Power amplifier THD+n tests

[mike7877]

What brought this on is I recently bought the Topping PA3s, and am disappointed with its performance.
Some people say all amps sound the same - this thread is not about that.

Supposedly the difference between the top THD+n figure (1kHz 5W / 4 ohm), and the THD+n in multi frequencty/power graph, is from test bandwidth

- the THD+n measurement at the very top is done @ 22kHz, multi tone: 45.

The two images below show a 13dB diference
*The edit is bolding and making larger the text above to make more obvious the reason for this thread


If there wasn't a measurement error in the PA3s review, it seems it would be best to do both measurements 22kHz bandwidth, so that the figure representsrepresents the most relevant distortion: distortion in the audible range
* The edit is underlined as it's the proposed solution.
...for clarity.
For further clarity, this thread is not about the PA3s at all - it's just the example which brought this issue to my attention

(The top number is into 4 ohms, the 6.3V p-p proves it. Often the 8 ohm measurement is ~3dB better, but not in the PA3s's case - there's a (very odd) 10dB increase in distortion between ~6 and ~8 watts into 4 ohms, and at 10 watts 4 ohms (= 5w 8), distortion seems to be 7 or 8dB worse)

 

[notsodeadlizard]

This amp is a kind of really single-chip design because ST Micro's TDA7498E $5 in volumes IC is a complete class D amplifier.
ST Micro claims 0.05% THD at 1W output power:

What do you want from this rattle that is used in TVs and other things that do not require a high quality audio path?
You can’t get better than what is stated in the datasheet from this amplifier (but you can always make it worse, the possibilities for this are almost limitless):

 

[notsodeadlizard]

And, of course, it's weird and ugly, THD goes up in the most audible frequency range, yep:

 

[voodooless]

The difference between dashboard and THD+N vs frequency and Power is because the latter has a wider measurement bandwidth. The multitone is clean enough:

 

[mike7877]

The difference between dashboard and THD+N vs frequency and Power is because the latter has a wider measurement bandwidth. The multitone is clean enough:

View attachment 293264

Please read the OP

 

[mike7877]

The difference between dashboard and THD+N vs frequency and Power is because the latter has a wider measurement bandwidth. The multitone is clean enough:

View attachment 293264

Apparently you didn't read the OP either. Or chose to just say something irrelevant

It's about the 13dB difference in 1kHz 5W distortion, and choosing 22kHz bandwidth for both of the tests in the OP


Or 45, if tat's what's deemed better. I think 22 because it's distortion in the audible range (like I said in the OP)

Do you have an opinion on this thread's subject?

Edit: sorry for being short. It's not you. Well it is a bit. It's more the trend of so many of the threads I start, the first replies are not replies, rather statements that are only somewhat relevant or even entirely irrelevant.
Then I have to correct the first replies because if I don't, other people skim the OP too, and then read the first replies, getting the wrong idea about what the thread is about, and BAM: thread?
Derailed!
This one's barely a paragraph, too

First forum rule should be "no reply until entire OP is read" lol

 

[voodooless]

You need 45 kHz bandwidth to measure distortion at 15 KHz and higher. Otherwise you’ll need to limit to a max of 10 kHz or so.

That’s why I rather look at the multitone and brought it up. It’s a much more realistic view of how it handles real music in the audible range.

If there is a big difference between the 22 kHz and 45 kHz bandwidth measurements, it means there is quite a lot of things going on above 20 kHz. That should not be audible indeed, but that’s not really what that graph is about. You’ll probably not hear any distortion above 6~7 kHz at all…

We also do don’t use A-weighting, which would make sense for audibility.

 

[mike7877]

You need 45 kHz bandwidth to measure distortion at 15 KHz and higher. Otherwise you’ll need to limit to a max of 10 kHz or so.

That’s why I rather look at the multitone and brought it up. It’s a much more realistic view of how it handles real music in the audible range.

If there is a big difference between the 22 kHz and 45 kHz bandwidth measurements, it means there is quite a lot of things going on above 20 kHz. That should not be audible indeed, but that’s not really what that graph is about. You’ll probably not hear any distortion above 6~7 kHz at all…

We also do don’t use A-weighting, which would make sense for audibility.

22 is used for the charted spec (hierarchy)

We don't really need 15kHz because 2nd harmonic is inaudible.

Maybe two charts, one with 22 and one with 45.
If only one it should be 22, because it's clearly more relevant

Edit: 31, 62, 125, 250, 500, 1k, 2k, 4k, 8k (Hz) would be best

 

[mike7877]

You need 45 kHz bandwidth to measure distortion at 15 KHz and higher. Otherwise you’ll need to limit to a max of 10 kHz or so.

That’s why I rather look at the multitone and brought it up. It’s a much more realistic view of how it handles real music in the audible range.

If there is a big difference between the 22 kHz and 45 kHz bandwidth measurements, it means there is quite a lot of things going on above 20 kHz. That should not be audible indeed, but that’s not really what that graph is about. You’ll probably not hear any distortion above 6~7 kHz at all…

We also do don’t use A-weighting, which would make sense for audibility.

Importance is look at how much more distortion there is for even 1kHz with 45kHz bandwidth: 13dB (100% inaudible distortion). That's huge! And so inaccurate. As these reviews are first supposed to reflect audio quality... unless I'm missing something.

Sure, unweighted and high bandwidth gives insight into build quality/characteristics, but that's only important insofar as the spec doesn't detract from the device's purchase in the first place.
(If someone doesn't buy it... does a tree make a sound if noone's around to hear it?)

1: 22
2: 22/45

 

[voodooless]

As these reviews are first supposed to reflect audio quality... unless I'm missing something.

They show engineering excellence regarding these measured quantities. No audibility claims are generally made about these measurements.

 

[fpitas]

They show engineering excellence regarding these measured quantities. No audibility claims are generally made about these measurements.

In fact, no one has ever demonstrated that human ears can detect anything below about 70dB SINAD.

 

[mike7877]

They show engineering excellence regarding these measured quantities. No audibility claims are generally made about these measurements.

When people read the reviews it's to determine how one particular audio device might sound.

Maybe it's just me though.

And before anyone gets excited, I said might, not will.

Q: In the device hierarchy, which bandwidth is used?
A: 22

Q: Why?
A: go for it

 

[mike7877]

In fact, no one has ever demonstrated that human ears can detect anything below about 70dB SINAD.

So the bottom of all charts here should be -70?

 

[mike7877]

Listen, I'm arguing for consistency and audibility. Consistency with what is already the benchmark, and audibility because these are audio devices. NOBODY should be disagreeing.

22 and 45 both have their place (as I've already said - you're not countering anything), but if there's only one chart, it should be 22: most audible - for the audio devices

 

[voodooless]

So the bottom of all charts here should be -70?

No, because audibility is not the (primary) objective of the review.

 

[fpitas]

So the bottom of all charts here should be -70?

If all you care about is audibility.

 

[fpitas]

Ultra extreme audio specs

I reckon that most If not all people do not have the physical ability to hear the difference between 120 SINAD vs. 95 SINAD, so why bother with 120 SINAD? Because I personally appreciate the engineering chops when you can get to 120 SINAD. It makes me feel good and have pride of ownership of...

www.audiosciencereview.com

Asks a lot of the same questions about measurements beyond audibility.

 

[mike7877]

If all you care about is audibility.

Disengaging

 

[mike7877]

Ultra extreme audio specs

I reckon that most If not all people do not have the physical ability to hear the difference between 120 SINAD vs. 95 SINAD, so why bother with 120 SINAD? Because I personally appreciate the engineering chops when you can get to 120 SINAD. It makes me feel good and have pride of ownership of...

www.audiosciencereview.com

Asks a lot of the same questions about measurements beyond audibility.

Of course, but what about a chart showing worse than -70 when in actuality it's better than -70? The PA3s for example. And a lot of other devices.

I already said having both 22 and 45 is better, but if there's one it should be 22.

Argue for only 45 being better, taking into account that a lot of devives are being pushed to worse than -70 when in reality (audibly) they're better, in your response.
Unless that's not your opinion.
I've said all that needs to be said, unless someone says something compelling

 

[mike7877]

No, because audibility is not the (primary) objective of the review.

Audio devices are for making audio. Of course you can investigate past what's audible.

Tell me why one wouldn't measure with 100MHz bandwidth