Chord upsampling

[Fluffy]

I recently went to Canjam in London. In Chord's booth I was shown the Hugo M Scaler, which to my understanding basically goes between the computer and the dac, and upsamples the PCM signal sent through it. They claim it takes the basic 44.1 khz sample rate and upscale it to something ridiculous like 768kHz and it makes it somehow sound better.

I listened to a couple of tracks and tried to compare between the "max" upscaling mode and the bypass mode that performs no upsampling. I didn't notice any dramatic difference, but I do think I heard something like a very tiny (like 1%) increase in "richness". I don't know if that's really connected to the upsampling, and it could equally be caused by some hidden EQ fuckery of some IMD effect that adds more harmonics thus making the sound perceivably richer.

Is there any truth to their claims of a better sound through upsampling? Why would that even be possible, giving that upsampling doesn't add any new information to the signal?

Or could it be just another $4,995 (!!!@#) snake oil product?

 

[kaka89]

If upsampling is so great, record companies should just upsample their music file once and then sell it as Hi-Res.

No point to upsample it again and again during playback in customer's end.

 

[solderdude]

Is there any truth to their claims of a better sound through upsampling? Why would that even be possible, giving that upsampling doesn't add any new information to the signal?

All DS DACs do 'upsampling'.. just not from 44.1/24 to 768k/24 but several MHz/4 (or something in that area)
In both cases there is a 22.05kHz (steep) filter active and neither options will (should) output relevant signals above 20kHz.

 

[Roasty]

I tried the mscaler+hugo tt2 combo with a Meze Empyrean with music from tidal via my phone. I do own the Empyrean and my setup at home plays through a dac3b and hpa4.
I played only music I am well accustomed to.

I must admit, I did very much enjoy the chord stack. Everything sounded much clearer. There was a nice sense of space around each instrument and voice in the music. Vocals sounded luscious, palpable, real. I wish I had better descriptors, but I guess I can best put this in the way it felt; like everything was in place, at the right time. The speed with which music seemed to appear out of nowhere was simply a joy. It gave the Meze a whole new sound. As cliche as it sounds, a veil was lifted. I did not know my headphones could output music like that. And I really enjoyed the experience.

I have read many posts on this forum saying how the chord stack is overpriced and just bull crap and how the purported sound can be replicated with software etc

But damn.. It does sound good.

 

[Fluffy]

Roasty, i'm not sure if you're joking or not...

 

[Roasty]

Roasty, i'm not sure if you're joking or not...

Haha.. well I'm not. I'm only describing what I heard and felt. I went into the shop skeptical, having read a lot of posts about the mscaler on this forum. My ears heard a difference, which I enjoyed, and that's all I'm saying.

 

[Music1969]

But damn.. It does sound good.

But did you hear the Hugo TT2 without M-Scaler, to hear what it adds (if anything)?

 

[Music1969]

All DS DACs do 'upsampling'.. just not from 44.1/24 to 768k/24 but several MHz/4 (or something in that area)
In both cases there is a 22.05kHz (steep) filter active and neither options will (should) output relevant signals above 20kHz.

And the current Chord DACs noise shaping runs at 5bit ~104 MHz SDM.

 

[SIY]

I went into the shop skeptical, having read a lot of posts about the mscaler on this forum. My ears heard a difference, which I enjoyed, and that's all I'm saying.

They probably did. Shops are well aware of the effect of small level changes on the perception of sound. Especially when there's some subliminal coaching going on.

 

[solderdude]

Chord makes nice performing but overly expensive stuff.
The audibility of it is another matter of course.

 

[Roasty]

Just to add on to my previous post.

I've had the Empyreans for several weeks now. My take on the Empys are they are a very "polite" set of headphones. The highs are smooth and rounded and never sibilant/bright. They have a sufficient amount of bass, but there is some blending into midbass which makes it sound occasionally bloated. Most music sounds good with the empy, but the overall presentation is seemingly quite flat/bland when I compare it to my Utopias, which I find a much more engaging headphone.

With TT2 alone, I did not perceive much of a difference from my own setup. Granted, I only listened to the TT2 (alone) for a few minutes before the mscaler was added to the equation.

With the tt2/mscaler, it felt like the empy had opened up. Speed and clarity almost on par with the Utopia. More upfront presentation. Less of a flat soundstage and less of a polite sound. I really didn't want to stop listening and I wanted to go through my whole "favourites" playlist but I just didn't have the time that day.

The TT2 and mscaler combo was offered to me at a very very attractive price, but I declined the purchase that day. The stack is really insanely priced.. I'm not sure that amount of wallet hurt is worth the audio sparkle. I can do a lot of things with that amount of money eg pay kids tuition. If i had like a billion dollars to spare, then yea why not.

 

[SIY]

Some double blind level-matched evaluation might prove enlightening to you. Your sonic claims are extraordinary and beg for some controls.

 

[Roasty]

Some double blind level-matched evaluation might prove enlightening to you. Your sonic claims are extraordinary and beg for some controls.

No doubt. I completely agree. I don't discount expectation bias or whatever other bias may have come into play during my listening session.

 

[SIY]

No doubt. I completely agree. I don't discount expectation bias or whatever other bias may have come into play during my listening session.

This makes you about 10,000x more intellectually honest than most "audiophiles."

 

[Fluffy]

Even if that is what you heard, honestly I'm not really interested in vague subjective descriptions. I want to know what is the objective basis to their claims, if there is even one. FFS, this device alone costs like a really high end amp or dac, it basically have one function, and cannot be used independently. What possible technology could there be in there that will justify that price tag?

 

[Roasty]

I'm interested to know too.
Perhaps some of the folks here should go into the mscaler thread on headfi and stir up all the Chord fan boys there and maybe one of them will send Amir the chord stack for review.

 

[SIY]

I'm interested to know too.
Perhaps some of the folks here should go into the mscaler thread on headfi and stir up all the Chord fan boys there and maybe one of them will send Amir the chord stack for review.

Nope, what will happen (based on my brief experience there) is that posts questioning advertisers will be censored, and bans will happen following the uproar of the flying monkeys that inhabit their forum section.

 

[M00ndancer]

Or could it be just another $4,995 (!!!@#) snake oil product?

Short answer: Yes.
If it works as advertised, there isn't any info to "fix", anything above 22050 Hz will be inaudible. Unless the source is broken or engineered badly.

 

[Fluffy]

I've had the Empyreans for several weeks now. My take on the Empys are they are a very "polite" set of headphones. The highs are smooth and rounded and never sibilant/bright. They have a sufficient amount of bass, but there is some blending into midbass which makes it sound occasionally bloated. Most music sounds good with the empy, but the overall presentation is seemingly quite flat/bland when I compare it to my Utopias, which I find a much more engaging headphone.

By the way, it's interesting what you are saying about the Empyreans. I listened to several of those with different equipment, and they all sounded to me like they have too much energy in the very high treble (10K+). I have a friend that also had them and he also claim that they sound smooth and rounded, with no hint of the high treble I was hearing. He happens to be 45+ years of age, while I'm 29. I wonder if the difference is due to age difference and age related hearing loss.

And plus, I have these two measurements from SoundStage, one is the Focal Clear which I own and the treble there sounds to me very flat and nice:

And the Meze Empyrean, that shows much elevated levels of high treble:

 

[Mtbf]

I don’t think that it is the upscaling that is the (most important) property of the M Scaler, it is the use of very long linear FIR filters that is supposed to make it special. The Chord designer is postulating for many years that a perfect sinc(x) function, and thus (t)his FIR filtering with a very large number of taps is essential for a perfect reconstruction of the original analogue signal from standard redbook 16 bits/44.1 kHz digital. He doesn’t concur in that respect with the use of minimal phase filtering (including MQA), as do others.

As I understand it the M Scaler is doing just the same upscaling and filtering as the other Chord DAC’s do, but with a much higher tap count, while all of the Chord DAC’s already use a higher number of filter taps than of-the-shelf DAC chips from ESS, AKM, and the like.

@amirm commented on this subject in another post, and mentioned the issue of diminishing returns:

The FPGA is not performing the DAC function.

In simple terms, the Chord DACs first upsample the audio to much higher rate so that they don't need as many bits to convert to analog. The FPGA performs this function as would any other upsampling DAC. Where it differs from other DACs is that its reconstruction filter uses far more audio samples to perform its filtering than off-the-shelf filters do. Rob Watts says this improves fidelity but that is a hand wave. What is implemented in mass market DAC chips is already at point of diminishing returns. Going 1000 times more taps (audio samples) doesn't amount to much improvement. The asymptop has already been reached.

FPGAs are inefficient to use by the way compared to hard silicon such as what is in a DAC chip. The main benefit of FPGA is that it can be programmed over and over again whereas a custom IC cannot. The price for that is higher cost and higher power consumption. An ASIC can be produced from an FPGA to get those benefits but it will have NRE up front costs north of a million dollars plus that much in engineering design cost (and tools). So it is not done in the scale that Chord sells these DACs.

As an analog, think of a car with 100 exhaust pipes instead of two. That is what Chord says will produce more horsepower. For measurements though, they show 0.01% more horsepower and can't show anyone actually feeling the improvement. They incur the expense, heat, etc. but the benefit simply is not there.