Matrix Audio X-Sabre Pro MQA

[asdf1000]

On 9/3/2019 at 6:35 AM, Miska said:

 

Most manufacturers will ignore it though. And improving results for this single measurement to improve ranking on this list can be counterproductive. It is just like anything else tested on single parameter, be it car tires, cars or audio equipment. And for example they test the DAC only with single input format.

 

 

I agree you can’t put everything into just one measurement.

 

But if we’re talking about this SINAD ranking, I do see an informal trend that the highest SINAD DACs also tend to do well in other measurements too...

 

It seems (just looking casually) that it takes good overall engineering competency to get a SINAD above 110dB...

 

Of course I’m not talking about sound quality at all. Just talking about meeting the manufacturer’s technical claims.

 

[Miska]

59 minutes ago, asdf1000 said:

But if we’re talking about this SINAD ranking, I do see an informal trend that the highest SINAD DACs also tend to do well in other measurements too...

 

It seems (just looking casually) that it takes good overall engineering competency to get a SINAD above 110dB...

 

SINAD (aka THD+N)... Within constraints of those tests... Those measurements don't tell for example how good performance the DAC is capable of when driven in a different way. But even looking at those measurements at ASR, the DAC in question wasn't so special.

 

For example I personally rather take modern AKM DAC chip than ESS Sabre, no matter how good 1 kHz THD+N figure someone squeezes out of ESS.

 

[asdf1000]

1 hour ago, Miska said:

or example I personally rather take modern AKM DAC chip than ESS Sabre

 

Why?

 

Is it because of the 'ESS Sabre IMD hump' that ASR has highlighted?

 

Or mainly because of 'DSD Direct' feature?

[Miska]

5 hours ago, asdf1000 said:

Why?

 

Is it because of the 'ESS Sabre IMD hump' that ASR has highlighted?

 

Or mainly because of 'DSD Direct' feature?

 

Mainly because of DSD Direct feature which allows getting proper performance. ESS modulator has some unfixable annoyances, for example related to it's ASRC/DPLL, as it's DSP cannot be bypassed completely, so the performance is more capped.

 

ESS digital filter has at best only about 100 dB attenuation, so it's reconstruction accuracy is really equivalent of about 16.5 bits. But that can be fixed externally.

 

[asdf1000]

4 minutes ago, Miska said:

ESS digital filter has at best only about 100 dB attenuation, so it's reconstruction accuracy is really equivalent of about 16.5 bits.

 

AKM is better in this regard?

 

[Miska]

10 minutes ago, asdf1000 said:

AKM is better in this regard?

 

Not with digital filters, it is the same. Both of course can be used with external one. But AKM has no ASRC/DPLL related problems. I prefer DACs that can be used as D/A converters, without built-in DSP. Otherwise their performance is capped by the DSP...

 

I'm very happy with my discrete DACs now though, both T+A and Holo Audio. I will also try Denafrips when I get around ordering one.

 

[Le Concombre Masqué]

23 minutes ago, asdf1000 said:

 

AKM is better in this regard?

 

and FPGA I guess?

[Miska]

Just now, Le Concombre Masqué said:

and FPGA I guess?

 

FPGA/CPLD is just plain programmable logic, so it doesn't have any performance as such. It totally depends on what goes there and what is around, since FPGA itself cannot be a D/A converter alone.

 

Most discrete DACs use FPGA/CPLD as a basis, because some digital logic is needed to make it happen and those programmable logics are a neat way to put it in a compact package that can be modified later.

 

 

P.S. I'm not criticizing ESS or Matrix as such, just the out of proportions hype ASR is making around SINAD ranking, based on single measurement result at single input format... IMO, it gives wrong kind of picture, performance and sonics are much more complex topic.

 

[opus101]

2 hours ago, asdf1000 said:

 

AKM is better in this regard?

 

 

Even the very latest AKM DAC (AK4499) has only -100dB stop-band rejection. Also note it uses an equi-ripple filter. These are known to exhibit pre-echoes - see a paper on this by Julian Dunn.

 

http://www.nanophon.com/audio/antialia.pdf

[Miska]

5 hours ago, opus101 said:

Even the very latest AKM DAC (AK4499) has only -100dB stop-band rejection. Also note it uses an equi-ripple filter. These are known to exhibit pre-echoes - see a paper on this by Julian Dunn.

 

Just like ESS... But equiripple doesn't have anything to do with that. You can also have minimum-phase equiripple filter that of course doesn't have pre-echo. Or any intermediate-phase one for that matter...

 

AKM also has minimum-phase (short delay) filters available on the chip, plus "low-dispersion short-delay" filter.

 

But I personally stay away from any DAC chip digital filters.

 

[edwoodjr]

I own the Matrix MQA and like it better than the Yggdrasil A2.  

[The Computer Audiophile]

On 9/2/2019 at 2:26 PM, asdf1000 said:

 

If it means manufacturers see it and try to improve measurements so they go to a higher “ranking”, that’s not a bad thing for end consumers?

 

How people buy things can be too complicated to even think about... whether there is a ranking or not.

 

 

This is the way many consumers think, but in reality it isn't the case. Think about rankings of anything and what manufacturers will do to be ranked higher. If it means making an unmeasured or unranked part of a product worse in order to make the measured part better, they will do it in a heartbeat. Before I went to law school I dissected the data in the US News Best Law Schools and rankings for all others. I new most schools info by heart. What a waste this was. When I learned what schools didi to get rankings, that had nothing to do with the quality of school or education, I realized I had wasted way too much time and anxiety. 

 

In audio, the most expensive / newest ESS chip is always  rated as the best by many people. Thus, some manufacturers put this chip in their DACs. However, the manufacturers may not implement it in the best way or make it any better than the least expensive ESS chip or chip from another manufacturer. Rankings / ratings always have unintended consequences. I'm not always against them, but I can't think of a time when I was for them. 

[The Computer Audiophile]

Oh wait, I'm 100% for rankings when it comes to @JoshM's The Best Version Of ... articles. I'll take more of those all day long. 

[asdf1000]

59 minutes ago, The Computer Audiophile said:

This is the way many consumers think, but in reality it isn't the case. Think about rankings of anything and what manufacturers will do to be ranked higher.

 

Noted but as I noted above, it appears (informally) that those DACs with SINAD above 110dB also do well in most of the other measurements he does...

 

So maybe it takes good overall engineering competency to achieve a SINAD above 110dB? 

 

 

Unless you mean all his measurements are useless?

 

[The Computer Audiophile]

8 minutes ago, asdf1000 said:

Unless you mean all his measurements are useless?

Oh no, I don't mean that.

[Miska]

12 minutes ago, asdf1000 said:

Noted but as I noted above, it appears (informally) that those DACs with SINAD above 110dB also do well in most of the other measurements he does...

 

I just don't agree about that...

 

[asdf1000]

21 minutes ago, Miska said:

 

I just don't agree about that...

 

 

Understood but can you share an example to discuss. 

 

Even the DAC that is topic of this thread?

 

None of the measurements matter?

[pkane2001]

30 minutes ago, Miska said:

 

I just don't agree about that...

 

 

THD+N is one, THD+N over frequency, IMD, jitter, SNR, multi-tone, IM over Level, linearity, DR, white noise with different filters, none of these are important? These are what Amir usually reports as part of his testing. 

[Josh Mound]

On 9/2/2019 at 5:24 PM, pkane2001 said:

 

Haven't seen AB's USB measurements. Mine was just to compare a Lush^2 to a generic USB cable. First post here, some follow-ups in the posts below it:

https://www.audiosciencereview.com/forum/index.php?threads/do-usb-audio-cables-make-a-difference.1887/post-216835

 

Comparison of a few balanced interconnects:

https://www.audiosciencereview.com/forum/index.php?threads/do-interconnect-cables-make-a-difference-a-null-test-result.7738/

 

 

 

Cool. Thank you for sharing. Here’s Bob’s post:

https://www.superbestaudiofriends.org/index.php?threads/usb-cable-shield-resistance-technical-measurements.5662/

[asdf1000]

52 minutes ago, The Computer Audiophile said:

Oh no, I don't mean that.

 

Understood. My point being it may not be that easy to fudge a SINAD of over 110dB and have all other measurements pretty decent too...

 

[pkane2001]

1 hour ago, JoshM said:

 

Cool. Thank you for sharing. Here’s Bob’s post:

https://www.superbestaudiofriends.org/index.php?threads/usb-cable-shield-resistance-technical-measurements.5662/

 

 

Interesting. Not sure how shield resistance translates into sound quality. Variable resistance that seems to change all by itself seems very strange, though

 

[opus101]

4 hours ago, The Computer Audiophile said:

Rankings / ratings always have unintended consequences. I'm not always against them, but I can't think of a time when I was for them. 

 

Summed up in Goodhart's Law : When a measure becomes a target, it ceases to be a good measure.

[asdf1000]

43 minutes ago, opus101 said:

 

Summed up in Goodhart's Law : When a measure becomes a target, it ceases to be a good measure.

 

No caveats?

 

So nobody should target lower distortion measurements then?

 

 

[opus101]

Looks like dodgy reasoning to me. When people target a particular distortion metric, that particular metric ceases to be one guaranteeing a good outcome for customers.

[asdf1000]

4 minutes ago, opus101 said:

Looks like dodgy reasoning to me. When people target a particular distortion metric, that particular metric ceases to be one guaranteeing a good outcome for customers.

 

It must be pure luck / coincidence that those DACs with SINAD over 110dB also have (mostly) very good other measurements... 

 

It can’t be good overall engineering... it must be luck / coincidence...