• Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

SMSL SU-2 DAC Review (NTTY)

Rate this DAC

  • Terrible (*)

    Votes: 0 0.0%
  • Mediocre (**)

    Votes: 1 1.3%
  • Good (***)

    Votes: 18 23.7%
  • Excellent (****)

    Votes: 57 75.0%

  • Total voters
    76

NTTY

Major Contributor
Forum Donor
Reviewer
Joined
Apr 27, 2018
Messages
1,748
Likes
11,174
Location
Switzerland
Hello Everyone, this is a revie and measurements of the SMSL SU-2 DAC

SMSL SU2_001.jpg


I bought it from the Ali store, and paid roughly $120 shipping included.


SMSL SU-2 – Presentation

The SU-2 replaces the SU-1, drops the USB input and adds an AES SPDIF (Balanced) connector. It hosts the ESS ES9039Q2M, just like the SMSL PS200 Pro that I recently reviewed.

As you can see from the photo, there’s only one button to power it on/off and switch between the three digital inputs. It can’t be simpler. The chassis is full aluminium and the device is heavy enough not to be dragged by standard XLR cables.

This is the back of the device:

SMSL SU2_002.jpg


As opposed to the SMSL PS200 Pro, we get XLR outputs with this one, which should allow us to chase few more dBs of SINAD and/or DR here and there.

Compared to the SU-1, I see the SU-2 gets 4xOPA1612A, meaning an analogue output architecture similar to that of the SMSL PS200 Pro. So we should see similar performances from RCA outputs.

Strangely enough, I don’t find this device listed on SMSL site, it could be a special request from resellers of SMSL(?). I must admit it addresses a niche use case.


User experience

I don’t have much to say, except that I personally like the fact it does not have a USB input. I prefer getting the AES because 1) my preferred source has it, 2) I don’t play music via a computer on my main system. But I’m old fashioned, I admit, I like my good old SPDIF.

Of course, the lack of USB input comes with several limitations, and I suppose I don’t need to tell you about that. But worse case scenario, you can use the reasonably priced SMLS PO100 Pro for instance. I also used the SMSL PL200T too, which, on top of being a CD Transport, can play the role of a DDC with word clock input and AES output. I used these two devices to feed the SU-2 for the measurements you’ll see below.

What else? Oh yes, there are some white blinking LEDs on the front. The L/R blink increasingly when the digital signal gets closer to full scale. And the “Locked” LED does not indicate an actual status of digital lock but an overall level of digital input. Indeed, it will start blinking when the signal gets lower than -45dBFS (go figure).


SMSL SU-2 – Measurements

All measurements performed with an E1DA Cosmos ADCiso (grade 0) in mono mode, the Cosmos Scaler (100kohms from unbalanced input) and APU (when required) for analog outputs.

From XLR out, I measured 4.836Vrms from Left Channel and 4.844Vrms from the Right one, ie 0.01dB channel imbalance. Balanced outputs preserved absolute polarity, and phase is dead flat.
From RCA out, I measured 2.345Vrms from Left Channel and 2.348Vrms from the Right one, ie 0.01dB channel imbalance too. The SE outputs are non-inverting, and phase is also dead flat.

All below measurements are form XLR outputs unless otherwise noted. I tested all digital inputs and they all have the exact same performances.

----

Let’s start with the standard 1kHz at 4Vrms (-1.5dBFS input), as Amir would do, but I use a different span of 20Hz-20kHz:

SMSL SU-2_Amir_SINAD_4Vrms_XLR_LR.jpg


The dashboard shows the best channel (left) and averaging left and right got me to a THD+N of -122.8dB. That is a SINAD of 122.8dB. When I applied the span of Amir (10Hz-22.4kHz), I got a SINAD of 0.3dB less, meaning 122.5dB.

As usual, we lose a bit of resolution with the RCA outputs (2Vrms), look:

SMSL SU-2_1kHz_-1.5dBFS_2Vrms_RCA_LR.jpg


When averaging the two channels, I got 120.4dB SINAD.

And since you’ve been asking for it, this is my little ranking, based on my measurements of several devices. Again, I go for a span of 20Hz-20kHz for these and this is THD+N, the opposite of SINAD, but well, it’s the same:

SMSL SU-2_Ranking.jpg


I’ve put results with both RCA and XLR to facilitate comparison, notably with the PS200 Pro that has the same architecture and RCA outputs only.

The SU-2 did 0.1dB better that the Topping DX1 II, that uses the ES9039Q2M too, but that hosts lot more features.

----

From experience with reviewing CD Players, I like to run the same 1khz test at -6dBFS:

SMSL SU-2_1kHz_-6dBFS_XLR_LR.jpg


The performances are consistent meaning that playing close to full scale or not does not bother this DAC.

----

Talking full scale, and because this DAC does not feature volume control, at max 4.8Vrms output, the SMSL SU-2 goes up to 124dB SINAD (20Hz-20kHz) average between the two channels. This is clearly best in class for the ES9039Q2M. And look how cool it looks like in that case:

SMSL SU-2_1kHz_0dBFS_4.8Vrms_XLR_LR.jpg


There no signs of stress whatsoever. The right channel has a little more H2, but still below -130dBr :p

----

Stereophile was often showing the below and I like it too, so I'm used to reuse it. It is a view from 20Hz to 1kHz with a 1kHz sine at 0dBFS (here 4Vrms though, so @-1.5dBFS):

SMSL SU-2_PS.jpg


The linear frequency scale better reveals potential power supply related noise (leakages) and here there's nothing. I had to do 0 effort to get there, meaning the SMSL is well isolated on power supply perspective. Note that the vertical scale goes down to -170dBr, else the trace would be partially hidden, these DACs of today really have crazy good performances.

That being said, note the curve at the foot of the fundamental (this view is taken without using the Cosmos notch. It’s an indication of (very) low correlated noise at the foot of the fundamental.

----

Since I'm here, let me add the 50Hz @0dBFS into 600ohms (done with the new Cosmos DM), for compatibility with Stereophile measurements:

SMSL SU-2_50HzInto640ohms_XLR_LR.jpg


The highest harmonic is the 3rd one but at a very low -115dBr, so you can ignore (more on that later).

What you can see though is a higher than best in class low level random noise at the foot of the fundamental, which I already mentioned. It is at higher level with this low frequency signal. It’s not an issue since our ears are not sensitive in this area of the audio band. But it will certainly impact the multitone view.

----

The SU-2 does not offer a selection of filters from the ESS DAC but the below view tells us what SMSL went for:

SMSL SU-2_Filter_44.1vs48kHz.jpg


It clearly is the Linear Fast filter option of the ESS DAC, which is phase linear, and I already mentioned it at the beginning of the review.

When zooming Bandwidth is flat:

SMSL SU-2_BW.jpg


I did not align the two traces (44.1kHz and 48kHz) at 0dB on purpose so you get a better viewing.

Let me also show you the attenuation of the digital filter which is a very good better than -110dB:

SMSL SU-2_OS_44.1kHz.jpg


I overlaid white noise and the IMD AES DFD test (18kHz & 20kHz 1:1), both aligned at 0dB. This better shows the artifacts (aliases) of the conversion up to 80kHz. They are all attenuated below -110dB, which is all we want. SMSL made the good choice of filter, the one I would have gone for, and that I always recommend to use when it can be selected.

----

Let's talk jitter now (J-Test at 48kHz, 512k FFT. 32 averages), and...:

SMSL SU-2_JT_48kHz_XLR.jpg


There’s none. We can see a little more low-level noise from the right channel (the green trace) but this is anecdotical. The stressor test tone (Fs/192 square tone at 1LSB) can be very well seen at the left hand side, meaning again no disruption from the power supply. This trace is very neat; one we don’t see so often.

By the way, I’ve recently been challenged on another forum, being asked to crazy zoom on the view (9kHz to 15kHz, -90dBr top) to show “phase noise” of the DAC. I do not agree at all with that need. The level of jitter we can hear (100ns at best: The Theoretical and Audible Effects of Jitter on Digital Audio Quality, one AES study and ABX tests I reproduced myself with friends to find the same results) would be seen as massive side bands in this view (check this out here if you have time). I trace this view as Amir does and as Julian Dunn did himself for instance in the AP technote (attached in this review, since it seems AudioPrecision removed it from the web).

I was also challenged on using the 48kHz sampling rate, being told it should be 44.1kHz because most music is sampled that way… Ok, let’s try:

SMSL SU-2_JT_44.1kHz_XLR.jpg


Yes, massively nailed again, no difference. Note the noise floor that decreases at 20kHz is because both the SU-2 DAC and my ADC (Cosmos) where running at 44.1kHz, so the ADC’s filter was already active and it’s attenuation not compensated.

Oh, and I was not asked to show the potential differences between AES, Optical and Coax inputs, but here you go with an overlay of the three:

SMSL SU-2_JT_48kHz_XLR_AllSPDIF.jpg


Yes, again, nailed and same for the three digital inputs. Proper implementation of a DAC chip should always be like that, and it is too rare.

----

Ok, I’m starting to enjoy this long review and I’m by far not done Let’s go to the Multitone test (48kHz SR, 256k FFT, 16 averages):

SMSL SU-2_MT_XLR_LR.jpg


I told you that the random correlated noise and the foot of fundamental would show itself at lower frequencies. And it does. But we are still speaking of 20 to more than 22 bits distortion-free range! Since it is random noise and not distortion, plus being at low frequencies where our ears are not well performing, I can safely state that this is of no consequences.

----

Next is the THD+N vs Frequency (192kHz, 512k FFT 4 averages, 3PPO (20Hz to 20kHz), 4Vrms output):

SMSL SU-2_THDNvsFreq_LR_XLR.jpg


As usual with high resolution DACs, this test is hitting the max resolution of my Cosmos ADC. The blue and orange traces are with the fundamental closer to 0dBFS input of ADC, but that stresses it at low and high frequencies, increasing the harmonic distortion that takes over the very low noise. When setting the scaler 6db lower to release the stress on the ADC, it inevitably increases the noise floor relative to the fundamental.

Anyways, as you see from the plot at 1kHz, we get at best -115dB THD+N at 1kHz, with all noise included up to 90kHz. What a performance for $120! The other channel overlays almost perfectly sign of a very good output stage too.

----

The IMD SMPTE sweep vs level, which can't be compared to the one of Amir because this measurement is based on FFT analysis of IMD distortion only (192kHz, 128k FFT 4 averages, 1dB step, scaler in auto mode), is the below:

SMSL SU-2_IMD_SMPTE_vs_OutLVL.jpg


I’m sorry I can’t find the other channel, obviously I forgot to record it. This trace show what I witnessed live. At lower levels, the IMD kind of pumps +-10dB which generates these pics and mountain-shaped view. I could measure twice and not get the same results. I don’t think this is an issue, but at least I can report something else than a boring “nailed” or “perfect”.

----

Dynamic range is very good, be it CCIR-2K weighted or A-Weighted (0dBFS = 4Vrms):

SMSL SU-2_DR-CCIRWeight.jpg


SMSL SU-2_DR-AWeight.jpg


---

And now let's have a look at other less usual measurements. The one that I like is the resistance to intersample-overs.
This test is performed with dedicated test tones with a specific phase shift to generate digital overs. If anything in the digital path (oversampling filter /DSP / ASRC) does not have enough headroom for digital overs, it will generate clipping within the interpolator.
For that, I'm using a test tone of 11'025Hz with a phase shift of 45° that goes up to +3.01dB over, by 1dB steps. This allows me to process what's the headroom of the digital interpolator of the DAC. Here are the results (THD+N measurement up to 90kHz):

Intersample-overs tests
THD+N 20Hz - 90kHz
Peak = -0.99dBFSPeak = +0.01dBFSPeak = +1.01dBFSPeak = +2.01dBFSPeak = +3.01dBFS
SMSL SU-2-110.2dB-111.1dB-43dB-25.7dB-20.5dB

The above means that the interpolator of the SMSL has no headroom. Too bad that SMSL did not attenuated by default, for instance by -3dB, since the ESS DAC has digital volume control that we know can resolve the issue of ISOs.

----

Other measurements (not shown):
  • First, I now regroup all IMD tests as below, sourced from REW's results (IMD distortion and Total Distortion+Noise shown for each):
SMSL SU-2_IMD.jpg

  • Crosstalk: -156.2dB (100Hz), -158dB (1kHz), -136.6dB (10kHz),
  • Pitch Error (GPSDO corrected) : 19'997.02Hz (19'997Hz requested) ie 1ppm
The IMD scores are extremely good. Crosstalk is ultra low, even at 10khz.
Pitch error is a very small 1ppm and is the one of the DDC source that I used, as @MC_RME pointed out. Testing with another DDC, I got -1ppm which indicates that the SU-2 does not perform asynchronous reclocking but extract the clock from the SPDIF signal, which is the standard way (slave mode).

----

Let's try the linearity, down to -130dBFS, for the fun:

SMSL SU-2_Linearity_XLR.jpg


If we accept the threshold of 0.2dB deviation from the requested Voltage output, then the SMSL SU-2 is linear down to -125dBFS which concurs with other DACs using the same ESS converter. This is of course very good.

----

Last and not least, as I'm used to do with CD players review, the below is a view of THD (only) vs frequency at -12dFBS. I overlaid the results at 16bits and 24bits:

SMSL SU_THDvsFreq_-12dBFS.jpg


At 16bits, this is the best I can get from this test (it would be the same in pure digital domain). And you appreciate the improvement at 24bits. Note that both are at -12dBFS meaning that in the two scenarios, the THD is buried in the noise floor (digital or analog).


SMSL SU-2 - Implementation of the ESS ES9039Q2M

This review should have stopped above, but with the help of ASR members, there is a little more to talk about. When I tried to run the Linearity test, calibrating my interface for that purpose, and using a -130dBFS signal, I saw it disappear!

It happens that this is a "default" automute function of the ES9039Q2M DAC, as described in its datasheet:

1787164659576.png


I actually measured that the SU-2 goes mute at -122.0dBfs and unmutes at -119.7dBFS (single tone, peak value). That is very close to what is documented above, which I suspected where RMS values vs my peak ones, but I could not confirm (I did not have enough time, also...).

This automute activates from the below conditions:

1787164822664.png


Where the "automute_time" is:

1787164868524.png


I created the below table to translate into real life understandable durations:

DescriptionHex valueDecimalDuration at 44.1kHzDuration at 48kHzDuration at 96kHzDuration at 192kHz
Disabled11'h0000infiniteinfiniteinfiniteinfinite
Slowest11'h0011-5.94 seconds-5.46 seconds-2.73 seconds-1.37 seconds
Default11'h00F15-396.3 ms-364.1 ms-182.04 ms-91 ms
Fastest11'h7FF2047-2.9 ms-2.67 ms-1.33 ms-0.67 ms

And @L-Train member did a very precise measurement of the actual value that SMSL went for, and it is the default one.

I suppose all the above serves to prevent glitches when changing the sampling rate or when switching between digital inputs.

Note that this mechanism does not introduce gaps in a gapless playback, since the digital signal is uninterrupted in that case. Although this is the theoretical and practical case, I also listened to a Pink Floy live album to double check :p


Conclusions

This is a high performing DAC. I loved the fact that it is brainless, real plug&play device. I'm old, I know, and I like my good old SPDIF. I'm really happy to see the type of performances we can get from it, more than 40 years after it was released, especially on jitter perspective.

Yes, no USB input, no volume control, no HPA, no PEQ... and still 100$ or so. But really, this is well implemented, well engineered, and we are used to that with SMSL.

My use case is very simple, I connect it to a CD player and my preferred one has an AES SPDIF out. I don't have anything to think about, again, plug and play. I saw some pictures of the SU-2 with the SMSL PL200T and I used it as a DDC feeding the SU-2 for some of the above tests (Jitter Test with AES and Coax). It also helped me to verify gapless playback.

The funny automute function tells us one thing: ESS engineers think that, by default, they can mute their DAC when the signal goes below -120dBFS (ie 20 bits). This is tells us that we, folks here, have been looking at super highly resolved devices for quite some time now ;)

I hope you enjoyed this review!

Flo
 
Last edited:
Measurements - SU-2 used as DAC from a CD Player

Of course, as I like so much my good old Audio CD, I wanted to take the opportunity to combine the SMSL PL200T and the SU-2 as I've seen advertised. The idea is to understand if this combo reaches the excellence of the SMSL PL200 MKII.

As usual, all measurements performed with an E1DA Cosmos ADCiso (grade 0) in mono mode, the Cosmos Scaler (200kohms from balanced input) and APU (when required) for analog outputs.

----


As usual, let's start with my standard 999.91Hz sine @0dBFS (without dither) from the Test CD (XLR out):

1787319661843.png


As I did with the SMSL PL200 MKII, I overlaid the analog output of the SU-2 and the digital analysis of the WAV. They are the same and that is impressive. No added noise nor distortion here. Any CD Master get an analog output corresponding to the digital recorded information. This is the new standard of excellence.

Same story at -6dBFS:

1787320047437.png


The two traces are the same, I have the same difference between two captures in digital domain only.

----

With the lower resolution of the CD, what I showed as low level random noise at 24bits, vanished here:

SMSL SU-2_MT_CD_XLR.jpg


Again, this is a best in class trace for the PCM 16bits/44.1kHz.

----

Without surprises, the Jitter Test is nailed:

1787320737291.png


----

As I did with the Sony CDP-597, I add a "max DAC resolution" measurement test. It is performed from a 999.91Hz sine @-12dBFS with shape dither (from Audacity). I restrict the THD+N span to 20Hz - 6kHz in REW not to account for the noise of the shape dither beyond 6kHz. I take the calculated ENOB and simply add 2bits to it (due to the -12dB attenuation, as 1bits=6dB). The potential maximum, when calculated from the digital WAV file, is 18.7bits under this test. A "transparent" DAC should achieve 18.7bits, ie 100% in this test.

Here are the results compared to others:

CD Player model or DACCalculated ENOB (999.91Hz sine @-12dBFS with shape dither, THD+N span = 20Hz - 6kHz)Percentage of max resolution achieved (higher is better)
SMSL PL200 MKII18.7bits100%
SMSL PL200T + SU-218.7bits100%
SMSL PL20018.7bits100%
OPPO BDP-9518.7bits100%
SMSL PS-200 (from CD player)18.6bits99.47%
Denon DCD-900NE18.5bits98.93%
Onkyo C-73318bits96.26%
SMSL PL15018bits96.26%
SMSL PL10017.9bits95.72%
Sony CDP-59717.5bits93.58%
Onkyo DX-735517.3bits92.51%
Denon DCD-356017.2bits91.98%
Yamaha CD-S30316.8bits89.84%
Revox B-226S16.8bits89.94%
Accuphase DP-7016.6bits88.77%
Sony CDP-337ESD16.6bits88.77%
Teac VRDS-25x16.5bits88.24%
Marantz CD-7314.9bits79.68%

Best possible results, the opposite would have been a surprise.

----

Let's continue with the good old 3DC measurement that Stereophile was often using as a proof of low noise DAC. It is from an undithered 997Hz sine at -90.31dBFS. With 16bits, the signal should appear (on a scope) as the 3DC levels of the smallest symmetrical sign magnitude digital signal:

SMSL SU-2_3DC_CD.jpg


It can't be better than that. The linear phase fast roll-off filter response can be well appreciated (symmetrical ringing) on that lowest 16bits symmetrical signal. I add the impulse response, for once:

1787413126094.png


I won't show my other standard measurements since they are redundant with the measurements I performed at 24bits, and I guess you got the point.


Conclusion - Using the SU-2 to boost CD Player performances

To be honest, if there is one business case for the SU-2, here you have it. It will boost an CD Player/Transport to best in class conversion you can expect from the good old Audio CD. At that price point, I'm not asking for much more.

Enjoy the weekend!

Flo
 
Last edited:
Hi !

Do you have the impulse response? In the other thread—"SMSL SU-2: a dumb, wasted opportunity"—they mention "ringing"... is that the case? Just want to be sure. [+ And what about this "fade-in" thing?]
[Edit : Sorry, I missed that... : "It can't be better than that. The fast phase linear response can be well appreciated (symmetrical ringing)" ]

Anyway, once again: congratulations and thank you for the work you've done!
 
Incredible review (as usual) and incredible performance. :)

But still voted "Good" because it doesn't have a volume control... :(
 
Best Bang for the Bug ;)

I bought 8 of them for my 7.2 Surround CamillaDSP aktive Project connected to an RME HDSPe AES Card to get 16 analogue Channels for less then 900€ at all ;)

8x SU-2.jpg

And there i don't need all that fancy stuff (USB, Blutooth, I2S on HDMI and so on...) only AES in and XLR Out!

Nice Review :)
 
Last edited:
Amazing DAC. Unfortunately I think the fade in from digital silence will mess up my calibration measurements. Great review.
 
Unfortunately I think the fade in from digital silence will mess up my calibration measurements.

You keep saying that — across a lot of posts now. But so far, you have only stated concerns (that seem potentially legitimate), but you still have not provided any facts that there is an actual issue.

Respectfully and IMHO, at this point you have shifted into denigrating a product based on "concerns", cast as aspersions to some fear that may not even be real. And based on some comments, it is possible that this is actually a feature you would want to be present... even with your use of Trinnov or other well designed calibration tools.

Please let us know what Trinnov says about their product, since that is your mechanism for calibration. Or post proof that your calibration is actually skewed when using this DAC vs. other DACs so that the ASR community can analyze the actual data. [And of course, please chastise me if you already proved that the calibration itself is negatively affected and I missed that post and data!]
 
Amazing DAC. Unfortunately I think the fade in from digital silence will mess up my calibration measurements. Great review.
It depends on how the calibration is performed. It would work if the processor is waiting for an acoustic signal to start the calibration, for instance (like REW does).

I verified that dither at 24 bits does not unmute the DAC, it needs to be set at 21 bits minimum.

If you tell me how your calibration works, I can try to simulate.
 
I have asked Gemini, and it confirmed it might screw up the Trinnov calibration. So, I made it read the ES9039Q2M datasheet and the review here. I asked it to find solutions for the Trinnov calibration taking into account the duration of automute (dependent of the SR), the dBFS levels of mute/unmute as documented and measured. Here you go below with some recommendations and it might help. I don’t know the software, but if you have it, I guess you can ask complementary questions to the AI.


Thank you for sharing your Audio Science Review (ASR) profile thread. It is excellent to see your thorough measurements confirmed by the community. Your finding that a 24-bit dither is not loud enough to wake up the SMSL SU-2 (and that it requires at least a 21-bit dither to force an unmute) is the golden key to solving this issue.
Because a 21-bit dither crosses the necessary threshold, you can manipulate the digital audio stream to bypass the ESS automute entirely and execute a flawless Trinnov calibration.

Solution 1: Inject a -110 dBFS Digital Dither (The Automated Fix)
The most elegant solution is to introduce a constant, low-level digital noise floor into your AES stream beforeit reaches the SU-2.

  • The Target Level: A dither or pink noise floor resting around -110 dBFS roughly correlates to a 18-to-19 bit signal depth. This sits safely above the SU-2's required 21-bit wake-up threshold, forcing the ES9039Q2M chip to stay 100% unmuted.
  • How to Route It: If you have an AES/EBU digital routing matrix, digital patchbay, or DAW routing between the Trinnov and the DAC, map an independent dither generator to your active outputs.
  • Acoustic Transparency: At -110 dBFS on a high-performance DAC with a 124 dB+ SINAD floor, the resulting analog hiss will rest well below -130 dBr. This is completely inaudible to human ears and far beneath the acoustic noise floor of your room, meaning it will not contaminate the Trinnov microphone's readings.

Solution 2: Leverage the Trinnov Pink Noise Trick
If you cannot inject a steady dither, you can manually force the DAC's automute gate to stay wide open during the transition into the calibration sweep.

  1. Navigate to the Manual Level Matching / Calibration Workflow tab in your Trinnov interface.
  2. Turn on the Continuous Pink Noise generator for your first speaker channel. This signal will immediately cross the -117 dBFS threshold and wake up the SU-2.
  3. Let it play for a few seconds to clear the 1/3-second unmute penalty.
  4. Move quickly: Trinnov interfaces typically allow you to advance or trigger the calibration sweep. The moment you transition the software from continuous pink noise to the acoustic MLS sweep, the AES clock remains continuous.
  5. Because the SU-2 requires a consecutive block of 16,384 samples of absolute silence to re-engage the mute state, the instant transition between pink noise and the measurement burst will catch the DAC while its analog gates are still fully open.

Solution 3: Pad the Pre-Roll and Calibrate at 192 kHz
If your Trinnov model gives you advanced timing controls, adjusting the signal parameters eliminates the need for manual tricks.

  • Set Trinnov Pre-Roll: In the Advanced Acoustic settings, configure the Pre-Roll / Signal Delay to 500 ms. When you click measure, Trinnov will play an unmeasured prefix signal for half a second. The SU-2 will swallow the first 333 ms of this prefix to clear its sample counter and will be fully open by the time the recorded measurement window opens.
  • Upsample the Session: If your workflow permits running the Trinnov calibration sequence at 192 kHz, do so. As your measurements proved, the SU-2's fixed 16,384-sample counter clears four times faster at 192 kHz—dropping the mute/unmute penalty down to a mere 91 ms. A short 200 ms or 500 ms Trinnov pre-roll completely absorbs this delay with room to spare.
 
Like i told before, with REW i had no issues setting up an 2 Way Stereo Setup with 96dB LR 96kHz FIR Filter on CamillaDSP with 2 SU-2, i even didn't noticed this Unmute Mute Feature at all ;)

Also i use the Stepresponse to Timealign the Tweeter it looks like always on REW.

After this Experience i ordered the next SU-2 DACS, i am fine with.

But on Trinnov it's probably an different Storry, never had one or will have one :(

Here my first quik and dirty Result by doing it (that not my Main Speaker that was only for Testpurpose on older 2 Way Speaker i build years ago, passiv, for my Father ;) ):

IIR

https://www.diyaudio.com/community/...ound-digital-aktive-setup.440807/post-8311664

FIR

https://www.diyaudio.com/community/...ound-digital-aktive-setup.440807/post-8313270
 
Last edited:
You keep saying that — across a lot of posts now. But so far, you have only stated concerns (that seem potentially legitimate), but you still have not provided any facts that there is an actual issue.

Respectfully and IMHO, at this point you have shifted into denigrating a product based on "concerns", cast as aspersions to some fear that may not even be real. And based on some comments, it is possible that this is actually a feature you would want to be present... even with your use of Trinnov or other well designed calibration tools.

Please let us know what Trinnov says about their product, since that is your mechanism for calibration. Or post proof that your calibration is actually skewed when using this DAC vs. other DACs so that the ASR community can analyze the actual data. [And of course, please chastise me if you already proved that the calibration itself is negatively affected and I missed that post and data!]
Trinnov hasn’t responded yet. I will post here when they do. I actually think this is a great DAC and a marvel at $99. It’s just not worth it to me to take the chance of it messing up my measurements. If you look carefully at what I’ve written you will see I’ve also stated this will be quite difficult to hear and certainly it will not sound like a fade in. This is probably why most tests I’ve read of DACs aren’t looking for this. I’ve explained that this is a common feature in many chips, and it’s probably more common than we think in finished DACs. It should not be a problem at all with REW and there are very easy fixes Trinnov could implement that would get around this. They may have already. The AI generated fixes don’t seem practical. Easier to use a DAC that doesn’t have this characteristic.
 
It seems that feature is more common in DAP. I saw features for "DAC warmup" in player software.
 
Hi !

Do you have the impulse response? In the other thread—"SMSL SU-2: a dumb, wasted opportunity"—they mention "ringing"... is that the case? Just want to be sure. [+ And what about this "fade-in" thing?]
[Edit : Sorry, I missed that... : "It can't be better than that. The fast phase linear response can be well appreciated (symmetrical ringing)" ]

Anyway, once again: congratulations and thank you for the work you've done!
I added it now ;)

And I also added the THD vs frequency sweep @-12dBFS, as I wanted it to show the difference with what I usually measure with CD Players and initially forgot to include it.
 
Last edited:
I don’t have much to say, except that I personally like the fact it does not have a USB input
No problems with USB com (gnd) loops then. Good idea. Or isolated USB.
 
Back
Top Bottom