Master Thread: Are measurements Everything or Nothing?

[pkane]

No, every time I say something here I get shit load of posts making assumptions about my ability and motives.

Yes, every time I say something here, or even just post measurement results without making any further statements about those, I get under hostile attack wave.

Yes, at most I will say that based on my experience -120 dB things are audible. I already said it is totally blind AB test. More details I won't divulge.

When someone questions your results, it's nothing personal, they may want to repeat your findings, or understand what you did. But your reaction to questions is violent, and purely personal, you treat everyone here as if they are the enemy. Even said it in so many words.

Here are a few statements that may help you understand the skepticism you're facing to your posts:

Extraordinary claims require extraordinary evidence​

That which can be asserted without evidence, can be dismissed without evidence.​

If you want to share information, you'll need to do more than just drop extraordinary claims and then refuse to discuss them.

 

[Miska]

If you have evidence for a groundbreaking observation, than why keep it for yourself? Are you running a business and don't want your competitors to know?

Yes

Unfortunately Chord Electronics is already working on the -300dB level (without evidence of course ).

In digital domain, nothing special. So do I, for example HQPlayer has some digital filters with > 300 dB stop-band attenuation.

 

[Miska]

That which can be asserted without evidence, can be dismissed without evidence.

Just dismiss it...

 

[Miska]

Because you fucked up that particular measurement and won't go figure out how you managed that and repeat it properly.

I did repeat it several times, but it doesn't become any better

 

[pkane]

Just dismiss it...

done.

 

[audio2design]

Ahh, always good excuse when the results don't please you. Somehow, on the exact same setup, tens of other DACs don't have these issues.

There was an obvious lack of consistency of the one measurement at -60db with higher noise flow and that is the most contentious result. I pointed that out, @SIY pointed it, perhaps others did as well. The FFT length or window function is obviously different from the others. Some of the other results look like artifacts from software volume control and hence could be system / setup related.

Extraordinary claims, extraordinary evidence.

This whole playing the victim thing is unattractive.

 

[SIY]

I did repeat it several times, but it doesn't become any better

So the error is systematic, assuming you didn't make that up as well. And as long as you refuse to get into details about your test setup, measurement process, and settings so that the error source can be determined, it means that the other measurements are, at best, questionable. But this particular graph was a red flag that you're doing it wrong.

 

[Blumlein 88]

If you go back, you can see, I posted graphs for 44.1k alt setting 1, 352.8k alt setting 1 and DSD256 alt setting 3 (PCM rate 352800). Mentioned in the graph headers.
View attachment 181245
View attachment 181246
View attachment 181247

Of those, 44.1k PCM was worst performer, 352.8k PCM was a bit better but had picketed noise floor, and DSD256 was cleanest but still abnormal noise floor modulation.

In addition I think I posted graph showing how 19+20 IMD test tone at 705.6k PCM trips up the output.
View attachment 181248

PCM input is 24-bit TPDF dithered data.

So can any of you with a Topping D90SE confirm the results? The easiest would be the 44.1 khz result.

 

[Miska]

So the error is systematic, assuming you didn't make that up as well. And as long as you refuse to get into details about your test setup, measurement process, and settings so that the error source can be determined, it means that the other measurements are, at best, questionable. But this particular graph was a red flag that you're doing it wrong.

Ahh nice, but I'm still happy to find that with many other DACs the same "doing it wrong" procedure works without problems.

Interesting that only with this product, my wrong doing is visible.

 

[pkane]

Ahh nice, but I'm still happy to find that with many other DACs the same "doing it wrong" procedure works without problems.

Interesting that only with this product, my wrong doing is visible.

Easy. If anyone else can duplicate your result, then it's confirmed. If they can't, you have a bad unit, messed up measurements, or the sun spots interfered with your measurements that day. In any case, why the animosity? I suspect you've gotten too used to the audience on AS where they believe anything and everything you say. That's not how things work in the real world.

 

[Killingbeans]

OK, let's put it this way. How much you see ESS, AKM or TI documenting and publishing stuff about their subjective audio testing?

Or for example my ex employer Nokia?

None.

Are you talking about marketing blurbs descibing sonic signatures? Don't think anybody in here is taking those at face value.

No, every time I say something here I get shit load of posts making assumptions about my ability and motives.

Personally I couldn't care less about you abilities and motives. I'm totally open to the idea of the D90 having some sort of defect that slipped through the cracks. But I'd like to have comfirmation of you measurements not being a glitch, and more importantly, that it's audible. Otherwise it doesn't really leave room for more than loose assumptions.

 

[rgdawsonco]

I don't think 5db is particularly significant

You and I are in agreement.

 

[Blumlein 88]

So Archimago has tested the Topping D90 SE. He didn't use 44.1 khz, but did use 96 khz at -60 db. It didn't show any problems. And in fact had a lower noise level than any of the DSD rates for the -60 db test tone at 1 khz.

MEASUREMENTS / REVIEW: Topping D90SE DAC - Part II: Resolution

A blog for audiophiles about more objective topics. Measurements of audio gear. Reasonable, realistic, no snakeoil assessment of sound, and equipment.

archimago.blogspot.com

About 2/3rds of the way down the page.

So is this only a problem for PCM multiples of 44.1 khz in this DAC?

 

[Miska]

Easy. If anyone else can duplicate your result, then it's confirmed.

Yes, so why so much noise?

If they can't, you have a bad unit, messed up measurements, or the sun spots interfered with your measurements that day.

Yeah, I spent entire one day measuring the device. If I have bad unit, then there's a QC problem. But it doesn't explain why Archimago got suspicious results too.

OTOH, I don't lightly leave a 900€ device in a box. So I hope at least some of the issues can be fixed with firmware update or such. I suspect at least some of the issues could be related to XMOS controller and the MQA code there.

In any case, why the animosity? I suspect you've gotten too used to the audience on AS where they believe anything and everything you say. That's not how things work in the real world.

Instead of getting response like "nice work, I will check my D90SE with same measurements" I get response in the spirit of "you fucking moron, you don't know what you are doing and you are full of shit, and even if it would be right it wouldn't matter because your cannot hear it, but still the Amir's SINAD ranking chart is the only thing that matters".

 

[Miska]

So Archimago has tested the Topping D90 SE. He didn't use 44.1 khz, but did use 96 khz at -60 db. It didn't show any problems. And in fact had a lower noise level than any of the DSD rates for the -60 db test tone at 1 khz.

But noise level changes for example between the PCM rates. Which is not what ESS DAC chips do. That's a hint. Also DSD256 should provide lower noise floor that DSD64. If it doesn't, something is going wrong. I suspected first that this is because the E1DA ADC has so low stop-band attenuation for it's decimation filter, that would make ultrasonic noise alias down to audio band in measurements. One shouldn't use ADC chip designed for recording music for measurement purposes directly, because the anti-alias filter wasn't designed with expectation that there would be notable levels of ultrasonic input (from micrphones). For this same reason, I also have a separate analog filter module (made by the same company making the measurement gear) for my measurement gear, for measuring class-D amplifiers. But instead of finding that noise change to be measurement glitch in his measurements, I found it to be (at least partially) issue in the DAC. I didn't expect such result.

Let's compare same multi-tone test with both 352.8k PCM DSD256 on two DACs with the same ES9038PRO chip:

You notice with D90SE the noise floor jumps. While with the M500 mkII it stays level (as expected).

 

[Miska]

Are you talking about marketing blurbs descibing sonic signatures? Don't think anybody in here is taking those at face value.

I'm talking about "what audiophiles are hearing", and listening tests of modulators in general.

 

[pkane]

Yes, so why so much noise?

Perhaps because you immediately attacked the people that asked for details?

Instead of getting response like "nice work, I will check my D90SE with same measurements" I get response in the spirit of "you fucking moron, you don't know what you are doing and you are full of shit, and even if it would be right it wouldn't matter because your cannot hear it, but still the Amir's SINAD ranking chart is the only thing that matters".

Maybe it's a cultural/language difference, but I didn't see anyone calling you names in this thread, but I did see you attacking others based on this straw-man "ASR member" that you apparently constructed in your head.

 

[Blumlein 88]

I'm talking about "what audiophiles are hearing", and listening tests of modulators in general.

I read one of the developers at Sony say very low level differences were surprisingly audible due to interactions between dither and modulators. And that the difference sometimes took a few minutes of playing to reveal itself. Is that related to the differences you say are audible at very low signal levels?

 

[Miska]

I read one of the developers at Sony say very low level differences were surprisingly audible due to interactions between dither and modulators. And that the difference sometimes took a few minutes of playing to reveal itself. Is that related to the differences you say are audible at very low signal levels?

Many things, also dynamic behavior with non-constant signals is different with different modulators. I would say large part of sonic differences between for example AKM and ESS account for differences in the modulator behavior. Another is likely just the conversion stage, such as SCF voltage output stages vs multi-element constant current source stages, plus effect of DEM stage on their behavior.

 

[Miska]

Perhaps because you immediately attacked the people that asked for details?

Maybe it's a cultural/language difference, but I didn't see anyone calling you names in this thread, but I did see you attacking others based on this straw-man "ASR member" that you apparently constructed in your head.

I only respond in similar tone.