SMSL D-6s Balanced DAC Review

[horstman]

I wished Smsl offered that model without these somewhat exotic balanced outputs to get the price down a little; it would make a great competitor to the Topping E30II.

 

[Toku]

My D-6s arrived this past Friday. After trying a few filter settings, I've settled on F3. It seems to have the best balanced of extension and attenuation. Currently paired with a WiiM Pro. Obviously, I've connected an amplifier since taking this photo.

View attachment 324312

I sometimes see reports of pop noise or the beginning of songs being cut off with recent SMSL DACs or amplifier products equipped with DACs.
Does your D-6s have any symptoms that are worrisome?
I am also interested in the D-6S and am considering purchasing one.

 

[dsnyder0cnn]

I wished Smsl offered that model without these somewhat exotic balanced outputs to get the price down a little; it would make a great competitor to the Topping E30II.

They do. Sort-of. It's called the SMSL SU-1. ;-)

 

[Esprit]

Why doesn't the product appear on the manufacturer's website yet?

 

[horstman]

They do. Sort-of. It's called the SMSL SU-1. ;-)

You haven't really looked at the specifications and components, have you?
That's a completely different product.

 

[wavetrade]

Why doesn't the product appear on the manufacturer's website yet?

Seems really slow to get to Amazon US as well. I bought a DL200 while waiting but want a D-6S as well.

 

[dsnyder0cnn]

You haven't really looked at the specifications and components, have you?
That's a completely different product.

I have. my reply was somewhat tongue-in-cheek.

I guess if you really need Bluetooth, a display, remote control, selectable filters and DPLL, or have a strong preference for the ES9039Q2M over the AKM AK4493S DAC chip, the SU-1 is not a suitable replacement.

But both DACs support all of the same formats, including MQA CD (why that's a thing, I'll never understand) and probably sound almost identical. If you're not planning to use the DAC as a digital preamp and only need RCA outputs, the SU-1 will get the job done at less than half the cost.

 

[horstman]

I have. my reply was somewhat tongue-in-cheek.

I guess if you really need Bluetooth, a display, remote control, selectable filters and DPLL, or have a strong preference for the ES9039Q2M over the AKM AK4493S DAC chip, the SU-1 is not a suitable replacement.

But both DACs support all of the same formats, including MQA CD (why that's a thing, I'll never understand) and probably sound almost identical. If you're not planning to use the DAC as a digital preamp and only need RCA outputs, the SU-1 will get the job done at less than half the cost.

For a moment I forgot where I am - of course they all sound the same! ...

But yeah, apart from the preamp thing, a remote, the dual chip design, a very different PSU setup, the bluetooth and balanced output options (I could well do without) ... wait...
Actually, the SU-1 is a terrible deal these days.

I got the Smsl C100 for 100€ shipped, and still wished the Topping E30 II and Topping E30 II Lite had sounded a little more mellow/forgiving on my system, after comparing those units.
Topping just gets the UI done far better.

But the SU-1, I can't imagine why anyone would buy that thing, it's not cheap enough to chose it over say a Topping E30 II Lite.

 

[Esprit]

Seems really slow to get to Amazon US as well. I bought a DL200 while waiting but want a D-6S as well.

There are all models but not this one.

 

[dsnyder0cnn]

For a moment I forgot where I am - of course they all sound the same! ...

But yeah, apart from the preamp thing, a remote, the dual chip design, a very different PSU setup, the bluetooth and balanced output options (I could well do without) ... wait...
Actually, the SU-1 is a terrible deal these days.

I got the Smsl C100 for 100€ shipped, and still wished the Topping E30 II and Topping E30 II Lite had sounded a little more mellow/forgiving on my system, after comparing those units.
Topping just gets the UI done far better.

But the SU-1, I can't imagine why anyone would buy that thing, it's not cheap enough to chose it over say a Topping E30 II Lite.

Haha. All fair points.

I have both the Topping E50 and the SMSL D-6s. Love them both but for different applications. That we can get this level of subjective and objective performance for under $200 is astonishing to me. The rest of the world is a mess, but life is good for audiophiles and Hi-Fi enthusiasts here in late 2023.

 

[Toku]

Why doesn't the product appear on the manufacturer's website yet?

Some SMSL products are manufactured for specific distribution channels.
Those products will not be posted on SMSL's website, but will be posted on each sales company's website.
Its representative sales channels include Shenzhen Audio and Aoshida Audio.

 

[Toku]

For a moment I forgot where I am - of course they all sound the same! ...

But yeah, apart from the preamp thing, a remote, the dual chip design, a very different PSU setup, the bluetooth and balanced output options (I could well do without) ... wait...
Actually, the SU-1 is a terrible deal these days.

I got the Smsl C100 for 100€ shipped, and still wished the Topping E30 II and Topping E30 II Lite had sounded a little more mellow/forgiving on my system, after comparing those units.
Topping just gets the UI done far better.

But the SU-1, I can't imagine why anyone would buy that thing, it's not cheap enough to chose it over say a Topping E30 II Lite.

SMSL's SU-1 and Topping E30 ll are very inexpensive but excellent products that I also use regularly.
However, people have different preferences when it comes to audio, and I think the modern high-definition sound resolution of these products just doesn't suit your tastes. SMSL D300 is a product with the opposite sound quality.

 

[ceiber]

I recently bought this DAC as an external "upgrade" the DAC section of my secondary set up which includes a USB input to integrated amp's TI PCM 2704, and optical and coaxial inputs to CD player's CS 4398. I didn't want to spend much, like more than $500, since I think I wouldn't discern any difference. I also considered other latests SMSL DAC featuring AK chipset but just went with D6S since it's cheaper and has the features I wanted only though I fancied AK chipset sounds ... better.

The CS 4398 section of my CDP is fine though not SOTA at all, but, since the source is an old iPad running Tidal, I needed a USB input DAC. XLR and no headphone out spec of D6S also well cater to my need of XLR and situation (have plenty of headphone amp/outputs).

My impression has been clearer and better structured imaging and sound staging than before when I just used the internal DAC sections (TI is noticeably worse than CS). Big audible difference between SMSL D6S and internal CS 4398? Well, maybe or maybe not. But overall, just more enjoyable than before and I feel like I don't miss much now, which is very important for my audiophile insecurity

I actually very like the low profile, small form factor. It doesn't feel cheap. Remote works well, too It works flawlessly so far - I'm using COAXIAL from CDP and USB from iPad. not using BT.

Form factor, spec, latest chipset and SOTA measurement for the price i paid, I'm happy to live with this for long as far as I doesn't break.

My primary set up is NAD M33 but both primary and secondary sets run Klipsch Heritage speakers (sound awesome to me but I know it's not super accurate in many ways to catch small differences made by DAC or amps). Adding SMSL D6S to my secondary system gave me enough feeling of upgrade and the flexibility to play around further with other amps and sources I have.

I briefly ran Benchmark 1 or 2 a few years ago to the same secondary system (so paid about $1,000 for DAC then) and now I think I get more or so the same performance with less than $200 for this SMSL DAC, which is great.

 

[jayapple]

If the DL200 indeed shows the measurements we expect (expectations based on comparing it with similar specs of other DACs/headphone amps) I think it has a chance to elevate SMSL to another level (like the Fosi did with the V3 in the amp section). The DL200 will, hopefully, end up being THE best price to performance DAC/headphone amp. Which really makes me wonder why the company did not send Amir a unit to measure LONG time ago.

Sure, but no one test DL200.

 

[nanook]

The D-6s is advertized to have MQA support. The actual DAC ES9039 does not have a MQA renderer according to the ESS overview (the ES9069 does).
Not that I really understand this feature - at least the marketing sounds suspicious to me - rather I'm scared the "decoding" might be activated inadvertently.
I'm also unclear which component in the D-6s does the MQA decoding if it's not the ES9039.

I did use HDCD in the past and this technique did make sense to me since - having the data 96/24 available - you can try to hide some information in the anyway dithered floor when you go down to 44.1/16. As far as I followed this topic, MQA got applied to material that was available in just 16 bit depth, such that the algorithm might "invent" some details. When even upsampling to higher fs this gets even more speculative.
Maybe my understanding of the MQA process is wrong. These thoughts only to make plausible why I'm cautious with the feature MQA.

Take this scenario:
I use foobar and I frequently sample up to 96/24 in order to have no sharp roll-off close to the actual audio band. Most of the actual material is ripped CD, so 44.1/16. Resampler usually is SOX which asfaik does not dither, but let's assume there even is dithering involved.
-> What would the SMSL D-6s do?
- Would it switch to MQA, modifying the roll-off to match the one required for the first MQA step? Or does this need a certain signature hidden in the data?
- In case it can get activated inadvertently, can I turn off this feature in the settings of the D-6s ?

I have not yet ordered a D-6s, so I cannot look into this by myself.
Any comments appreciated.

 

[nanook]

Ok., I have read into MQA a bit in the meantime and it seems to me that
- properly applied, MQA may indeed improve sound quality
- the risk that the DAC switches its digital filter inadvertently is minimal. There is a signature in the audio data that is unlikely to be matched by coincidence.

 

[PyramidElectric]

The D-6s is advertized to have MQA support. The actual DAC ES9039 does not have a MQA renderer according to the ESS overview (the ES9069 does).
Not that I really understand this feature - at least the marketing sounds suspicious to me - rather I'm scared the "decoding" might be activated inadvertently.
I'm also unclear which component in the D-6s does the MQA decoding if it's not the ES9039.

I did use HDCD in the past and this technique did make sense to me since - having the data 96/24 available - you can try to hide some information in the anyway dithered floor when you go down to 44.1/16. As far as I followed this topic, MQA got applied to material that was available in just 16 bit depth, such that the algorithm might "invent" some details. When even upsampling to higher fs this gets even more speculative.
Maybe my understanding of the MQA process is wrong. These thoughts only to make plausible why I'm cautious with the feature MQA.

Take this scenario:
I use foobar and I frequently sample up to 96/24 in order to have no sharp roll-off close to the actual audio band. Most of the actual material is ripped CD, so 44.1/16. Resampler usually is SOX which asfaik does not dither, but let's assume there even is dithering involved.
-> What would the SMSL D-6s do?
- Would it switch to MQA, modifying the roll-off to match the one required for the first MQA step? Or does this need a certain signature hidden in the data?
- In case it can get activated inadvertently, can I turn off this feature in the settings of the D-6s ?

I have not yet ordered a D-6s, so I cannot look into this by myself.
Any comments appreciated.

According to the ESS datasheet, this DAC chip supports MQA (listed as an input on the schematic): https://www.esstech.com/wp-content/uploads/2023/03/ES9039Q2M_Datasheet_v0.1.3.pdf I think the MQA rendering will only be applied to files with an MQA flag, not sure how else it could work really.

 

[nanook]

I think that's a copy/paste. It's the signal path illustration that is identical to the figure in the ES9069 datasheet.
In the ES9039 datasheet MQA appears only in two figures
In the ES9069 datasheet MQA is described in more detail (37 occurrences of "MQA")
On the first page of the datatsheets MQA is in the feature list for the ES9069, not however in the one of ES9039
There is a nice overview: https://www.esstech.com/products-overview/digital-to-analog-converters/sabre-audiophile-dacs/

This raises the question: Does the D-6s indeed support MQA?
Since specs and pinout of the two chips are otherwise identical at first glance, it could as well be the D-6s is build around an ES9069, but marketing mentions the well known ES9039?

 

[staticV3]

properly applied, MQA may indeed improve sound quality

It may not. It is lossy compression which inadvertently reduces fidelity.
Their claims of compensating for studio gear make no sense in the context of how a song is created.

For more info, search "Goldensound MQA" on YouTube.
It's a two-part analysis of what MQA is actually doing to the audio file.

This and especially the second part about MQA's response should tell you everything you need to know about the company.

 

[staticV3]

I think that's a copy/paste. It's the signal path illustration that is identical to the figure in the ES9069 datasheet.
In the ES9039 datasheet MQA appears only in two figures
In the ES9069 datasheet MQA is described in more detail (37 occurrences of "MQA")
On the first page of the datatsheets MQA is in the feature list for the ES9069, not however in the one of ES9039
There is a nice overview: https://www.esstech.com/products-overview/digital-to-analog-converters/sabre-audiophile-dacs/

This raises the question: Does the D-6s indeed support MQA?
Since specs and pinout of the two chips are otherwise identical at first glance, it could as well be the D-6s is build around an ES9069, but marketing mentions the well known ES9039?

MQA decoding can be done either on the DAC chip, or on the XMOS USB bridge, which should be the case on the D-6s.