SMSL SU-10 DAC Review

[Robbo99999]

DAC performance surpassed the threshold of audibility many, many years ago. Manufacturers will continue to build better performing DACs and no one will be able to hear any differences, not only due to the DACs’ absolute performance but also particularly because downstream equipment is almost always noisier and produces more distortion. So it’s not about the ability to “eke out the last remnants of SINAD,” but rather the ability to produce gear with lower amounts of noise and distortion. In other words, what manufacturers have always done. Marketing hi-fi is often numbers-driven. Remember the 80’s when every single amplifier sell sheet proudly proclaimed its THD percentages? Same thing.

Even so, that doesn't mean we have to go along with the "numbers-driven advertising". I'm basically saying that I'd like to see these companies offer a real point of difference as we know the SINAD race is won already in practical terms.......so real point of difference would be features/connectivity/(aesthetic)/DSP/etc. I don't mind companies offering simple DACS that measure great and are well-priced, but expensive DACS that don't have special features beyond squeezing out the last possible SINAD don't figure with me, and probably not with experienced users here on ASR, or even anyone that understands ASR.

 

[xnor]

It is a parody that we already have three poor votes…

If I had to score just based on a few measurements then I'd also vote for great, but there's also things like the price or MQA...

 

[kaopad999]

Wow! well done SMSL. Dynamic range of 130 db is bloody fantastic! and thank you so much for the review @amirm

 

[kaopad999]

hope to see a DAC that lets you apply EQ, maybe even FIR EQ!

That would be an excellent implementation to future DACS!

 

[phoenixdogfan]

I haven't done the math but I think we are at the wall.

Time for Audio Precision to step up their game.

 

[sarumbear]

If I had to score just based on a few measurements then I'd also vote for great, but there's also things like the price or MQA...

Do you mean any DAC that doesn’t support MQA is poor, irrespective of anything else?

 

[pseudoid]

Based on company recommendation, I upgraded the unit to latest firmware.

I assume you did this firmware upgrade before the start of testing...
[I am the apprehensive-type; when products are released into the consumer market, before fully engineered/designed.]
I wonder if there were any measurable differences between old and new 'firmware'.
[Only because I have always had an axe to grind about never-ready-for-prime-time software, especially firmware.]

Would it be proper for you to add the software versioning information as a part of the measurements, for historical perspective?

 

[NiagaraPete]

If this had an aes output it would be the cat’s meow.

 

[SalutaryGraveness]

The same company also offers some cheap models as low as $92~$200. I wonder if they can be as good in most aspects. Particularly in terms of distortion and noise.

 

[57gold]

Looks pretty tidy inside. Beefy looking transformers, clean layout...

 

[Spocko]

This is a review and detailed measurements of the SMSL SU-10 Balanced DAC with Bluetooth. It was sent to me by the company and costs US $900.
View attachment 238217
Front panel is smooth and slick and sports the new colorful display from SMSL with large volume level which I very much appreciate. Interestingly volume goes up another 2 dB (or so) so you have a bit of headroom. Back panel shows some new features:
View attachment 238218
For the first time we have a choice of USB-B and -c inputs which you can select from the menu. Another new feature is ability to turn on the balanced or RCA outputs independently or together. Nice!

The high cost is partially due to use of not one, but two ESS ES9038Pro DAC chips. The 8 channels in each chip is combined to get the best dynamic range for each channel.

Based on company recommendation, I upgraded the unit to latest firmware.

SMSL SU-10 Measurements
Let's start with our usual balanced output after letting the device warm up (SMSL recommendation):
View attachment 238219
Distortion is extremely low at -140 dB so SINAD is noise dominate as usual. It is hitting the limits of the analyzer in this front but it still manages to squeak by to get #1 place:
View attachment 238220
Of course no one should worry about fraction of dB SINAD as the measurements are highly variable in that regard.

RCA output performance is a bit lower as usual but still excellent:
View attachment 238221

Dynamic range is superb as we would expect:
View attachment 238222

At max volume, balanced dynamic range got close to 132 dB.

IMD distortion and noise is very low although there is more than a hint of "ESS IMD Hump:"
View attachment 238223

Linearity is perfect:
View attachment 238224

Multitone test shows the very low distortion:
View attachment 238225

Jitter performance is uniform and excellent on all three inputs:
View attachment 238226

The usual set of filters are provided:

View attachment 238227

As you can see, there is also an "off" mode available which oddly was the default.

Frequency response suggest Fast Linear as one of the better filters to use:
View attachment 238228

Using that we get excellent broadband THD+N vs frequency:
View attachment 238229

Conclusions
The SU-10 is an excellently engineered DAC. It aims to squeeze the last bit of performance available. Such improvement is limited by how well I can measure it due to analyzer noise itself. We clearly have state of the art performance here. Whether that justifies the higher cost is something that is your business and not mine.

Besides performance we have nice new functionality in the form of two independent USB inputs and support for independent audio outputs. One or the other may be critical for some of you.

I am happy to recommend the SMSL SU-10 DAC.

----------
As always, questions, comments, recommendations, etc. are welcome.

Any donations are much appreciated using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/

I suspect this level of performance will be stuffed into a $150 dongle in 2 years, oh wait forgot about inflation, $190 dongle.

 

[DanielT]

I suspect this level of performance will be stuffed into a $150 dongle in 2 years, oh wait forgot about inflation, $190 dongle.

Dongle? Take that step further, standard in all computers, standard in all tablets, mobiles and so on. The level that SMSL SU-10 is at, or close to. In any case, guaranteed to be inaudible. No external DACs are therefore needed. It will probably happen a little after two years from now but it will happen sooner or later, probably sooner.
(within five years is my guess)

Just a guess from my side. But those of you who have been around for a while you know the explosive development of computers in recent decades, I'm just saying.

Edit:
But, for example, balanced analog audio outputs on a mobile phone may not happen in and of themselves in the near future. Thus will external DACs, in which case such balanced outputs need to exist ..I can guess.

 

[mctron]

The same company also offers some cheap models as low as $92~$200. I wonder if they can be as good in most aspects. Particularly in terms of distortion and noise.

Might want to check out the other DAC reviews on this fine website

 

[beagleman]

Given that this product is well into the "inaudible" range, I cannot imagine the usefulness of further delineation using the above technique. Granted, it would be an "intellectual" exercise - which has a range of values depending on ones viewpoint.
If I may make a couple of suggestions. Show an "inaudible" line on your ratings chart - where all products to the left have measurements better than we can hear.
Secondly, introduce a "reliability" metric - if possible. Reliability data would be extremely valuable. Where to get it is a very good question. The OEM knows this based on warranty claims and repairs. I can see an OEM sharing this data if it reflected well on the product. Relying on end user experiences can be good or bad - need a large sample set for any reasonable accuracy.

Maybe 2 lines
1. Inaudible
2. Inaudible in real world circumstances.

In my real world circumstances, I have found mediocre SINAD figures not truly audible in many situations, but they COULD be humanly audible in the right circumstance.
Of course, there is quite a gap between what is humanly possible to hear, and what we would tend to actually hear in real world situations.

 

[amirm]

Secondly, introduce a "reliability" metric - if possible.

It is not possible. Yes, I can do teardowns of amps and such but beyond that, this type of evaluation simply is not in the cards. To get a sense of reliability multiple units need to be tested in various conditions for long periods of time. That just not going to happen and is outside of my charter.

 

[Eulipian]

Does it add to cost or detract from performance substantially to add a headphone out circuit? Or is it just a ploy to force one to buy another box?

 

[Spocko]

It is not possible. Yes, I can do teardowns of amps and such but beyond that, this type of evaluation simply is not in the cards. To get a sense of reliability multiple units need to be tested in various conditions for long periods of time. That just not going to happen and is outside of my charter.

i love this response! "...is outside of my charter." I need to use it more often.

People: hey, can you review this cheapo TV from Walmart?
Me: sorry, it's outside my charter
People: wait, what's in your charter?
Me: sorry, that's above my pay grade

 

[Talisman]

I say something that may be a little annoying here. But for me, with so many absolutely transparent dacs already at 100 euros or less, spending more than 500 euros for this feature set (let's go 900) makes no sense, but I repeat FOR ME.
Many times I have had the trivial need to connect two spdif or two coaxial inputs, it is nonsense but only the dx3pro + (which in fact I bought) had two coaxial inputs.
The research of the last sinad decibel paying 6X for a product that has the same functions and is equally transparent does not appeal to me.
The review is fine, the panther is fine, praising engineering is fine, but does this direction of technological development make sense?
If at the same price they had given me a dac with the performance of the Topping E30 but with a pair of inputs of all kinds, with a performing equalization software, with one or two analog inputs to be able to use it at 360 degrees as a preamp.
But right now anyone who already has pretty much any balanced dac 110 sinad and up would upgrade just for the personal pleasure of doing so, replace their unit with this gorgeous, flamboyant new, and their music would be EXACTLY the same to their ears.

 

[pseudoid]

Thank you @amirm,
Is there much audible differences between Fast1 and Fast3 filters?
The 23.6kHz hump should be inaudible and their website graphs do not depict any waveform differences.

SOTA has a new champ!

 

[DanielT]

Does it add to cost or detract from performance substantially to add a headphone out circuit? Or is it just a ploy to force one to buy another box?

Now I think you are unfair to SMSL SU-10. Consider those who are to buy a SOTA DAC.

We should be happy that there are different options and that there is competition in the HiFi market.

Even those of us who are NOT looking, in need of, nor are interested in SOTA benefit. Sooner or later it trickles down to the "simpler" models and we will benefit from better performance. That is the consequence (the indirect one?) of the development of the top models, the SOTA models.

The level of what is (was) ok, good for price X is indirectly raised thanks to these SOTA DACs.