Do solid state amplifiers sound the same ?

[XTF1]

i cant speak about apps on android but a tweaked Linux Kernel definitely makes a minor difference
i tested the default debian kernel vs liquorix and xanmod kernel
both tweaked kernels sound better than default, with liquorix a smitch better than xanmod

I love your posts. Everything turns out to have a noticeable impact.

@Ghoostknight Whatever you are reporting about, does it sound different after you post about it as compared to before posting?

 

[medon]

..., more so now that modern DACs use their own clock instead of the computer's.

I don't intent to go into the question of audible effects (proper ABX tested and so on...).
Another disclaimer: I am NOT an electronics engineer. Nor DAC designer.
That having said: Even with asynchronous transfer of data (USB) from source to DAC, some sort of timing alignment HAS to happen.
With synchronous transfer (everything else) anyway --> PLLs.

So far, so true?

 

[BS5711]

People need to better understand correlation versus causation.

Sometimes the correlation is founded only in psychology.

 

[bigshot]

I’ve never heard of any home audio component with audible levels of jitter. Everything I’ve heard of is an order of magnitude below audibility. Jitter was a popular selling point over five years ago, but it isn’t mentioned much any more because it’s turned out to not be an issue.

 

[Ghoostknight]

@Ghoostknight Whatever you are reporting about, does it sound different after you post about it as compared to before posting?

no usually many A/B tests were involved before writing about stuff like this just sharing my expierence YMMV

 

[medon]

I’ve never heard of any home audio component with audible levels of jitter. Everything I’ve heard of is an order of magnitude below audibility. Jitter was a popular selling point over five years ago, but it isn’t mentioned much any more because it’s turned out to not be an issue.

Again -- no ABX testing done.
Still I'd advise anyone to try different digital interconnect in his/her system, if available. Most streamers / DACs offer more than one choice of being connected.

No audible difference to be heard? Fine.

Jitter not being an issue any longer might be true for many devices, but for all?
Sure mine are broken...

 

[bigshot]

There have been audibility studies by the AES. I don’t have an account there, but if you are interested you might want to open one.

 

[medon]

There have been audibility studies by the AES. I don’t have an account there, but if you are interested you might want to open one.

Nope... don't have one either.
Well then, this is also settled -- some studies about jitter, done at a time in place, done maybe decades ago, with certain equipment, certain people, revealed that jitter was not an issue in the sense of statistical relevance within that testing setup.
This does IMHO not rule out the possibility that other equipment, in other circumstances, behaves differently.

 

[castleofargh]

Nope... don't have one either.
Well then, this is also settled -- some studies about jitter, done at a time in place, done maybe decades ago, with certain equipment, certain people, revealed that jitter was not an issue in the sense of statistical relevance within that testing setup.
This does IMHO not rule out the possibility that other equipment, in other circumstances, behaves differently.

You try stuff, casually, you feel a change, and you talk about jitter. I'm not even sure we can acknowledge correlation from that, let alone causation. Plus, what does that have to do with amplifiers?
If you need to know for a fact about your own system in your own house, get a measurement rig(for jitter I'm not sure you can hope for something credible below maybe 400$, and ideally, real expensive gear like some Audio Precision rig would be better as this is such a tiny thing to try and measure very accurately(why it doesn't justify making a big deal out of it in the first place).
If you plan to experiment with actual listening tests, you can start with Distort https://distortaudio.org/ and try various settings for yourself. Of course, even with that, one can use it casually and come up with any sort of conclusions from "hearing" all the scary looking graphs. I can try to explain the need for actual listening tests all my life, I can't force anybody to actually do something correctly from time to time.

Just worrying because you think there's a link between casual impressions and variable X, without even a clear concept of the magnitudes involved or when we humans can detect them, I cannot see how that's ever going to serve your enjoyment of music, your curiosity about objective stuff, or your desire for fidelity.

As a special case, if you go fool around with inputs that are locked on the computer clock, and it sucks, when there's a fine USB right next to it, I'd say you're not demonstrating that jitter matters, but that you might have a secret passion for masochism.



Here is what my SS amplifier sounds like playing Sultan of swing, a real wire with gain:

 

[Vamp898]

You try stuff, casually, you feel a change, and you talk about jitter. I'm not even sure we can acknowledge correlation from that, let alone causation. Plus, what does that have to do with amplifiers?
If you need to know for a fact about your own system in your own house, get a measurement rig(for jitter I'm not sure you can hope for something credible below maybe 400$, and ideally, real expensive gear like some Audio Precision rig would be better as this is such a tiny thing to try and measure very accurately(why it doesn't justify making a big deal out of it in the first place).
If you plan to experiment with actual listening tests, you can start with Distort https://distortaudio.org/ and try various settings for yourself. Of course, even with that, one can use it casually and come up with any sort of conclusions from "hearing" all the scary looking graphs. I can try to explain the need for actual listening tests all my life, I can't force anybody to actually do something correctly from time to time.

Just worrying because you think there's a link between casual impressions and variable X, without even a clear concept of the magnitudes involved or when we humans can detect them, I cannot see how that's ever going to serve your enjoyment of music, your curiosity about objective stuff, or your desire for fidelity.

As a special case, if you go fool around with inputs that are locked on the computer clock, and it sucks, when there's a fine USB right next to it, I'd say you're not demonstrating that jitter matters, but that you might have a secret passion for masochism.



Here is what my SS amplifier sounds like playing Sultan of swing, a real wire with gain:

I use my PreSonus Studio 68c. It costs roughly 200~250 bucks and is good enough for the job. Here is an Jitter test from my 50$ Dongle which performs excellent by the way

 

[VNandor]

I don't intent to go into the question of audible effects (proper ABX tested and so on...).
Another disclaimer: I am NOT an electronics engineer. Nor DAC designer.
That having said: Even with asynchronous transfer of data (USB) from source to DAC, some sort of timing alignment HAS to happen.
With synchronous transfer (everything else) anyway --> PLLs.

So far, so true?

Not really.

The timing of sending data asynchronously can not have an influence on the timing of digital to analog conversion in any way at all. The data is sent from the PC's memory to the DAC's memory. The packet usually contains a couple hundred of audio samples along with some control data to make to protocol work. For the most part, no data is being sent to the DAC, it only needs the next packet whenever the couple hundred of samples in the buffer has been converted. So instead of sending 1 audio sample every 1/48000 seconds, the data is sent every ~512/48000 (as an example) seconds instead. This gives plenty of leeway to send the packet again if not received correctly. If somehow the packet is not received in time, the DAC can't wait around because the conversion is controlled by the DAC's clock. The data in the DAC's buffer will get converted into an analog voltage even if the memory is just filled with garbage/random bits instead of the correct audio sample. So instead of the DAC somehow "slowing down" to wait for receiving the correct audio sample just outputs noise, cracking audio, or the last samples stuck in the buffer over and over again.

 

[Vamp898]

Not really.

The timing of sending data asynchronously can not have an influence on the timing of digital to analog conversion in any way at all. The data is sent from the PC's memory to the DAC's memory. The packet usually contains a couple hundred of audio samples along with some control data to make to protocol work. For the most part, no data is being sent to the DAC, it only needs the next packet whenever the couple hundred of samples in the buffer has been converted. So instead of sending 1 audio sample every 1/48000 seconds, the data is sent every ~512/48000 (as an example) seconds instead. This gives plenty of leeway to send the packet again if not received correctly. If somehow the packet is not received in time, the DAC can't wait around because the conversion is controlled by the DAC's clock. The data in the DAC's buffer will get converted into an analog voltage even if the memory is just filled with garbage/random bits instead of the correct audio sample. So instead of the DAC somehow "slowing down" to wait for receiving the correct audio sample just outputs noise, cracking audio, or the last samples stuck in the buffer over and over again.

With some DAC and devices, you can even select the buffer size. Some Models have insane buffer sizes and modify it, when there is an underrun.

The Walkman for example usually has 350ms of Buffer but extends endlessly if there is an underran. My Record with an really bad unstable USB Cable that was wiggling a lot during a flight was 1.5s of Buffer.

I constantly had to adjust the delay in mpv

 

[VNandor]

Interesting, so this DAC does exactly what I've claimed no DAC does. My answer to that would be to get a DAC that has its driver made by a sane person.
The buffer size selection is typical feature though. In some use cases even 10ms delay could be untolerable while for playback only, hundreds of milliseconds can be fine and giving the option for the user to select the preferred buffer size accomodates that.

 

[Vamp898]

Interesting, so this DAC does exactly what I've claimed no DAC does. My answer to that would be to get a DAC that has its driver made by a sane person.
The buffer size selection is typical feature though. In some use cases even 10ms delay could be untolerable while for playback only, hundreds of milliseconds can be fine and giving the option for the user to select the preferred buffer size accomodates that.

It has no DAC though, it is an full digital amp and i am sure that this buffer is implemented in the OS layer.

So i should have made that clearer. If you use the Walkman as an DAC (It has an DAC Mode) than it does have this behavior. The same is true for the FiiO M11 Plus where you can select the buffer size in the driver.

But those are both Android devices, so i am pretty sure this buffer is even before the decoding is done and an OS implementation.

My DAWN PRO for example doesn't have this feature.It just starts to stutter if there is an lack of data without increasing any buffers

 

[medon]

Not really.

The timing of sending data asynchronously can not have an influence on the timing of digital to analog conversion in any way at all. ...

Well...
... on an abstract level.

Still, there's various degrees of measured jitter in USB audio, see the measurements published by Golden One.
Link to some MA3 DAC below, others easily found on his website.

How come, if everything's perfect and "The timing of sending data asynchronously can not have an influence on the timing of digital to analog conversion in any way at all"?
(and NO, it's not broken)

https://goldensound.audio/2024/06/08/meitner-ma3-dac-measurements/