• 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 M500 DAC and HP Amp Review

mi-fu

Addicted to Fun and Learning
Forum Donor
Joined
Jun 7, 2018
Messages
589
Likes
660
Location
New York
#41
Hi @amirm , do you think it is possible to test if there is any difference between hardware-enabled MQA by M500 vs software-enabled MQA by Roon?

Probably it will be an interesting comparison, as the cost of M500 ($399) is similar to the lifetime subscription of Roon ($499).
 

gvl

Major Contributor
Joined
Mar 16, 2018
Messages
2,336
Likes
2,277
Location
SoCal
#43
I am fully aware of this and chose my words carefully in the review. The original file was at 192 kHz sampling and the decoding therefore remains faithful to that. There is no useful information above 96 kHz to preserve. So simply upsampling to get to the original sample rate is just fine in my book. This is what a perceptual high-res coder needs to do and MQA is doing.
Is it known if the MQA decoder in this DAC does anything above what you get in the native Tidal app or Roon? That is, does it do the second unfold?
 

gvl

Major Contributor
Joined
Mar 16, 2018
Messages
2,336
Likes
2,277
Location
SoCal
#44
I was pleasantly surprised to see a high output of nearly 7 volts at max volume. For fairness to other DACs measured, I cut the output down to 4.3 volt. Our worst case distortion product is below -120 dB, assuring complete transparency as that is well below our threshold of hearing. Same is true of SINAD which sums in the noise with it and still lands in truly inaudible distortion+noise.
Were the numbers different at 7V?
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
30,926
Likes
95,742
Location
Seattle Area
Thread Starter #46
This is not a weakness. It ensures that the user may enjoy this unit also tomorrow.
I have content that is not encoded at 0 dBFS. In those cases you may need higher gain/output. As well as someone using a headphone above 300 Ohm.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
30,926
Likes
95,742
Location
Seattle Area
Thread Starter #47
Were the numbers different at 7V?
Will measure and post later today. Just came in from digging dirt out of our drainage pipes and am out of breath. :)
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
30,926
Likes
95,742
Location
Seattle Area
Thread Starter #48
Hi @amirm , do you think it is possible to test if there is any difference between hardware-enabled MQA by M500 vs software-enabled MQA by Roon?
You mean with test signals? If so, I don't have those. Analysis with music is a lot more involved which I plan to do one day but not now.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
30,926
Likes
95,742
Location
Seattle Area
Thread Starter #49
Is it known if the MQA decoder in this DAC does anything above what you get in the native Tidal app or Roon? That is, does it do the second unfold?
That was the purpose of testing with 192 kHz. Below that Roon handles but above 96 kHz, you need a hardware decoder which this has.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
30,926
Likes
95,742
Location
Seattle Area
Thread Starter #50
Thanks for your reply. I don't doubt you chose your words carefully, but the question of whether or not there is any useful info about 96k is irrelevant to the question of whether MQA is decoding the original 192k or simply fabricating it. It's the latter, not the former.
No. Once again, the original file was at 192 kHz. So it makes sense that the reproduced version of it also be 192 kHz. If the content was at 96 kHz, you would have a point. But that is not the case here.

Whoever believed the original should have been at 192 kHz, wants to see the same come out of MQA process. It is missing the market dynamics to instead downsample that to 96 kHz.
 

LTig

Major Contributor
Forum Donor
Joined
Feb 27, 2019
Messages
2,767
Likes
4,230
Location
Europe
#51
I have content that is not encoded at 0 dBFS. In those cases you may need higher gain/output.
You need more gain, but not more power. I have no problem with a volume pot which gives 20 dB gain above 0 dBFS at max setting. It will distort of course when volume is set too high but this is just the same as with normal amplifiers.

As well as someone using a headphone above 300 Ohm.
More voltage, yes. Do such headphones still exist?
 

gvl

Major Contributor
Joined
Mar 16, 2018
Messages
2,336
Likes
2,277
Location
SoCal
#52
No. Once again, the original file was at 192 kHz. So it makes sense that the reproduced version of it also be 192 kHz. If the content was at 96 kHz, you would have a point. But that is not the case here.

Whoever believed the original should have been at 192 kHz, wants to see the same come out of MQA process. It is missing the market dynamics to instead downsample that to 96 kHz.
I seem to remember reading someone's analysis of MQA (was it Archimago?) that given a full unfold of the original 192kHz material the resulting equivalent "quality" is actually somewhere between 96kHz and 192. Wrong?
 

restorer-john

Major Contributor
Joined
Mar 1, 2018
Messages
5,895
Likes
14,182
Location
Gold Coast, Queensland, Australia
#53
I seem to remember reading someone's analysis of MQA (was it Archimago?) that given a full unfold of the original 192kHz material the equivalent "quality" is actually somewhere between 96kHz and 192. Wrong?
Just get hold of an MQA encoded series of pure test tones and publish full test results of that little experiment. I'm not holding my breath to see that one...

Just the mention of MQA alongside a D/A converter with this level of performance makes me smile.

Great review. :)
 

frogmeat69

Addicted to Fun and Learning
Forum Donor
Joined
Apr 25, 2018
Messages
680
Likes
1,049
Location
Western New York, USA
#54
So looks like SMSL has figured out the ESS 9038 hump issue as well?
 

tmtomh

Addicted to Fun and Learning
Joined
Aug 14, 2018
Messages
771
Likes
1,898
#56
No. Once again, the original file was at 192 kHz. So it makes sense that the reproduced version of it also be 192 kHz. If the content was at 96 kHz, you would have a point. But that is not the case here.

Whoever believed the original should have been at 192 kHz, wants to see the same come out of MQA process. It is missing the market dynamics to instead downsample that to 96 kHz.
Amir, if the market dynamics are a concern, then the MQA encoder shouldn't downsample the original 192k to 96k in the first place. The fact that the original file was 192k is completely irrelevant because the MQA encoder first converts that 192k PCM file to a 96k PCM file, and then does all its perceptual encoding and "folding" on that 96k PCM file. It is no different than if a human being manually ran the 192k PCM file through a downsampling app to convert it to 96k PCM and then ran that 96k PCM file through an MQA encoder. In that case, there's no way in the world you would ever try to justify MQA lighting up the "192k" light on the decoded playback.

The "192k" MQA file is misleading precisely because the original PCM file was 192k - it creates the false impression that MQA is restoring the original 192k sample-rate data when it is not.

The fact that the market expects/"likes" to have a 192k PCM original still be labeled as 192k when it comes out of an MQA Decoder does not mean that's what MQA should do - in fact, it makes it fraudulent for MQA (or any MQA device) to give the impression that the MQA file retains the original sample rate.

Either the ultrasonics afforded by a 192k sample rate are irrelevant, in which case there's no need for MQA to claim it does 192k; or else those ultrasonics are relevant, in which case it's a problem that MQA throws them out and totally fraudulent for MQA to claim that it retains them. You can't have it both ways.
 
Joined
Apr 24, 2019
Messages
37
Likes
31
#57
Can the quality/reduction of bit depth some how be measured when used at low volumes on this device? Does it alter the original digital signal when the digital volume control attenuates beyond a certain point?
 

BDWoody

Chief Cat Herder
Moderator
Forum Donor
Joined
Jan 9, 2019
Messages
3,800
Likes
8,881
Location
Mid-Atlantic, USA. (Maryland)
#59
Amir, if the market dynamics are a concern, then the MQA encoder shouldn't downsample the original 192k to 96k in the first place. The fact that the original file was 192k is completely irrelevant because the MQA encoder first converts that 192k PCM file to a 96k PCM file, and then does all its perceptual encoding and "folding" on that 96k PCM file. It is no different than if a human being manually ran the 192k PCM file through a downsampling app to convert it to 96k PCM and then ran that 96k PCM file through an MQA encoder. In that case, there's no way in the world you would ever try to justify MQA lighting up the "192k" light on the decoded playback.

The "192k" MQA file is misleading precisely because the original PCM file was 192k - it creates the false impression that MQA is restoring the original 192k sample-rate data when it is not.

The fact that the market expects/"likes" to have a 192k PCM original still be labeled as 192k when it comes out of an MQA Decoder does not mean that's what MQA should do - in fact, it makes it fraudulent for MQA (or any MQA device) to give the impression that the MQA file retains the original sample rate.

Either the ultrasonics afforded by a 192k sample rate are irrelevant, in which case there's no need for MQA to claim it does 192k; or else those ultrasonics are relevant, in which case it's a problem that MQA throws them out and totally fraudulent for MQA to claim that it retains them. You can't have it both ways.
Dead horse time...
Plus that misses the point... The little light came on, indicating a full unfold of the 192k file, per MQA.
 

tmtomh

Addicted to Fun and Learning
Joined
Aug 14, 2018
Messages
771
Likes
1,898
#60
I seem to remember reading someone's analysis of MQA (was it Archimago?) that given a full unfold of the original 192kHz material the resulting equivalent "quality" is actually somewhere between 96kHz and 192. Wrong?
Not exactly. MQA's "in-between" status is about bit-depth: it's supplied as high bit-depth (24-bit), but because it has to fold the ultrasonics into the lower bits in order to achieve its compression, the actual bit depth (aka noise floor) of an MQA file is less than 24-bit but more than 16-bit (when properly decoded).

As far as sample rate, though, MQA has a hard limit of 96k. Any 192k, or 384k, PCM file you feed into an MQA encoder will be downsampled to 96k before the encoder does any of its work. That downsampling step is not merely lossy - it's totally destructive: half the samples are thrown out and the upper Nyquist frequency is cut from 96k or 192k (half of 192k and 384k samples rates, respectively), to 48k (half of the new, downsampled 96k sample rate).

For files with sample rates of 176.4k or 352.8k, MQA downsamples them to 88.2k.

Personally I don't care about the downsampling - IMHO 88.2k and 96k are plenty. It's the claim that the full 192k or 176.4k sample rate of the original PCM file is being "restored" or "decoded" by MQA that I have a problem with - it's a demonstrably false claim.
 
Last edited:
Top Bottom