CHORD M-Scaler Review (Upsampler)

[tmtomh]

This is a conversation that will go nowhere because I cannot provide the measurements or data to support Chord's / Rob Watts' claims.

I am also not saying I know that it does what he says. I am merely saying that something is happening that results in a sound that I prefer and I am accepting his explanation in lieu of any other reasonable one. If someone could demonstrate negative traits like added harmonics, I'd accept that happily - it's not about it having to prove or disprove Rob's claims.

There's been no clarity provided about how the theory/hypothesis that the WTA filters are built on (i.e. the benefits of extending the sinc function as close as possible to its infinite products) is wrong. Instead we have this circular argument about how it can't be possible because it isn't possible.

Let's take this to a slightly different, but related topic for a moment. Is the takeaway from all of this that upsampling in general does nothing? Does software like HQ Player offer no benefit to the sound quality?

Thanks again for sticking with this conversation. I agree it is likely to end in an impasse, but at the same time I do think some points of disagreement are coming into sharper focus, and that in and of itself can be a good and useful thing.

You don't improve the sound by upsampling. Once again, the reason you don't improve the sound by upsampling is not because I can't hear any improvement or someone else can't hear any improvement. The reason is that upsampling simply, and literally, just duplicates samples. So if you upsample from 44.1kHz to 176.4kHz, that's 4x oversampling, which means every original sample is replicated three more times - it's copy-pasted three times so that every original sample is now four samples. All four samples are identical - there is no "interpolation" like you might get with a TV that increases the frame rate by creating new frames that are combinations of the frames before and after them. That's not how upsampling works - it just copies the existing sample exactly.

It might be helpful or useful to ask, Why not create a digital upsampling system that DOES interpolate samples, that looks at Sample 1 and Sample 2 and creates a new Sample 1.5 in between them that is a combination of them, to "smooth" the transition from Sample 1 to Sample 2 just like video frame interpolation on a TV smooths the perceived motion of the moving video?

The answer is that digital audio doesn't work that way: there is only one way to get from Sample 1 to Sample 2. There is no need for interpolation in the digital domain, and in fact the idea of synthesizing such an interpolated new sample doesn't even make any sense. The reconstruction filter does the smoothing during the digital-to-analogue conversion step. It doesn't - and can't - happen within the digital domain.

The only scenario in which there is more than one way to get from Sample 1 to Sample 2 is if the original audio signal contains frequencies higher than 1/2 the sample rate of the digital recording device. That's why analogue signals going into an ADC (analogue to digital converter aka a digital recorder) have to be (or at least should be!) band-limited so that they contain no frequencies higher than 1/2 the sample rate.

This is the same principle that dictates that any frequency needs to be sampled only twice - and that sampling it more than that does nothing and changes nothing.

So no, HQ Player's upsampling does nothing - or more precisely, increasing the sample rate does not change or refine the sound, and it doesn't enhance time accuracy because sample rate has nothing to do with time accuracy - it has to do with which frequencies can be accurately encoded. If you try to record a 15kHz signal using a 20kHz sample rate, then there will be a "time inaccuracy" because a 20kHz sample rate can only encode frequencies up to 10kHz. So it will encode that 15kHz signal as 5kHz (10kHz minus 5kHz instead of the original and proper 10kHz + 5kHz = 15kHz). So the "time inaccuracy" will be that the frequency - the speed, the timing - of the signal will be aliased; reproduced at 5000 cycles per second instead of the original 15000 cycles per second. But as I'm sure you can see, that "timing inaccuracy" would not be subtle (to say the least!) and would be an indication of a fundamental operational error rather than a subtle change or improvement in the "refinement," "pacing," or "precision" of the sound.

As has been explained by folks with far more knowledge than I have, increasing the sample rate does allow the reconstruction filter to operate in a larger frequency range above the limit of human hearing but below the Nyquist limit of 1/2 the sample rate. However, as those with more knowledge than me - including Amir - have shown, Chord products do not need to take advantage of that extra "wiggle room" - instead they use their gazillion taps to create a very well-implemented reconstruction filter that has tons of ultrasonic attenuation and does its work in a very small/narrow frequency band (Amir has repeatedly acknowledged that Chord products have great reconstruction filters). So with Chord products, to the best of my knowledge even this one potential implementation benefit of oversampling is not necessary or taken advantage of.

 

[Dogcoop]

@PassionforSound I am interested in your reply to post #1201.
If you are not able to accept the conclusion, then please provide the technical explanation for why you don’t agree. If you cannot disprove the conclusions, then maybe you could forward it to rw for his response and explanation.

 

[chesebert]

You zero pad and not duplicating when you upsample.

 

[Blumlein 88]

So the position taken by passion for Sound is "I don't understand sampling theory, you cannot explain it to me, so I believe Mr. Watts."

 

[PassionforSound]

So the position taken by passion for Sound is "I don't understand sampling theory, you cannot explain it to me, so I believe Mr. Watts."

No, you're misrepresenting what I've stated clearly at least twice now.

 

[Dogcoop]

No, you're misrepresenting what I've stated clearly at least twice now.

Could you state it one more time for clarity. This way we will all be discussing the same points. Thnx.

 

[PassionforSound]

Thanks again for sticking with this conversation. I agree it is likely to end in an impasse, but at the same time I do think some points of disagreement are coming into sharper focus, and that in and of itself can be a good and useful thing.

You don't improve the sound by upsampling. Once again, the reason you don't improve the sound by upsampling is not because I can't hear any improvement or someone else can't hear any improvement. The reason is that upsampling simply, and literally, just duplicates samples. So if you upsample from 44.1kHz to 176.4kHz, that's 4x oversampling, which means every original sample is replicated three more times - it's copy-pasted three times so that every original sample is now four samples. All four samples are identical - there is no "interpolation" like you might get with a TV that increases the frame rate by creating new frames that are combinations of the frames before and after them. That's not how upsampling works - it just copies the existing sample exactly.

It might be helpful or useful to ask, Why not create a digital upsampling system that DOES interpolate samples, that looks at Sample 1 and Sample 2 and creates a new Sample 1.5 in between them that is a combination of them, to "smooth" the transition from Sample 1 to Sample 2 just like video frame interpolation on a TV smooths the perceived motion of the moving video?

The answer is that digital audio doesn't work that way: there is only one way to get from Sample 1 to Sample 2. There is no need for interpolation in the digital domain, and in fact the idea of synthesizing such an interpolated new sample doesn't even make any sense. The reconstruction filter does the smoothing during the digital-to-analogue conversion step. It doesn't - and can't - happen within the digital domain.

The only scenario in which there is more than one way to get from Sample 1 to Sample 2 is if the original audio signal contains frequencies higher than 1/2 the sample rate of the digital recording device. That's why analogue signals going into an ADC (analogue to digital converter aka a digital recorder) have to be (or at least should be!) band-limited so that they contain no frequencies higher than 1/2 the sample rate.

This is the same principle that dictates that any frequency needs to be sampled only twice - and that sampling it more than that does nothing and changes nothing.

So no, HQ Player's upsampling does nothing - or more precisely, increasing the sample rate does not change or refine the sound, and it doesn't enhance time accuracy because sample rate has nothing to do with time accuracy - it has to do with which frequencies can be accurately encoded. If you try to record a 15kHz signal using a 20kHz sample rate, then there will be a "time inaccuracy" because a 20kHz sample rate can only encode frequencies up to 10kHz. So it will encode that 15kHz signal as 5kHz (10kHz minus 5kHz instead of the original and proper 10kHz + 5kHz = 15kHz). So the "time inaccuracy" will be that the frequency - the speed, the timing - of the signal will be aliased; reproduced at 5000 cycles per second instead of the original 15000 cycles per second. But as I'm sure you can see, that "timing inaccuracy" would not be subtle (to say the least!) and would be an indication of a fundamental operational error rather than a subtle change or improvement in the "refinement," "pacing," or "precision" of the sound.

As has been explained by folks with far more knowledge than I have, increasing the sample rate does allow the reconstruction filter to operate in a larger frequency range above the limit of human hearing but below the Nyquist limit of 1/2 the sample rate. However, as those with more knowledge than me - including Amir - have shown, Chord products do not need to take advantage of that extra "wiggle room" - instead they use their gazillion taps to create a very well-implemented reconstruction filter that has tons of ultrasonic attenuation and does its work in a very small/narrow frequency band (Amir has repeatedly acknowledged that Chord products have great reconstruction filters). So with Chord products, to the best of my knowledge even this one potential implementation benefit of oversampling is not necessary or taken advantage.

Once again, you've provided a thoughtful, insightful and interesting response - thank you!

This is exactly the kind of discussion I was hoping for and you've raised some really good points that I cannot respond to (which is great).

My current lay person's understanding is that, in a complex musical signal with many superimposed frequencies, the precise timing of an upswing or downswing in the waveform is not necessarily as cleanly predictable as an individual frequency in isolation. It has been my understanding that the upsampling process and WTA filters are intended to better resolve these details in time. As I said though, I don't have the foundations to suggest that my understanding is right - it is just the understanding I was working with. I am thrilled to be able to explore this further. Thank you.

 

[chesebert]

You need DSP 101

 

[amirm]

My current lay person's understanding is that, in a complex musical signal with many superimposed frequencies, the precise timing of an upswing or downswing in the waveform is not necessarily as cleanly predictable as an individual frequency in isolation. It has been my understanding that the upsampling process and WTA filters are intended to better resolve these details in time. As I said though, I don't have the foundations to suggest that my understanding is right - it is just the understanding I was working with. I am thrilled to be able to explore this further. Thank you.

Let me ask you this: does it not sound like you are getting something for nothing? That when you listen to that music as produced, said "resolve" is not there. But add this box and it pops out of ether. Right?

 

[Lukino]

Maybe it would be good to put two in a row. What the first breaks, the next fixes. The debaters' forest is starting to burn.

 

[Blumlein 88]

Once again, you've provided a thoughtful, insightful and interesting response - thank you!

This is exactly the kind of discussion I was hoping for and you've raised some really good points that I cannot respond to (which is great).

My current lay person's understanding is that, in a complex musical signal with many superimposed frequencies, the precise timing of an upswing or downswing in the waveform is not necessarily as cleanly predictable as an individual frequency in isolation. It has been my understanding that the upsampling process and WTA filters are intended to better resolve these details in time. As I said though, I don't have the foundations to suggest that my understanding is right - it is just the understanding I was working with. I am thrilled to be able to explore this further. Thank you.

You are mistaken about complex vs simple waveforms. The timing accuracy is 1/(sample rate x number of levels x 2pi). For 16 bit audio the number of levels is approximately 65 k. Read mansr links provided up thread for a more detailed explanation. It isn't just an idea. It is known to be true.

Do you disagree with this? Do you not understand it? It is okay if you don't, but obviously you not in a good position to argue persuasively otherwise.

 

[CapMan]

Hello:

I've been using TT2 & M Scaler through KEF Blade Speakers and an A-S3200 Yamaha Integrated for a while now.
M SCALER required around 100 hours to come around (even more pronounced than Hugo TT2).

DAVE with or without M SCALER was a bit too analytical for my liking but I can understand the appeal.

My Source is a home-built (pc-based) A/V Server with SoTM components (clock and dedicated USB).
I primarily listen to FLAC and DSD files from SSD.

With M Scaler turned onto "High", it sounds fantastic; there exists a musical realism that simply isn't there with it bypassed or eliminated from the equipment chain.
The "Maximum" Upscale setting on M SCALER I didn't care for.

I've just glanced at a few of the comments above and I don't know what else to say except I'd like to drop a quote from A. Einstein- "Not everything that can be counted counts and not everything that counts can be counted".

On a side-note, I did read an article about Speaker Cables and because the MEASURABLE differences were small between the various cables, the human-ear couldn't possibly detect a difference, according to the Author of the testing. Several cables later I found this simply wasn't true and finally settled on what I liked.

That article reminded me of the quote.

Lots of variables and opinions in high(er)-end Audio which (in part), makes it an interesting Hobby.

Please could you post a photo of your listening room - 360 degrees . Thanks

 

[voodooless]

Also, not being able to measure something doesn't make it inexistent. The Higgs boson couldn't be found/measured prior to 2012 but was theorised in 1964.

No, it was theorized using actual science, and later proven correct beyond a shadow of a doubt using experimentation.

Similarly, many of the medicines we are familiar with today (Lithium, Tylenol, Penicillin) are still not understood in terms of how they work. We just know that they do.

That is the whole point: we know! How do you think we know? By just popping a few and then concluding: wow! Or did people do a scientific double-blind study to figure out the efficacy of these drugs?

 

[KSTR]

You don't improve the sound by upsampling. Once again, the reason you don't improve the sound by upsampling is not because I can't hear any improvement or someone else can't hear any improvement. The reason is that upsampling simply, and literally, just duplicates samples. So if you upsample from 44.1kHz to 176.4kHz, that's 4x oversampling, which means every original sample is replicated three more times - it's copy-pasted three times so that every original sample is now four samples. All four samples are identical - there is no "interpolation" like you might get with a TV that increases the frame rate by creating new frames that are combinations of the frames before and after them. That's not how upsampling works - it just copies the existing sample exactly.

It might be helpful or useful to ask, Why not create a digital upsampling system that DOES interpolate samples, that looks at Sample 1 and Sample 2 and creates a new Sample 1.5 in between them that is a combination of them, to "smooth" the transition from Sample 1 to Sample 2 just like video frame interpolation on a TV smooths the perceived motion of the moving video?

The answer is that digital audio doesn't work that way: there is only one way to get from Sample 1 to Sample 2. There is no need for interpolation in the digital domain, and in fact the idea of synthesizing such an interpolated new sample doesn't even make any sense. The reconstruction filter does the smoothing during the digital-to-analogue conversion step. It doesn't - and can't - happen within the digital domain.

This is so clearly wrong I don't know even where to start.

For ages now, all DACs (unless exotic NOS) do internal digital upsampling and reconstruction filtering, creating the required intermediate sample's values.

Why in digital? Because it is almost impossible to make a halfway decent analog filter, notably for low sample rates.
After upsampling and filtering, the analog filter only has to take care for stuff above the upsampled conversion rate.

 

[KSTR]

What constitutes best, though? Is it stopband attenuation? Passband flatness? Transition sharpness? Since we can't maximise all of those at the same time, what compromises are acceptable?

That's wrong, we can maximise all of those at a time (the closer we get to true sinc() the better all three properties will become) until calculation artifacts kick in...

 

[KSTR]

It's also paradoxical. From what you say, the higher the resolution of the source file, the better an upscaler should work because it has more to work with.

Must be a misunderstanding here, I never meant to say this.

 

[Clavius]

Once again, you've provided a thoughtful, insightful and interesting response - thank you!

This is exactly the kind of discussion I was hoping for and you've raised some really good points that I cannot respond to (which is great).

My current lay person's understanding is that, in a complex musical signal with many superimposed frequencies, the precise timing of an upswing or downswing in the waveform is not necessarily as cleanly predictable as an individual frequency in isolation. It has been my understanding that the upsampling process and WTA filters are intended to better resolve these details in time. As I said though, I don't have the foundations to suggest that my understanding is right - it is just the understanding I was working with. I am thrilled to be able to explore this further. Thank you.

This is not a question of tech though, it’s even beyond objective vs subjective, it’s a question of logic.

Information lost is lost forever, that’s really the gist of what people here are saying.

Even if we disregard this fact, when I’m looking at your argument using the complexity of a music signal vs a simple wave in reverse; how would the up-sampler be able to re-create the musicality/goodness of a complex signal when it won’t work on a simple tone? To me this is the fruits of magical thinking in the vein of how blurry blobs of surveillance cams are “enhanced” by secret algorithms and cool nerds in movies. This is nothing but fantasy in a box sold at a fantastic price.

 

[CapMan]

The Higgs boson couldn't be found/measured prior to 2012 but was theorised in 1964.

As a newish member I’m pleased to see the particle physics defence being deployed, it’s my first time. I feel all grow’d up now .

 

[Dogcoop]

If an impulse measurement would be demonstrative and easy to produce, that's the type of data I am seeking. That is my one and only request - to see actual data that proves or disproves the claims that the M-Scaler does nothing.

You will definitely need to ask RW to perform that test because:

However, it seems the M-Scaler has test-detection measures built in, as when I feed an impulse through it it simply spits out a ‘Zero Order Hold’ result, copying that impulse a few times and giving a square result as shown below:

M-Scaler Impulse response test doesn’t work due to anti-testing measures built in

Raw samples shown
This is of course not the actual upsampling method the M-Scaler normally uses, it’s simply a method of preventing anyone from being able to potentially gain too much information or closely examine the impulse response. Potentially even being able to reverse engineer the exact design of the filter and windowing given enough time and expertise.
-Golden Sound

Only RW will be able to provide the request for impulse behavior.

Please request it as follow-up to your interview.

 

[solderdude]

There's no level difference. The M-Scaler pads the output in all modes to allow for the same volume across the board.

I am told bypass is at a different level than 2x,4x and 16x making any comparison between bypass and active impossible.