SMSL M500 DAC and HP Amp Review

[gvl]

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 (have 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.

Thanks. So sounds like the 192kHz illuminating doesn't necessarily mean the DAC is actually up-sampling to 192?

 

[gvl]

We'll have to agree to disagree about what the point is: the little light indeed indicates a full unfold - but even there your language (which is MQA's language) is revealing: the "final unfold" is not an unfold - it's just upsampling. Nothing wrong with upsampling - lots of playback software offers that as an option. But only MQA claims it's a restoration of something that was in the original file, when it most certainly is not. That's fraudulent IMHO.

I read once that the second unfold in hardware is a way to "correct" DAC's impulse response. It came from a company selling NOS DACs as an explanation why their DACs don't do the second unfold as there is nothing to correct. So the statement is a bit of a suspect.

 

[BDWoody]

We'll have to agree to disagree about what the point is: the little light indeed indicates a full unfold - but even there your language (which is MQA's language) is revealing: the "final unfold" is not an unfold - it's just upsampling. Nothing wrong with upsampling - lots of playback software offers that as an option. But only MQA claims it's a restoration of something that was in the original file, when it most certainly is not. That's fraudulent IMHO.

No disagreement with any of the substance. I had just hoped the dead horse from the MQA thread wouldn't get dragged into this one, or every MQA capable device going forward. I have no love for it, but if I bought a unit that did have the capability, I'd like to know that it at least functions as it was designed to.

 

[amirm]

You need more gain, but not more power.

One comes with the other. I have not seen a headphone amp with separate gain knob.

 

[tmtomh]

Thanks. So sounds like the 192kHz illuminating doesn't necessarily mean the DAC is actually up-sampling to 192?

Actually, the DAC is upsampling to 192k. The problem is that the MQA software/firmware is telling the DAC to upsample the 96k output to 192k because the original PCM file from which the MQA file was made happened to be a 192k file.

Here's another way to look at it, which might help. Here are two different scenarios that will lead to the DAC behaving differently even though the data the DAC is being fed and is processing with MQA is the same:

1. 24/192k PCM original. MQA encoder downsamples to 24/96k PCM, and then encodes it as MQA, "folding" the samples between 48 and 96k into a 24/48k MQA container file. DAC receives that file, decodes/unfolds to 24/96k, then upsamples to 192k, applies a digital filter before the analogue output stage, and displays "192k" as the output.
2. 24/96k PCM original. MQA encoder encodes it as MQA, "folding" the samples between 48 and 96k into a 24/48k MQA container file. DAC receives that file, decodes/unfolds to 24/96k, does not upsample, applies a digital filter before the analogue output stage, and displays "96k" as the output.

Here the problem becomes clear (I hope!): In both cases, the MQA encoder is "folding"/encoding the identical file: a 24/96k PCM file. And in both cases, the decode/"unfold" is identical, as is the digital filter applied. The only difference is that one file gets upsampled to 192k and the other doesn't. The only reason the 1st file gets upsampled is that the original PCM file happened to be 192k. But since MQA destroys the samples above 96k before doing its encoding/folding, that 192k doesn't matter - the MQA encoder is always encoding 96k PCM data, nothing more.

So by indicating "192k" for one encoded 96k PCM file while indicating "96k" for another encoded 96k PCM file, MQA is giving the impression that its "192k" mode is restoring something from the original file, when it most certainly is not - it is only manufacturing a higher sample rate by just doubling every sample.

And if this kind of manufactured upsampling is so great, then why not do it for the MQA file created from a 96k PCM original too? The answer has been given by Amir (whether or not he did so intentionally): it's marketing and nothing more.

 

[amirm]

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.

MQA "claims" that you give it a master at 192 kHz, and it gives you a file with backward compatibility at 192 kHz. If you are going to look under the wraps and say there was nothing there, then your argument is with the original guy who created the 192 kHz as well. So maybe we should outlaw high sample rate in DACs too?

Ultimately, those of you who are not customers of MQA don't matter. It is like a steak eater complain about the price of tofu. Members here wanted to see if this device fully decodes to MQA at 192 kHz and I tested and reported that it did. Those of us who play MQA content and the tag in Roon says 192 kHz but the DAC says 96 kHz find that odd. Having it match the same makes sense to us as customers. To the rest of you, none of it matters including high-res content so let' just move along.

 

[tmtomh]

MQA "claims" that you give it a master at 192 kHz, and it gives you a file with backward compatibility at 192 kHz. If you are going to look under the wraps and say there was nothing there, then your argument is with the original guy who created the 192 kHz as well. So maybe we should outlaw high sample rate in DACs too?

I'm sorry Amir but this is just a flat-out incorrect statement. If a 192k PCM file has ultrasonic information above 48kHz, the MQA encoder still will throw out that information when it downsamples the 192k original to 96k.

Then, when the MQA decoder shows "192k," it will in no way, shape or form be restoring or decoding any of the original ultrasonics above 48kHz, because those ultrasonics were forever thrown away when the MQA encoder first started processing the 192K PCM file and downsampled it to 96k.

Displaying "192k" is not backwards-compatibility: it's giving the impression to people who care about high ultrasonics (and I agree that's not you or me) that the "192k" MQA file contains/restores the full ultrasonics of the 192k PCM original. You know full well that's the impression MQA gives by using this system, and you know full well that that's the intent of their marketing. And it's fraudulent.

Many software and hardware vendors offer upsampling - but even the worst peddlers of woo-woo claims that upsampling enhances the sound still are up-front that they are upsampling. MQA will not even be honest at that baseline level: MQA has strongly implied that they are restoring or decoding the full original sample rate, which is not true - and you used the same language in your review of this DAC/HP amp.

 

[amirm]

I'm sorry Amir but this is just a flat-out incorrect statement. If a 192k PCM file has ultrasonic information above 48kHz, the MQA encoder still will throw out that information when it downsamples the 192k original to 96k.

What information? Music?

 

[GaryS]

I'm sorry Amir but this is just a flat-out incorrect statement. If a 192k PCM file has ultrasonic information above 48kHz, the MQA encoder still will throw out that information when it downsamples the 192k original to 96k.

Then, when the MQA decoder shows "192k," it will in no way, shape or form being restoring or decoding any of the original ultrasonics above 48kHz, because those ultrasonics were forever thrown away when the MQA encoder first started processing the 192K PCM file and downsampled it to 96k.

Displaying "192k" is not backwards-compatibility: it's giving the impression to people who care about high ultrasonics (and I agree that's not you or me) that the "192k" MQA file contains/restores the full ultrasonics of the 192k PCM original. You know full well that's the impression MQA gives by using this system, and you know full well that that's the intent of their marketing. And it's fraudulent.

Many software and hardware vendors offer upsampling - but even the worst peddlers of woo-woo claims that upsampling enhances the sound still are up-front that they are upsampling. MQA will not even be honest at that baseline level: MQA has strongly implied that they are restoring or decoding the full original sample rate, which is not true - and you used the same language in your review of this DAC/HP amp.

Hey man, if you want go over to Audiophilestyle and join their debate fine. They have 18,732 posts on MQA.

 

[BDWoody]

Hey man, if you want go over to Audiophilestyle and join their debate fine. They have 18,732 posts on MQA.

Or our very own thread here, to see what ground hasn't been covered...

https://audiosciencereview.com/foru...-of-controversies-concerns-and-cautions.2407/

 

[amirm]

Many software and hardware vendors offer upsampling - but even the worst peddlers of woo-woo claims that upsampling enhances the sound still are up-front that they are upsampling.

That is a distinction with a difference. Here is MSB on their upsampling:

People read such things and then put value on a file that plays at 192 kHz versus 96 kHz. This is why the original digital capture is done at 192 kHz. And why MQA's value proposition needs to be reproduction of the same.

This all goes back to people having such a myopic view of MQA. A new format is about technology, business proposition and marketing. It needs to have all of these to have a shot at success. You all keep wanting them to neuter their value proposition to appease people who would not be a customer regardless. It makes no sense to me.

 

[gvl]

I don't see tmtomh's MQA comments to be out of context here.

 

[amirm]

In general, I don't like us to have debates in review threads. I will let this go on a bit more but after that, people should go and post in the other thread.

 

[MZKM]

The low output of 50 millivolt which aims to measure the noise level for sensitive IEMs is a bit above average

Did you ever do this measurement on the Apple dongle?

 

[amirm]

Did you ever do this measurement on the Apple dongle?

No, this is a more recent test. All the ones I have done are in this table per review:

 

[AndrovichIV]

Off topic, but I guess there's two main reasons that people hate MQA:

1. The technical aspects which are a little bit shady as has been covered here. The fact that no one has been able to prove that there's an SQ improvement using MQA doesn't help
2. ... and the commercial aspect which involves replacing open source standards and codecs for proprietary ones that unlike the FOSS versions are not platform agnostic. Plus, the fact that Meridian, the company behind MQA, will get a cut in each part of the music process, thereby imposing some sort of tax on consumers

 

[tmtomh]

That is a distinction with a difference. Here is MSB on their upsampling:
View attachment 37222

People read such things and then put value on a file that plays at 192 kHz versus 96 kHz. This is why the original digital capture is done at 192 kHz. And why MQA's value proposition needs to be reproduction of the same.

This all goes back to people having such a myopic view of MQA. A new format is about technology, business proposition and marketing. It needs to have all of these to have a shot at success. You all keep wanting them to neuter their value proposition to appease people who would not be a customer regardless. It makes no sense to me.

No, I don't want them to neuter their value proposition. I want them not to lie in order to artificially bolster their value proposition.

"Enriched" white bread removes the bran and germ from the wheat to get the white color and soft texture. It then adds vitamins and minerals back in to compensate for those lost during processing. But the full range of nutrients are not restored, and the fiber and other beneficial elements are gone forever. By calling itself "enriched," however, it gives the impression that something is added to the original. It does not give the accurate impression that something is "added" only to an "original' that is a diminished version of the actual original.

Out of respect to your (totally reasonable) desire to keep these kinds of arguments out of review threads, I will stop pursuing this argument in this thread after this. But my point is that it is worse, not better, that a vendor lies to the people who care about the thing being lied about. If no one cared whether an MQA file really was 96k or 192k then their lie would be of no consequence. But they are lying because many of their potential customers do care about it - and those customers will unwittingly repeat the lie in their own social networks, and if the false claims of MQA gain momentum then R&D dollars and product dollars will be invested in MQA's BS and not in other aspects of product design that you and I both value (for example power supply quality and isolation; quality, low-noise analogue-stage components).

Any vendor lying about their product should ipso facto be of concern to the members of this site, period. It is precisely because "people read such things and then put value on a file that plays at 192kHz vs 96kHz" that we should be concerned that MQA is lying to those people and perpetrating a fraud.

 

[amirm]

Any vendor lying about their product should ipso facto be of concern to the members of this site, period.

It is not a lie. It is the way it needs to work. The entire encoding needs to fit in the few bits in the in-the-clear file. As such, you don't want to try to encode the spectrum above 48 kHz with your precious few bits and wind up distorting what is below that. That is exactly what I would implement if I had the same constraints. There is no musical information above 48 kHz even in the best case situation. So best to preserve its sampling rate and throw out whatever noise is up there. Engineering is about the right compromises and they have made the best one.

 

[confucius_zero]

I'm looking at this DAC and I'm very wet.

 

[Hugo9000]

@amirm Can you test how this device handles transitions from MQA material to non-MQA and vice-versa? Specifically, do the filters change automatically between the MQA filter and whichever is the default or user-selected filter for non-MQA material, or does the MQA filter "take over" and continue to be used for other material? (Mytek had issues pointed out in several reviews, can't remember other specific makes/models off-hand.)

What about gapless playback, as that has been mentioned as an issue with some MQA-capable DACs, where identifying the type of material (MQA or not, and if it needs to change the filter accordingly) causes a delay that interferes with true gapless playback? An easy test of gapless playback would be to use an opera recording, since they typically have many tracks that are simply access points in a continuous music stream, unlike most pop/rock albums.

Can MQA be disabled altogether if the device doesn't handle filter changes easily/correctly without glitches or audible issues (pops, clicks, gaps introduced when the material should have gapless playback, etc.). I realize that some people might suggest not buying this product at all if one doesn't want MQA playback, but there may be people who are interested due to the rest of its combined features and price point (DAC/preamp with balanced output/headphone amp/etc. for less than $500).

Thank you!