maybe it was infamously.

 

[DonH56]

@NTK
There it is again. "Continous" random sinewaves. What if the signals are not continous? If anyone has ever taken higher math, you know that impulse responses are a special class of functions. Also the Fourier theorem cannot be applied to non repeating signals.

Ah, there's the problem, nobody here has ever taken higher math.

The Fourier Theorem can be applied to non-repetitive signals. At least when I went to school. And used it on the job. For things like radar processing, which uses pulses, or sonar ("One ping only." - Red October). Edit: Take the university course Stuart @SIY teaches, or check out any number of references.

It is easy enough to predict the spectral response of impulses, step functions, and so forth in addition to sine waves, square waves, and all that jazz. Since real audio systems are bandlimited, Fourier works pretty well to predict time and frequency domain performance, and ideal impulses and steps cannot exist. Single-tone testing over frequency is a reliable predictor of performance. Two-tone (for IMD) and multitone testing is common, and of course there are things like the noise-power ratio tests and such that can be performed. There are a number of tests defined by the IEEE.

 

[Doodski]

Where I live the air conditioning runs 7 months a year. Just switching from incandescent to LED lighting made a difference in how the room cools.

I'm onboard with you. I had 37C (98.6F) the other day and I lay in bed in a pool of sweat. Had to take 1.5 snoozing pills that my doctor prescribed for me to get to sleep. 98.6F is not too harsh by your Texan standards but for me up north here it was frustrating and overbearing for all involved. I have fans but a AC unit would be so nice. I'm getting ready for ski season so money is going there. Regardless of the temps though I would love to live in Texas or Florida.

 

[MAB]

@NTK
There it is again. "Continous" random sinewaves. What if the signals are not continous? If anyone has ever taken higher math, you know that impulse responses are a special class of functions. Also the Fourier theorem cannot be applied to non repeating signals.

Outrageously and embarrassingly wrong!

Fourier Transform--Delta Function -- from Wolfram MathWorld

The Fourier transform of the delta function is given by F_x[delta(x-x_0)](k) = int_(-infty)^inftydelta(x-x_0)e^(-2piikx)dx (1) = e^(-2piikx_0). (2)

mathworld.wolfram.com

Which happens to be a derivative of another discontinuous function.

Every statement you made is wrong. What is your purpose here?

 

[Koeitje]

I look for artifacts in images and TV movies. Can’t stop myself. Too much time using Photoshop.

But there’s a cure for old eyes. I see better than I did at 14.

For video its much easier. I constantly see artifacts from video processing on TVs, mismatched framerates between source and display, frame pacing issues in games. You name it. You also don't lose vision that badly over age, because of glasses and such. High frequency hearing is just gone. You can boost it up with hearing aids, but compared to glasses they are just ass.