• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

SMSL SU-8 Version 2 Balanced DAC

Joined
Mar 4, 2019
Messages
10
Likes
1
So, my SU-8 just arrived... Connected it to my set of JBL 305Ps using XLR cables. USB input. MusicBee with WASAPI in exclusive mode.

And my question is: what input sensitivity do I have to use for the JBLs? +4dBu or -10dBv?

I first tried +4dBu (for "professional" equipment), the speakers sound fine, however the max volume seems rather low.

I then tried with -10dBv (for "consumer-grade" equipment) and max volume seems OK, I think I don't hear any distortion.

Any ideas?

By the way, how can I correlate the DAC maximum XLR output (4Vrms) to the speaker max input voltage (not in the manual) and the input sensitivity settings?
 

JohnYang1997

Active Member
Joined
Dec 28, 2018
Messages
226
Likes
62
So, my SU-8 just arrived... Connected it to my set of JBL 305Ps using XLR cables. USB input. MusicBee with WASAPI in exclusive mode.

And my question is: what input sensitivity do I have to use for the JBLs? +4dBu or -10dBv?

I first tried +4dBu (for "professional" equipment), the speakers sound fine, however the max volume seems rather low.

I then tried with -10dBv (for "consumer-grade" equipment) and max volume seems OK, I think I don't hear any distortion.

Any ideas?

By the way, how can I correlate the DAC maximum XLR output (4Vrms) to the speaker max input voltage (not in the manual) and the input sensitivity settings?
It's neither. Would probably use +4dbu. This dac outputs 2Vrms which is much higher than the old standards. I highly doubt the max volume was low? Idk man.
 
Joined
Mar 4, 2019
Messages
10
Likes
1
It's neither. Would probably use +4dbu. This dac outputs 2Vrms which is much higher than the old standards. I highly doubt the max volume was low? Idk man.
The DAC output is actually 2Vrms for RCA and 4Vrms for XLR... But where can I find the maximum speaker input rms voltage for each sensitivity for no distortion?
 
Joined
Nov 29, 2018
Messages
5
Likes
1
Hi all,

Received my SU-8 in December direct from SZA and I’m happy to report it sounds phenomenal paired with my THX 789.

I use Audirvana for Mac, and I’ve wondered for the longest time now why it is that this DAC processes/outputs all of my music in 32-bit over USB, regardless of the native bit depth of a given track. The sample rate is left unchanged. This behavior occurs with “Integer Mode” enabled in Audirvana’s preferences — when disabled, it processes all tracks in 24-bit instead.

When running the SU-8 with Audirvana for Windows (Bootcamp), I observe that over WASAPI everything is processed in its native bit depth, but over ASIO (with driver) it exhibits the same all-32-bit behavior.

I’ve looked it up many times but have never found an explanation for this behavior. My previous DAC, the Modi 2, was my first and always processed tracks in their native bit depth.

This might not be an issue for some, but it’s bothered me greatly because I’m a bit OCD and after extensive listening it seems that tracks processed in 32-bit offer a slightly wider soundstage at the expense of a slight loss in clarity. These are subjective impressions, of course, but I am certain there is a difference — my friend concurs.

I just can’t determine which I prefer and I end up having to choose between two different versions of each track in my library — it’s an OCD nightmare. I’m OK with 16 -> 32 (I literally have no other choice on Mac), but I go back and forth on 24 -> 32 regarding whether or not I should disable Integer Mode for native 24-bit playback.
I could just use WASAPI on Windows and be done with it, but I always use macOS and prefer it.

I suppose my question is: Why is this a thing? Do other 32-bit DACs do this? Is everything just supposed to sound “better” processed in 32-bit?
 
Last edited:

Veri

Major Contributor
Joined
Feb 6, 2018
Messages
1,372
Likes
1,183
This might not be an issue for some, but it’s bothered me greatly because I’m a bit OCD and after extensive listening it seems that tracks processed in 32-bit offer a slightly wider soundstage at the expense of a slight loss in clarity. These are subjective impressions, of course, but I am certain there is a difference — my friend concurs.
Converting 24 bits to 32 bits should be lossless, unless any processing or volume control is done it should be merely padding some zeros.
I would find it extremely unlikely that 32-bit sounds any different.
 
Joined
Dec 25, 2018
Messages
37
Likes
46
Hi all,

Received my SU-8 in December direct from SZA and I’m happy to report it sounds phenomenal paired with my THX 789.

I use Audirvana for Mac, and I’ve wondered for the longest time now why it is that this DAC processes/outputs all of my music in 32-bit over USB, regardless of the native bit depth of a given track. The sample rate is left unchanged. This behavior occurs with “Integer Mode” enabled in Audirvana’s preferences — when disabled, it processes all tracks in 24-bit instead.

When running the SU-8 with Audirvana for Windows (Bootcamp), I observe that over WASAPI everything is processed in its native bit depth, but over ASIO (with driver) it exhibits the same all-32-bit behavior.

I’ve looked it up many times but have never found an explanation for this behavior. My previous DAC, the Modi 2, was my first and always processed tracks in their native bit depth.

This might not be an issue for some, but it’s bothered me greatly because I’m a bit OCD and after extensive listening it seems that tracks processed in 32-bit offer a slightly wider soundstage at the expense of a slight loss in clarity. These are subjective impressions, of course, but I am certain there is a difference — my friend concurs.

I just can’t determine which I prefer and I end up having to choose between two different versions of each track in my library — it’s an OCD nightmare. I’m OK with 16 -> 32 (I literally have no other choice on Mac), but I go back and forth on 24 -> 32 regarding whether or not I should disable Integer Mode for native 24-bit playback.
I could just use WASAPI on Windows and be done with it, but I always use macOS and prefer it.

I suppose my question is: Why is this a thing? Do other 32-bit DACs do this? Is everything just supposed to sound “better” processed in 32-bit?
I have the exact same setup as you. But I always get a native playback in Mac/Audirvana. Check your Audirvana Preferences. You must have enable upsampling.
 
Joined
Nov 29, 2018
Messages
5
Likes
1
I have the exact same setup as you. But I always get a native playback in Mac/Audirvana. Check your Audirvana Preferences. You must have enable upsampling.
Huh, that’s very interesting. Unfortunately no, I don’t have any upsampling enabled whatsoever under the Audio Filters tab. But I do notice that without my SU-8 connected, Audirvana is forcing all my tracks to play in 24-bit — surely that can’t be normal?

I’m running macOS Sierra and I do have the Direct Mode hack installed — perhaps that has something to do with it? I can’t imagine what else could be causing it. I’ve tried toggling all sorts of settings on and off but to no avail, with the exception of Integer Mode which forces 32-bit when enabled with the SU-8 connected.
 

Feyire

New Member
Joined
Mar 9, 2019
Messages
3
Likes
9
I went ahead and purchased the SMSL SU-8 based on the review and measurements recorded here - thanks again for your professional work @amirm and to the ASR community.

Here is how I weighed up the SU-8 when I was deciding:

Pros:
  1. Exceptional price (approx. $200-$249)/performance when it comes to SINAD as well as other key measurements and hardware features
  2. Fully balanced circuit design
  3. RCA/unbalanced and XLR/balanced outputs
  4. Various inputs - USB, Optical and Coaxial
  5. Volume control (including remote!)
  6. Aesthetically pleasing design
Cons:
  1. Linearity could be better for RCA/unbalanced
  2. ESS IMD Hump

The plan is to use it as a DAC for living room speakers and for my headphones (HD650). Inputs will be USB via computer and optical from TV.

I initially wanted to order from Massdrop, but in the end decided against it and went with Shenzhenaudio because they at least have a 1 year warranty and immediate stock availability.
 
Last edited:
Joined
Mar 29, 2018
Messages
19
Likes
12
Location
Bilbao, Spain
Hi, I received the SU-8 three weeks ago and today I tested it in a recording studio and considering my experience with it in two sound setups before, I can say that I am disappointed with it. It does sound quite inferior to a 2012 M-Audio Pro Fire 2626 audio interface.
Its sound is mid-frequency centered, there is a bump around 2.5-4 Khz, the top end is rolled-off and bass lacks intensity, and has a dark and narrow soundtage or channel separation. I couldn't stand it for very long compared to the M-Audio. I tested my own library of quality compressed music, as well as .wavs from recorded studio projects and frequency tone tests. We used balanced output and usb connection with included cable to a DAW PC computer. I must say I clearly discern its audio qualities. We used foobar200 with ASIO drivers n Windows 10 with default settings for SU-8 and the ProFire.

When I tried it first at home with my tv set I was not surprised either, I used Tv optical out to the SU-8 and to my Onkyo A-9010 integrated amplifier and I noticed bass was lacking compared to the TV-analog out and also the Onkio's optical-digital in. I noticed its mid frequency character but I couldn't discern its real quality because of my system limitations. I use two Q-Acoustics 3010 for tv and blu-ray player. I wasn't surprised to keep it. And I have to say the TV rca outs are of good quality since they sound very similar to the Onkyo's Wolfson WM8718 DAC.

At the main workplace in an auditorium with a DLCP Projector and out of coaxial from a Sony BDP-S790 blu-ray player to an analog mixer line-in (unbalanced) and then to powered 400W stage speakers, there is an improvement from the blu-ray player in clearer mid frequencies and dynamics in movies and music but the top-highs are also rolled off, which doesn't bother much for movies. Bass is also less full than blu-ray rca outs although stage monitors are bass sentitive and it results in better controlled bass sound profile, more 'isolated'.

However I must say that in the recording studio they have disappointed me. It sounded quite inferior to the circa 2012 ProFire 2626 audio interface (around 700$ then), which has 192KHz digital converters and balanced outs. The recording studio owner has custom-made wooden monitors and quality power amplifier with acustically rectified equalizers (defeated for the SU-8). I have listened several times music there, my own files as well as his music projects, and classical music. He has given me his outlook of the unit, and we played some pure tones and instrument samples that revealed some distortion compared to the M-Audio. I trust the set-up connection from the unit. Sound filters are misleading, there were bass and treble differences but none did sound clearly better. Some filter was even noticed adding strange patterns according to the engineer.

I expect that a present-day anynchronous balanced dac should hold its own to a medium quality hardware mixer, but it isn't the case.
studio.jpg
 
Last edited:

maxxevv

Senior Member
Joined
Apr 12, 2018
Messages
499
Likes
327
Have you compared the M-Studio with anything else directly ?

Because these can be all relative as in the M-Studio might have a V or U-shaped frequency profile and the SU-8 a flat one. When you the listen to two such devices next to each other, the flat one will sound like it has a mid-hump and the V or U-shaped one will sound like it has recessed mids depending on which way round you interpret it. Its a very common phenomena with earphones and headphones when people listen to them.

I have no idea how the M-Studio sounds like but just suggesting a possibility here based on what I have experienced with other audio products.
 
Last edited:
Joined
Mar 29, 2018
Messages
19
Likes
12
Location
Bilbao, Spain
Have you compared the M-Studio with anything else directly ?

Because these can be all relative as in the M-Studio might have a V or U-shaped frequency profile and the SU-8 a flat one. When you the listen to two such devices next to each other, the flat one will sound like it has a mid-hump and the V or U-shaped one will sound like it has recessed mids depending on which way round you interpret it. Its a very common phenomena with earphones and headphones when people listen to them.

I have no idea how the M-Studio sounds like but just suggesting a possibility here based on what I have experienced with other audio products.
No, but I have compared the SMSL to four different dacs including the M-Audio. I understand what you mean about V and U shape. The M-Audio is equalised for flat response in the studio using a professional microphone, white-pink noise generator and a hardware spectrum analyser. I heard the M-audio unequalised and equalised, and I get an idea of the improvement. However, the SMSL didn't sound well from the start although it might improve by EQ. The tonal balance is not right. I sensed it at home too. Mid frequencies have detail and don't sound digital but just not cohesive and resolved enough as a whole.
Channel separation was not good and timbre was dark, saturated in mids. A good piece of equipment just sounds well from the start, it doesn't need 10 different filters and just sounds true. Also low frequency tones and isolated instrument samples like snare drums made a difference. My friend has good sounding gear and years of experience. I have never heard audio as well as there, custom engineered monitors and a vintage quality power amp. You get to notice compressed tracks easily. I have musical training, go to theatres and classical music is a good test for flat response and tonal representation. The M-audio sounded right.

We tried all sound profiles and filters, and although there are bass and treble balances, the inherent timbre and resolution was unchanged. Some filters didn't sound right. Tube profiles added artificiality.
 
Last edited:

Similar threads

Top Bottom