• 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 D1 - ROHM DAC for everyone

Dear science-based community, could you be so kind as to provide me with a link to a description of what you consider to be the most correct process for setting equal loudness levels for a blind test?..
You take a 1 kHz test tone and measure the voltage with a multimeter in AC mode. The voltage should be as close to the same as possible for both devices.

It's obviously a good idea to set one device to your desired listening volume beforehand and use that as a guide.
 
Thanks Roland68

So, as I understand, a white noise signal is not necessary and a frequency of 1kHz will be enough.
I went a little further and measured the sound pressure level directly in the headphones with an accuracy of 0.1 dB by switching between two DACs.
I hope this doesn't violate the audio science laws either...)
 
Thanks Roland68

So, as I understand, a white noise signal is not necessary and a frequency of 1kHz will be enough.
I went a little further and measured the sound pressure level directly in the headphones with an accuracy of 0.1 dB by switching between two DACs.
I hope this doesn't violate the audio science laws either...)
Chapter 10 of Sound Reproduction would be a good reference point, at this juncture. Sound Reproduction
 
Thanks Roland68

So, as I understand, a white noise signal is not necessary and a frequency of 1kHz will be enough.
I went a little further and measured the sound pressure level directly in the headphones with an accuracy of 0.1 dB by switching between two DACs.
I hope this doesn't violate the audio science laws either...)
Spl measurements are not normally considered accurate enough, due to reflections, ambient noise and general fluffyness of SPL meaters. Might be OK in a headphone though - I don't know. Ideally you need to be measuring in the electrical domain (In this case output of the headphone amp). And the levels need to be within 1% of each other (about 0.1dB)
 
Spl measurements are not normally considered accurate enough, due to reflections, ambient noise and general fluffyness of SPL meaters. Might be OK in a headphone though - I don't know. Ideally you need to be measuring in the electrical domain (In this case output of the headphone amp). And the levels need to be within 1% of each other (about 0.1dB)
I agree, that's why I use headphones for testing.
I'll definitely measure the sine wave voltage at the headphone amplifier's output too.
 
Here's a great, inexpensive power supply option for this DAC. Of course, you should re-solder the connector to USB Type-C or use an adapter.

MEAN WELL, CLASS I, GST18A05-P1J​

I'm using this stupid-cheap LM7805 5V linear (not switching) step down regulator which uses 7V to 24V input.
It runs very clean with this no-name 12V 2A psu that came with my FX-Audio tube preamp (buffer).

There is zero noise with the volume pot on my Douk A5 amp set to 100% and my ear up to the horn loaded titanium compression tweeter (they are over 100db/1w efficient).

With my tube amp (w/cheap chinese tubes) in the mix there's a little noise at 100% but it's barely audible any more than 3 feet away and it goes away completely if the amp volume is turned down to 2-3 o'clock versus 5 o'clock at full volume.
 

Attachments

  • 2-D_NQ_NP_2X_658112-MLA99766474300_122025-F.webp
    2-D_NQ_NP_2X_658112-MLA99766474300_122025-F.webp
    19.4 KB · Views: 55
  • 3-no-name-12V-2!-wall-wart-smps.jpg
    3-no-name-12V-2!-wall-wart-smps.jpg
    158 KB · Views: 56
  • 4-LM7805-12V-2A-SMPS.jpg
    4-LM7805-12V-2A-SMPS.jpg
    129.9 KB · Views: 53
  • LM7805-AliExpress-2026-02-04_17-11.png
    LM7805-AliExpress-2026-02-04_17-11.png
    217.5 KB · Views: 56
Last edited:
The DAC song, to the tune of Do-Rei-Me;

DAC! – a bit, it’s ones and zeros
Re! – the rate your samples run
Me! – “trust the data, not your ears”
Fa! – blind tests are never fun
So! – a clock that’s running smooth
La! – the numbers all agree
Ti! – still someone hears a “veil”
That will bring us back to DAC! DAC! DAC! ... ©


JSmith
 
This thread went off topic and I closed. Recent posts that were not focused on the product have been moved to the Are Measurements everything or nothing thread.

At the OP’s request, have re-opened this thread. If you can’t stay on topic, there are more appropriate threads or start your own. Thanks!
 
I had the opportunity to compare the sound of the SMSL D1 and D200 DACs for quite a long time. I use the HQPlayer Embedded to resample all input streams to DSD512 and the DACs operate in direct DSD mode. There was no significant difference when changing devices in my particular audio configuration.

As a result, I kept the SMSL D1 and bought Sennheiser HD800 headphones, which more unlocked the capabilities of the BD34352EKV chip. Moreover, I don't need MQA support, volume control, Bluetooth Audio, PSM filter selection, etc. Replacing headphones is always more valuable for the sound than replacing a DAC.. :)

The only drawback of this device is not support for the DSD 48x format, which makes it impossible to use integer oversampling filters. To solve this task, manufacturers need to create new firmware for the USB XMOS-316. Hopefully this won't be such a difficult problem and will be implemented in the near future.
 

Attachments

  • smsl.jpg
    smsl.jpg
    342.8 KB · Views: 34
I had the opportunity to compare the sound of the SMSL D1 and D200 DACs for quite a long time. I use the HQPlayer Embedded to resample all input streams to DSD512 and the DACs operate in direct DSD mode. There was no significant difference when changing devices in my particular audio configuration.

As a result, I kept the SMSL D1 and bought Sennheiser HD800 headphones, which more unlocked the capabilities of the BD34352EKV chip. Moreover, I don't need MQA support, volume control, Bluetooth Audio, PSM filter selection, etc. Replacing headphones is always more valuable for the sound than replacing a DAC.. :)

The only drawback of this device is not support for the DSD 48x format, which makes it impossible to use integer oversampling filters. To solve this task, manufacturers need to create new firmware for the USB XMOS-316. Hopefully this won't be such a difficult problem and will be implemented in the near future.

that's interesting given the D1 is like 3X less expensive. Or maybe not, if you believe that cost and sound quality doesn't always have to correlate.
What speakers and amplifier did you use in the comparison? just curious.
 
Last edited:
What speakers and amplifier did you use in the comparison?

I'm currently using the iFi Zen Can amplifier with Sennheiser HD650 and HD800 headphones.

Both the D1 and D200 uses the exact same output components and design.

This isn't entirely true, the D200's analog output contains more components and uses higher-quality parts. However, this doesn't significantly affect the sound.
 

Attachments

  • d1.jpg
    d1.jpg
    71.9 KB · Views: 35
  • d200.jpg
    d200.jpg
    83.3 KB · Views: 33
I'm currently using the iFi Zen Can amplifier with Sennheiser HD650 and HD800 headphones.



This isn't entirely true, the D200's analog output contains more components and uses higher-quality parts. However, this doesn't significantly affect the sound.
more components is often worse, not better if there are more in the signal path.
Also longer PCB traces (larger circuit board) is also usually worse.
How do you know they are higher-quality, the chip labels are all sanded down in the D1.
I saw in one review they said the D200 uses OPA1612 opamps but we don't know what the D1 uses (if it matters) and the D200 also uses some unknown/sanded preamp chips.

what looks obviously good in the D200 to me is the power supply.
1772575315532.png
 
Last edited:
I'm currently using the iFi Zen Can amplifier with Sennheiser HD650 and HD800 headphones.



This isn't entirely true, the D200's analog output contains more components and uses higher-quality parts. However, this doesn't significantly affect the sound.
more components is often worse, not better if there are more in the signal path.
Also longer PCB traces (larger circuit board) is also usually worse.
How do you know they are higher-quality, the chip labels are all sanded down in the D1.
I saw in one review they said the D200 uses OPA1612 opamps but we don't know what the D1 uses (if it matters) and the D200 also uses some unknown/sanded preamp chips.

what looks obviously good in the D200 to me is the power supply.
View attachment 515057
The D200 has a more sophisticated I/V stage/filtering after the DAC IC and additional buffer stages in conjunction with the volume control ICs. Based on NTTY's measurements for the D200, this seems to be very well implemented, especially for the balanced output. The single-ended output is indeed somewhat weaker.
Both devices use unknown, sanded op-amps directly after the DAC chip.
The transformer is a rather cheap type with simple filtering. I would prefer a good switching power supply with high-quality filtering, but I'd rather have the transformer than a cheap switching power supply soldered directly onto the board.
In a brief initial comparison, I didn't notice any difference in sound, but I will test this more thoroughly in the coming weeks.

In the digital domain, circuits and paths, whether internal or external, are not a problem and have no influence as long as no 1 or 0 is changed. Even if they send the digital signal five times around the world and into space and back.

Of course, components and circuits can influence the analog music signal, just as the signal can be compromised by interference on the circuit boards.
However, with good implementations and developments, this is absolutely negligible and definitely without any audible and usually also measurable influence.
This can also be demonstrated with two simple practical examples.

On ASR, you can find recordings of an 8-way AD/DA loop, i.e., a recording in which the original signal passed through all the components of an analog-to-digital conversion and a digital-to-analog conversion eight times. Audible difference? No.

In the studio, the signal travels countless meters of cable, undergoes many balanced-to-single-channel and single-channel-to-balanced conversions, travels many meters of circuit traces on circuit boards, and passes through tens, thousands, or more components.
And then those few centimeters and components in a well-implemented and developed DAC are supposed to make a noticeable difference?
That definitely belongs in audiophile fantasy land.
 
Hello everyone, I have just registered, in order to participate in your forum and in particular the topic dedicated to the DAC SMSL D1. I have had this DAC since the end of December 2025, I have no problem with the controls, no switching noise, in short everything is normal and the sound suits me perfectly, it replaces a Micromega DAC of more than 20 years, the gain is obvious. Curiosity pushed me to open D1 and the version is different, the springs are fixed on a circuit at the bottom of the box and connected to the main circuit using a cable.
7B251BD9-8E3E-4BC2-B9F6-E60E33998B84_1_105_c.jpeg
 
Hello everyone, I have just registered, in order to participate in your forum and in particular the topic dedicated to the DAC SMSL D1. I have had this DAC since the end of December 2025, I have no problem with the controls, no switching noise, in short everything is normal and the sound suits me perfectly, it replaces a Micromega DAC of more than 20 years, the gain is obvious. Curiosity pushed me to open D1 and the version is different, the springs are fixed on a circuit at the bottom of the box and connected to the main circuit using a cable.
View attachment 515163
The changes only affect the operation of the touch buttons/fields on the front panel.
Contrary to the initial statement that the D1 might be discontinued, possibly due to the original touch issue, this is clearly a new revision, so the D1 will likely remain available to us all in the future.

Instead of springs, the "touch events" are now transmitted via ribbon cable, resulting in much more consistent operation.
This has no effect on the actual function or sound.

So, for everyone who wants to buy a D1, the touch issue should be finally resolved with this revision.
The rest of the components appear completely unchanged.
 
Back
Top Bottom