AK4499 vs ES9038Pro / Gustard DAC-X26 vs Gustard DAC-A22

[Martin]

I've wondered how these two flagship DAC chips compare but it's difficult to find them implemented in similar hardware. I thought the Musical Paradise MP-D2 MkII ($899) might be an interesting comparison vehicle as it has swapable DAC modules.

Musical Paradise has an ESS ES9038PRO module ($300) and is releasing an AK4499 module in mid-2020. This would allow one to measure the chip behavior in the same hardware. Unfortunately the tubed output may limit performance enough to negate any measurable differences.

Then along comes Gustard with the DAC-X26 (dual ES9038Pro):

And the DAC-A22 (dual AK4499):

Not sure how similar the architectures are but we do know Gustard knows how to build a well measuring DAC. So how do they compare:

SINAD X26 (ES9038Pro)

SINAD A22 (AK4499)

Dynamic Range X26

Dynamic Range A22

Jitter X26 left A22 right

Multitone:

Linearity:

THD+N-Freq:

IMD vs Level:

They look pretty evenly matched across the board. a few dB here, a few percentage points there, all-in-all they both measure very well and all differences definitely look to be below the threshold of human hearing.

I've read plenty of subjectivists comparing the "sound" of ESS and AKM DAC chips. A DBX test using these two DACs would either put those claims to rest or, inconceivably, prove they do indeed sound different.

Hope you found this at least enjoyable if not enlightening.

Martin

 

[Lotus97]

its like arguing that music from a PC with intel and amd sound different, dac chip is just a chip, and these days they can all be perfect at doing zero's and one's to put the music together.

it all comes down to many other components and how it is all put together, pre-out section is what makes all these units sound different and have their own sound signature.

I think we can all stop talking about DAC chips, and talk about everything else but the DAC chips in these units. I have seen the same DAC chip used in a $4k device and one in a $120 device, one sounded like pure crap, and other was pure bliss.

 

[Sak]

I think the digital filters of these DACs affect the perception of sound, so the reviews vary.

 

[Lotus97]

I think the digital filters of these DACs affect the perception of sound, so the reviews vary.

that is true but it goes beyond digital filters, the components and circuitry used outside of the DAC chip is what makes the music of one unit sound so much better than the other.

 

[solderdude]

Musical Paradise has an ESS ES9038PRO module ($300) and is releasing an AK4499 module in mid-2020. This would allow one to measure the chip behavior in the same hardware. Unfortunately the tubed output may limit performance enough to negate any measurable differences.

Most likely yes...

 

[FireLion]

Anyone ever use a Gustard X12? I picked up one and was highly impressed.

 

[Harmonie]

I've wondered how these two flagship DAC chips compare but it's difficult to find them implemented in similar hardware. I thought the Musical Paradise MP-D2 MkII ($899) might be an interesting comparison vehicle as it has swapable DAC modules.


I've read plenty of subjectivists comparing the "sound" of ESS and AKM DAC chips. A DBX test using these two DACs would either put those claims to rest or, inconceivably, prove they do indeed sound different.

Hope you found this at least enjoyable if not enlightening.

Martin

Hi Martin,

Digging out your old post above. What would be your conclusion 9038Pro<>AK4499 today ?

 

[Martin]

Hi Martin,

Digging out your old post above. What would be your conclusion 9038Pro<>AK4499 today ?

I’m of the mind that the chip inside a well implemented DAC should be completely transparent. I don’t believe I could distinguish between ESS Sabre and Asahi Kasei Microdevices DAC chips, or even Realtek, Burr-Brown, Wolfson or well designed R2R DACS.

Martin

 

[Harmonie]

Well there are some basic difference s no ? Like the velvet sound a bit softer less detailed but more bass in the AK vs the ESS for example ?

 

[Jimbob54]

 

[Jimbob54]

Well there are some basic difference s no ? Like the velvet sound a bit softer less detailed but more bass in the AK vs the ESS for example ?

If an AKM was the only DAC you ever heard, would you call it velvety or believe it was transparent? Also, I've never listened to velvet. Or air (apart from wind. Why is hifi never described as windy, only airy?)

 

[Lance768]

its like arguing that music from a PC with intel and amd sound different, dac chip is just a chip, and these days they can all be perfect at doing zero's and one's to put the music together.

it all comes down to many other components and how it is all put together, pre-out section is what makes all these units sound different and have their own sound signature.

I think we can all stop talking about DAC chips, and talk about everything else but the DAC chips in these units. I have seen the same DAC chip used in a $4k device and one in a $120 device, one sounded like pure crap, and other was pure bliss.

Most bullshit I ever heard. Dac chips has its own characteristic sound . An ess dac will sound like ess dac no matter where you put it.

 

[Harmonie]

Point is , what happens after the dac chip >> once analogue converted ... how is the analogue/preamp section implemented ?

 

[zzz2496]

A well designed DAC chip SHOULD be transparent, it SHOULD convert the digital data to analog in the most precise way as designed by the DAC chip designer. One thing that we mostly forgot, DAC chips has reconstruction filter. This is the filter that transforms digital data into it's analog shape. These reconstruction filters has their own quirks, that sometimes will show up in the music you hear - though I doubt it's THAT noticeable, but it's there. This can be seen in multi tone distortion test, those distortions are the result of reconstruction filter in action. Do not forget, Digital to Analog Conversion will introduce some sort of distortion to a certain degree, be it observable by human ear or not. I assume this is affects the resulting analog signal with "AKM" or "ESS" signature sound.

 

[Lance768]

But all the modern dac chips have thd+n less than -110db which is way less than what is noticable with human ear. Yet I can hear the difference in character from an akm and ess chips . Akm chips feels little livelier in midrange(well implemented ones). Ess ones are well extented with little bit of glare in mids.

 

[Newman]

Now match levels and repeat under blind listening conditions.

 

[Harmonie]

A well designed DAC chip SHOULD be transparent, it SHOULD convert the digital data to analog in the most precise way as designed by the DAC chip designer. One thing that we mostly forgot, DAC chips has reconstruction filter. This is the filter that transforms digital data into it's analog shape. These reconstruction filters has their own quirks, that sometimes will show up in the music you hear - though I doubt it's THAT noticeable, but it's there. This can be seen in multi tone distortion test, those distortions are the result of reconstruction filter in action. Do not forget, Digital to Analog Conversion will introduce some sort of distortion to a certain degree, be it observable by human ear or not. I assume this is affects the resulting analog signal with "AKM" or "ESS" signature sound.

Agree with your points.
But after analog conversion, there is still the output stage part and that part may be different from DAC to DAC, no ?

 

[zzz2496]

Agree with your points.
But after analog conversion, there is still the output stage part and that part may be different from DAC to DAC, no ?

I guess, if we really want to know, find a DAC chip implementation that we can tap directly to the DAC's chip output, and measure from there? That's the logical step we can try...

 

[zzz2496]

But all the modern dac chips have thd+n less than -110db which is way less than what is noticable with human ear. Yet I can hear the difference in character from an akm and ess chips . Akm chips feels little livelier in midrange(well implemented ones). Ess ones are well extented with little bit of glare in mids.

Sometimes we forgot that -110dB is just a number. The distortion, however minuscule, when it's additive to the content we're listening, we can heal weird artifact when we listen WITH content. Distortion cannot be measured without content. When you play something through the DAC, the distortion will distort the content, depending on the nature of the distortion, it can alter the content's sound to something "gritty", or "harsh", or "sibilant", or whatever term you want to use for it. To a sensitive ears, a 100dB content (say a music concert), but with added distortion on the "sibilant" frequency range around 3dB, will manifest as "This DAC is a tad bright". The distortion adds up to the total of 100dB, maybe the content it self is around 97dB with 3dB of distortion, but once you listen to that kind of loudness, anything can be heard.

I haven't tried anything else but my DAC (NAD M51), after 8 years of usage, I finally understood it's "voice", it is not transparent at very high level, it has really bad "crunch" sound to it (it is manifesting in the 2KHz+ freq graph when fed ONLY 1KHz test signal, you can refer your self to the test result, here at ASR - https://www.audiosciencereview.com/...ments-of-nad-m51-dac-and-digital-preamp.6681/ ), and it's quite bothersome. I am in the look out for my next DAC, still reading away at the forums looking for the one. If only MATRIX AUDIO made a multi channel Element X with DSP processing, so I can integrate my subwoofer properly (and perhaps analog inputs, both in balanced and unbalanced one?), I'll be a happy camper. Basically a super duper miniDSP SHD with more analog/digital outputs. I can only imagine...

 

[JohnYang1997]

Agree with your points.
But after analog conversion, there is still the output stage part and that part may be different from DAC to DAC, no ?

That's why we have measurements..