SMSL SU-9 Balanced DAC Review

[MediumRare]

It's not a streamer. They'd have to add a whole additional layer of hardware and software. The prices would probably be 400 more, at least. And then they'd have to support it.

Sorry, that’s just not factual. A $300 Yamaha amp can do all that with 100 wpc and four outputs.

 

[VinceS]

I see selectable inputs but i am really interested if it has a selectable outputs so you can switch between monitors and a headphone amp without having to pull any cables.

 

[Jimbob54]

Amazingly, a $300 amp can do all that at 100 wpc (Yamaha R-N303) but not a $440 stand-alone DAC!

True . And yet people are still buying separates including expensive standalone DACs. I think there is a consumer mindset (here included ) that cant move away from separates being better .

 

[MediumRare]

True . And yet people are still buying separates including expensive standalone DACs. I think there is a consumer mindset (here included ) that cant move away from separates being better .

With AVRs testing mediocrely I can’t convince myself to go that integrated route. Also I’d rather choose my own multi zone player/library app.

 

[Jimbob54]

Amazingly, a $300 amp can do all that at 100 wpc (Yamaha R-N303) but not a $440 stand-alone DAC!

In fairness - for $100 or so a large (relatively) number of years ago (pre chromecast audio)- you could get an Apple Airport Express that did the streaming and DAC thing if you could accept Apple's infrastructure. Im sure there were others

 

[somebodyelse]

Why no one licensing this tech from Google to use in their devices?

They may not like Google's terms, something we're unlikely to see. That's assuming you can license it at all given Google have discontinued it.

Or an LG Rebel tracFone. Qualcomm Snapdragon. Android 8. 2.4g wifi. Touch screen.. BT. replaceable battery. microSD, USB OTG capability..$29

And no security updates (not listed on the list of phones receiving updates, which does list security issues affecting their phones with that Android version)

 

[Vasr]

THD+N vs Frequency is a little messy for 9038pro

Sibling rivalry?

 

[Vasr]

With AVRs testing mediocrely I can’t convince myself to go that integrated route. Also I’d rather choose my own multi zone player/library app.

With AVRs in congested chassis - perhaps. But these are stereo DACs at $400. I would fully expect this type of performance to creep into stereo pre-amps and integrated amps as manufacturers move slowly into digital. Or the DAC makers here move up the value chain into pre-amps and/or integrated amps. At that time, $400 for just the DAC portion looks silly relative to the functionality.

On the other hand, manufacturers may not want to do any more effort beyond what is clearly not an audible concern and so these would remain an enthusiast early/adopter product for people who wish to pay a premium for paper specs and bragging rights.

This unit is really overpriced because of the early adopter premium for latest chip. Like latest Nvidia chip cards being sold at a premium or the latest memory chips bumping up prices for new SSDs or memory cards. At least the latter examples have tangible benefits not just lab measurements.

My calculation for current fair value of stand-alone DACs based on current prices
Base balanced USB DAC with optical inputs ($200) + Usable Display ($30) + Volume Control ($50) + BT/DSD/MQA ($40) + Streamer ($40)

The rest is new gizmo premium.

 

[voodooless]

It's not just the number of filters. Look at the Pro-Ject filters I referred to above. They have the same shape as the SMSL ones, and are different only in the ultimate attenuation (and one extra option). The Pro-Ject also decodes MQA BTW.

@amirm are the ADC settings for capturing the stopband always the same? Always using the same device? It almost looks like the stopband is drowned by HF noise in various ways. The Pro-Ject data is actually interesting in that the HF noise seems to start later than on the SMSL. On the DAC8 stereo for instance, the noise level is much lower. Also there the stopband should actually go even lower than captured for some of the filters)

If it’s not the ADC or sampling settings, does that mean that the noise is actually coming from the devices? If so a measurement with zero signal should show the same level of noise? But then how can the noise on the multi tone be below -130 dB, even in the HF?

 

[PeteL]

With AVRs in congested chassis - perhaps. But these are stereo DACs at $400. I would fully expect this type of performance to creep into stereo pre-amps and integrated amps as manufacturers move slowly into digital. Or the DAC makers here move up the value chain into pre-amps and/or integrated amps. At that time, $400 for just the DAC portion looks silly relative to the functionality.

On the other hand, manufacturers may not want to do any more effort beyond what is clearly not an audible concern and so these would remain an enthusiast early/adopter product for people who wish to pay a premium for paper specs and bragging rights.

This unit is really overpriced because of the early adopter premium for latest chip. Like latest Nvidia chip cards being sold at a premium or the latest memory chips bumping up prices for new SSDs or memory cards. At least the latter examples have tangible benefits not just lab measurements.

My calculation for current fair value of stand-alone DACs based on current prices
Base balanced USB DAC with optical inputs ($200) + Usable Display ($30) + Volume Control ($50) + BT/DSD/MQA ($40) + Streamer ($40)

The rest is new gizmo premium.

ESS introduced the 9038Pro in 2017

 

[manoy385]

This looks like it would go really well with the SMSL DA-8s. I wish someone would send Amir the DA-8s for measurement. I really want to know if it measures bad like SMSL's other power amps. Reviews seem to say otherwise.

 

[amirm]

@amirm are the ADC settings for capturing the stopband always the same? Always using the same device?

Of course. I simply change the filter setting and run the same test in "append mode" to add the next line and so on.

I use a small FFT but average 32 times as otherwise the lines will be very thick with noise. This results in a different noise floor in other tests such as Multitone where I use very long FFTs.

I think the secondary impact here is the noise of the DAC itself. That gets added to the attenuation and sets a lower limit. We see this in AVRs for example with their relatively high noise floor.

There may also be noise floor modulation.

 

[Harmonie]

Prices trend is so much down ...
In a year or two, they'll give you a Dac FREE for every amp you buy ....

 

[TNT]

32 tone IMD for the different filters would be interesting...

//

 

[amirm]

32 tone IMD for the different filters would be interesting...

//

That test runs at 192 kHz sampling so these filter aspects won't impact it.

 

[bennetng]

I like that there is a proper brickwall filter. I used that for the following test:

If the "brickwall" filter is identical to the one mentioned in datasheet, then it has 20dB poorer stopband attenuation than "linear phase fast".

 

[Jerry Bruckheimer Fan]

What is the voltage at the max volume 99?

 

[voodooless]

Of course. I simply change the filter setting and run the same test in "append mode" to add the next line and so on.

I use a small FFT but average 32 times as otherwise the lines will be very thick with noise. This results in a different noise floor in other tests such as Multitone where I use very long FFTs.

I think the secondary impact here is the noise of the DAC itself. That gets added to the attenuation and sets a lower limit. We see this in AVRs for example with their relatively high noise floor.

There may also be noise floor modulation.

Thanks, that explains quite a bit. It also gives a good indication of the extra engineering invested to keep HF noise low.

 

[Xyrium]

What is the voltage at the max volume 99?

5.2V? Estimating based on what he posted in the opening dashboard image.

Wow, what excellent performance, even with the minor ESS hump. Pretty soon that blue section will have to be limited to SINAD of >120dB!!!

 

[Nango]

Looks like a "Volks-ADI-2" once it arrives the 200$ price tag.