SMSL D200 DAC Review

[proud]

Not really. And headphones have far less soundstage than loudspeakers.

Do you use convolution filters to compensate for reflections from the walls of your listening room?

How wide is the actual stage during a live outdoor concert, with the huge crowd acting as a sound absorber?

 

[RandomEar]

Well, if you used an HD650 and matched to within 0.1 V, that might result in one DAC playing at 85 dB and the other playing around 90.5 dB (if you measured the voltage difference at such a comfortable listening level). That would explain any and all heard differences - you even said yourself one DAC sounded louder

Just to reiterate: To be in the safe side, you need to be below 0.2 dB of difference. 0.1 dB (not Volts!) would be ideal. You might have had a difference of over 5 dB in your test.

 

[proud]

Well, if you used an HD650 and matched to within 0.1 V, that might result in one DAC playing at 85 dB and the other playing around 90.5 dB (if you measured the voltage difference at such a comfortable listening level). That would explain any and all heard differences - you even said yourself one DAC sounded louder

Just to reiterate: To be in the safe side, you need to be below 0.2 dB of difference. 0.1 dB (not Volts!) would be ideal. You might have had a difference of over 5 dB in your test.

Yes, I completely agree with you. So, to reach 80 dB for my headphone, it only needs about 0.15 volts of sine wave voltage!
I asked about the blind testing conditions in this thread: https://www.audiosciencereview.com/...s/smsl-d1-rohm-dac-for-everyone.62806/page-26. You can read the responses for yourself. I didn't argue or try to prove anything because I wasn't treated very kindly after my desire to conduct a subjective evaluation. I simply agreed. No one even noticed that I wrote 0.1V instead of 0.01V, even though I indicated I used headphones for testing.

Of course, I used the Sound Meter app and compared sound pressure levels to an accuracy of 0.1 dB.

 

[RandomEar]

Yes, I completely agree with you. So, to reach 80 dB for my headphone, it only needs about 0.25 volts of sine wave voltage!
I asked about the blind testing conditions in this thread: https://www.audiosciencereview.com/...s/smsl-d1-rohm-dac-for-everyone.62806/page-26. You can read the responses for yourself. I didn't argue or try to prove anything because I wasn't treated very kindly after my desire to conduct a subjective evaluation. I simply agreed. No one even noticed that I wrote 0.1V instead of 0.01V, even though I indicated I used headphones for testing.

Of course, I used the Sound Meter app and compared sound pressure levels to an accuracy of 0.1 dB.

A difference of 0.01 V would still amount to up to 0.9 dB of difference in this case. An app measurement on a phone is not accurate to within 0.1 dB - you will never be able to hold the phone steady enough and environmental sounds influencing the measurement can account for much higher deviations.

That being said, you sound like you want people to be wrong or assume you were given bad advice on ASR. So, good luck with that.

I can only state that doing double blind testing correctly with solid, repeatable results is not an easy thing to do and in this case, it doesn't look like the test setup was solid.

 

[proud]

...you will never be able to hold the phone steady enough and environmental sounds influencing the measurement can account for much higher deviations

Why hold the phone in your hands? I placed it between the earpads of the headphones and took measurements in complete silence.

A difference of 0.01 V would still amount to up to 0.9 dB of difference in this case.

OK, I'll pick a time today, do the measurements again, and show here the voltage values with an accuracy of 0.001V depending on the sound pressure level.

 

[Svend P]

I can only state that doing double blind testing correctly with solid, repeatable results is not an easy thing to do and in this case, it doesn't look like the test setup was solid.

Wouldn't a null test be more telling in this situation? Record the electrical output of both devices, and see what the difference is.

 

[RandomEar]

Wouldn't a null test be more telling in this situation? Record the electrical output of both devices, and see what the difference is.

Yes. Far better and more precise. But you won't ever convince an audiophile that measurements are better than their ears.

 

[proud]

OK, I'll pick a time today, do the measurements again, and show here the voltage values with an accuracy of 0.001V depending on the sound pressure level.

I reran the measurements, taking into account all the comments and concerns. The results are as follows:
SPL 79.8 dB -> Vrms 0.131 V
SPL 80.0 dB -> Vrms 0.134 V
SPL 80.2 dB -> Vrms 0.137 V
I did not use a professional sound level meter and RF voltmeter. Therefore, the absolute values of the SPL and voltage levels may differ slightly from the reference values. But these measurements allow you to set the same volume levels from two signal sources with an accuracy of 0.1 dB.
I believe it would be more correct to use a white noise signal instead of a single 1 kHz tone. Then the full audio frequency band would be measured, and this is a more professional approach, but I am not going to impose my opinion on the esteemed audio science-based community.

 

[drewdawg999]

I'm about to pull the trigger on this sucker, also interested in DSD Direct functionality. Spent the night perusing this topic here, found it fascinating. There's a couple of guys there that are complete DSD junkies, and are very picky about DSD DACs, saying there's very few that offer true Direct DSD. The D200 is one of them, only other affordable options are iFi, which use an old BB chip.

I have a D-6 which is supposed to have Direct DSD, but they say the non-Direct mode measures the same, so something stinks there. Personally and subjectively I find the D-6 sounds fantastic, and even better on DSD. I want my DSD and I want it served up properly, so I believe this D200 will be mine in the near future.

 

[proud]

I want my DSD and I want it served up properly, so I believe this D200 will be mine in the near future.

For a first taste, you can listen to the SMSL D1. Believe me, there's little difference in sound when playing direct DSD, as they only use the output low-pass filter of the DAC chip. This is just a small part of the BD34352EKV's circuit diagram.
To fully utilize the chip's potential, I recommend using HQPlayer Embedded, which runs on Linux OS. And, of course, using the NAA network adapter to transfer source files for upsampling or DSD streaming, depending on the audio path configuration.

 

[drewdawg999]

For a first taste, you can listen to the SMSL D1. Believe me, there's little difference in sound when playing direct DSD, as they only use the output low-pass filter of the DAC chip. This is just a small part of the BD34352EKV's circuit diagram.

Okay, except I like balanced connection since my amp and Pre90 are already equipped for it. I did think the D-6 sounded better (yet softer) in Direct DSD mode, but last time I tried on headphones it didn't seem to make a difference. According to this one guy's measurements, he only guaranteed decent DSD performance for the D200 as far as Rohm chips go.

To fully utilize the chip's potential, I recommend using HQPlayer Embedded, which runs on Linux OS. And, of course, using the NAA network adapter to transfer source files for upsampling or DSD streaming, depending on the audio path configuration.

I'm only using foobar2000 on a Windows laptop. Will look into HQPlayer for Windows 11, not sure of the NAA stuff. Seems like they suggest upsampling to DSD256, I've been skeptical of that before but I guess I can give it a try.

 

[drewdawg999]

Went and dug up some DSD measurements on another site here. It shows performance at DSD512 to be very good. I'm intrigued, but alas I might have to wait for next month, AliExpress coupon expires in 45 minutes and I'm hesitant. I like the casework though, was gonna get the D400 Pro someday but this looks new and improved.

 

[Roland_606]

Yes, I completely agree with you. So, to reach 80 dB for my headphone, it only needs about 0.15 volts of sine wave voltage!
I asked about the blind testing conditions in this thread: https://www.audiosciencereview.com/...s/smsl-d1-rohm-dac-for-everyone.62806/page-26. You can read the responses for yourself. I didn't argue or try to prove anything because I wasn't treated very kindly after my desire to conduct a subjective evaluation. I simply agreed. No one even noticed that I wrote 0.1V instead of 0.01V, even though I indicated I used headphones for testing.

Of course, I used the Sound Meter app and compared sound pressure levels to an accuracy of 0.1 dB.

I don’t recall you stating 0.1V in that thread. You were given some decent advice I thought. Also, your sound meter app and mic are unlikely to even meet Class 2 specification (BS EN 61672) even if you use a Class 1 or 2 mic attached. Precision of Class 1 at 1kHz is +/- 1.1 dB.

 

[proud]

I did think the D-6 sounded better (yet softer) in Direct DSD mode, but last time I tried on headphones it didn't seem to make a difference.

Which model are you talking about, the SMSL D-6? It uses an ESS chip and doesn't have a direct DSD.

According to this one guy's measurements, he only guaranteed decent DSD performance for the D200 as far as Rohm chips go.

Yes, I constantly review measurement results in this area of interest on other forums. But these measurements are done primarily to be 100% certain that direct DSD is being used, and not some kind of conversion or processing. Some manufacturers don't provide specific information about DSD.

I'm only using foobar2000 on a Windows laptop. Will look into HQPlayer for Windows 11, not sure of the NAA stuff. Seems like they suggest upsampling to DSD256, I've been skeptical of that before but I guess I can give it a try.

I also started from the same starting point 6 years ago...
Please read the DSD discussion in more detail on other forums, and you will see that my setup tips may be useful.
Yes, it does require additional investment, but with the right approach it may not be that expensive.

 

[proud]

I don’t recall you stating 0.1V in that thread. You were given some decent advice I thought. Also, your sound meter app and mic are unlikely to even meet Class 2 specification (BS EN 61672) even if you use a Class 1 or 2 mic attached. Precision of Class 1 at 1kHz is +/- 1.1 dB.

Do you think I made such a huge mistake because of the microphone's characteristics, and that only the difference in volume affected the test results?

And if I use the most advanced measurement equipment to equalize volume, will two completely different DACs sound absolutely identical?

 

[VientoB]

Expectation bias is one hell of a thing.

 

[proud]

Expectation bias is one hell of a thing.

The fact that the Earth is round also came as a surprise to many. Some were even burned at the stake for it..

 

[Roland_606]

Do you think I made such a huge mistake because of the microphone's characteristics, and that only the difference in volume affected the test results?

And if I use the most advanced measurement equipment to equalize volume, will two completely different DACs sound absolutely identical?

Not just the microphone but the whole measurement system. When anybody wants to discuss accuracy within 0.1 dB and then states they used a phone app, I can’t help but think the results are not going to be, well you know, very scientific. I’m sure you had fun and the subjective observations you make of the D200 kind of follow my experience with this dac. But then again, it could be expectation bias on my part. Thank you for posting your observations.

 

[proud]

Not just the microphone but the whole measurement system. When anybody wants to discuss accuracy within 0.1 dB and then states they used a phone app, I can’t help but think the results are not going to be, well you know, very scientific.

What exactly can be the source of error when comparing loudness levels?
Is it really that difficult to measure a 1 kHz sine wave?
What exactly in the system can lead to unacceptable errors in obtaining the correct result?

Let's say the error of the microphone + app is 10%. How will this affect a simple comparison of two levels, especially since this was confirmed by parallel measurements of the signal's electrical parameters?

 

[VientoB]

The fact that the Earth is round also came as a surprise to many. Some were even burned at the stake for it..

True. One day you'll get it.