Jump to content
IGNORED

Chords New M -Scaler


Recommended Posts

On 8/5/2018 at 4:22 PM, auricgoldfinger said:

 

Search your HQP thread.  I believe it has been reported many times by various users, and you have explained something along the lines of the way the software is designed causes this problem. 

I'm not sure why you are having glitches/hangs with HQPlayer. But I'd say this, I've been using HQPlayer non-stop for the last three years. It has never happened to me (not even once) that HQPlayer hangs. Any software is prone to hangs/unresponsiveness. Even the mighty Google Chrome browser hangs now and then on my mac. There might be something you are doing wrong with your OS settings perhaps.

Link to comment
2 hours ago, Advieira said:

How do you guys compare PSAUDIO Direcstream with Hugo 2/Qutest "scaled" by Blu Mkii dual BNC?

I would expect folks to hear big significant between the DS and Chord DACs, after all, the PS Audio DS uses a single bit/DSD approach, where Chord uses a multi bit approach.

It might take awhile for people to report in on this, as the M Scaler is only just becoming available, and I doubt anyone is using the Blu with Qutest or Hugo 2. 

SO/ROON/HQPe: DSD 512-Sonore opticalModuleDeluxe-Signature Rendu optical with Well Tempered Clock--DIY DSC-2 DAC with SC Pure Clock--DIY Purifi Amplifier-Focus Audio FS888 speakers-JL E 112 sub-Nordost Tyr USB, DIY EventHorizon AC cables, Iconoclast XLR & speaker cables, Synergistic Purple Fuses, Spacetime system clarifiers.  ISOAcoustics Oreas footers.                                                       

                                                                                           SONORE computer audio

Link to comment
On 9/18/2018 at 10:31 PM, barrows said:

I would expect folks to hear big significant between the DS and Chord DACs, after all, the PS Audio DS uses a single bit/DSD approach, where Chord uses a multi bit approach.

It might take awhile for people to report in on this, as the M Scaler is only just becoming available, and I doubt anyone is using the Blu with Qutest or Hugo 2. 

There are already some reports. Plrase see:

fi.org/threads/hugo-m-scaler-by-chord-electronics-the-official-thread.885042/

Jensen VRD-iFF>Router>Rj45>opticalModule>
SFP>Buffalo2016>SFP>opticalModule >Rj45>

IZen Mk3>Rj45> Delock62619>Rj45>
etherRegen (Master Clock+ Mini-Circuits BLP)>SFP>opticalRendu>USB>IsoRegen>

USB>Phoenix>USB>OPPO 205 (Modded)>HMS “the Perfect Match”>Proac Tablette Reference 8 Signature.
 

Link to comment
On 9/24/2018 at 12:26 AM, audio.bill said:

Sorry for the bad link. Yes this is the one. MScalers are used with the little brothers. ?

Jensen VRD-iFF>Router>Rj45>opticalModule>
SFP>Buffalo2016>SFP>opticalModule >Rj45>

IZen Mk3>Rj45> Delock62619>Rj45>
etherRegen (Master Clock+ Mini-Circuits BLP)>SFP>opticalRendu>USB>IsoRegen>

USB>Phoenix>USB>OPPO 205 (Modded)>HMS “the Perfect Match”>Proac Tablette Reference 8 Signature.
 

Link to comment
  • 2 weeks later...
On 7/29/2018 at 12:42 PM, Miska said:

 

Hmmh, what am I misrepresenting or promoting? Since 99+% of content available is RedBook, I think it is sensible to focus on repduction of this content when discussing upsampling.

 

 

It's the same, here's AK4490:

NT503-sweep-pcm441-sdl-sh.thumb.png.e8a689a6ece56fdae660c3362a6364a2.png

 

ES9038, more down, but still visible (linear interpolation?):

PreBoxS2-44k-sweep-wide.thumb.png.990b148022f9146e7140660f95df3cec.png

 

 

Some interesting plots of Chord Mojo. 1 kHz tone at 705.6k input rate:

Mojo-1k_705-wide.thumb.png.5ca4fce07e6de78b8293f7f1df5ec2fd.png

Note the curious ripple in 1.25 - 2 MHz range.

 

Mojo-1k_705-narrow.thumb.png.7c280c10b32531136ee5836f6235d6d7.png

Zoom in, you can see some spurious tones riding the modulator noise bump.

 

1 kHz tone, but 768k input rate:

Mojo-1k_768-wide.thumb.png.06c15e17be8ea28f337d17ce9360adb3.png

You can see the ripple kind of absent, but more noisy high frequency band overall.

 

0 - 22.05 kHz sweep, both 705.6 and 768k input rate:

Mojo-sweep_705-wide.thumb.png.0d0aa6bf7b84d3dc04dd5f6e6268c4d0.png

Mojo-sweep_768-wide.thumb.png.02147de7c4572309b335c9fdb4b67813.png

 

I wonder what is the bump at 3 MHz.

 

@Miska

Do you have any measurements for Benchmark  DAC 3. It uses an FPGA to sample up to 250GHz which is not fed to the ESS 9028 but simply used to adjust clock timing to 4 picosec steps via a register shift. The time adjusted 250 GHz sample rate is then downsampled to 211KHz and fed to the DAC.  The choice of 211KHz sample rate is to keep the ESS built in linear phase filter further away from the audible band (than at 192KHz). Looks like it is Benchmark’s strategy to get what they regard as the best performance in the audible pass band region out of the ESS 9028 chip.

 

I guess what I am getting at - only a complete set of measurements and a look at passband as well as phase response is going to tell a more complete story.

 

Eliminating HF noise is a worthy goal but it might need to be balanced with other factors/priorities and the limits of analog performance of the design. (-192 dB digital stop band attenuation seems like an awful lot more than should be necessary.....but maybe I am missing something)

 

I am unable to eek out any benefit from upsampling with Roon with the DAC 3. It sounds just as good at 44.1 as it does at 192. It sounds worse if I use minimum phase filters in Roon (hardly a surprise given the damage that minimum phase does to harmonic relationships). Perhaps I should be trying HQ Player? Feel free to PM me if you feel inclined to answer but think this is to far off topic and too specific to belong in the M-Scaler thread. Thx! And thx for your contributions to these and many other threads. I certainly appreciate your input.

 

Link to comment
1 hour ago, Shadorne said:

I am unable to eek out any benefit from upsampling with Roon with the DAC 3

Right. Is not this because no matter what you do externally, the DAC 3 is always sending the ESS 9028 211 kHz?  If you cannot defeat the built in OSF of the benchmark, then you cannot hear (potential) benefits of oversampling externally...  Or can you defeat the Benchmark's OSF? 

SO/ROON/HQPe: DSD 512-Sonore opticalModuleDeluxe-Signature Rendu optical with Well Tempered Clock--DIY DSC-2 DAC with SC Pure Clock--DIY Purifi Amplifier-Focus Audio FS888 speakers-JL E 112 sub-Nordost Tyr USB, DIY EventHorizon AC cables, Iconoclast XLR & speaker cables, Synergistic Purple Fuses, Spacetime system clarifiers.  ISOAcoustics Oreas footers.                                                       

                                                                                           SONORE computer audio

Link to comment
2 hours ago, barrows said:

Right. Is not this because no matter what you do externally, the DAC 3 is always sending the ESS 9028 211 kHz?  If you cannot defeat the built in OSF of the benchmark, then you cannot hear (potential) benefits of oversampling externally...  Or can you defeat the Benchmark's OSF? 

Good point. I can feed it 44.1, 48, 88, 96 or 192 or DSD64. However the upsampling it uses must be at least as good as anything else I have tried so far. Sounds the same except when you start using minimum phase filters which seem to muck up the sound entirely (phase distortion). Everything is upsampled to 211 KHz using an FPGA Xilinx board that has been programmed by Benchmark. If I upsample to 192 then Benchmark still does the conversion to 250GHz and then back down to 211KHz. So it is running the ESS chip at what Benchmark obviously feels is the optimal rate. I cannot defeat this.

 

I am wondering if the benefits of software upsampling are limited to DACs that just use a DAC chip as it is and feed that chip the basic input sample rate and that any DAC devices with FPGA upsampling like M-scaler, DCS and others are likely using a better upsampler than you find on a cheap $70 Chip? It isn’t Benchmark’s first rodeo - they have been using FPGA upsampling with their own code for as long as Chord has. Stereophile rate it A+ recommended component along with DCS and others - it measures as good as or better than anything in Stereophile’s A+ category.

 

Sorry this is off topic but I am trying to understand if software upsampler or M-sampler is the way forward and where do they add the most benefit. What kind of DAC is needed?

Link to comment
1 hour ago, Shadorne said:

 

Sorry this is off topic but I am trying to understand if software upsampler or M-sampler is the way forward and where do they add the most benefit. What kind of DAC is needed?

Let's start with this 250 Ghz thing ... perhaps go back and check everything you've written

Custom room treatments for headphone users.

Link to comment
8 hours ago, Shadorne said:

Do you have any measurements for Benchmark  DAC 3. It uses an FPGA to sample up to 250GHz which is not fed to the ESS 9028 but simply used to adjust clock timing to 4 picosec steps via a register shift. The time adjusted 250 GHz sample rate is then downsampled to 211KHz and fed to the DAC.  The choice of 211KHz sample rate is to keep the ESS built in linear phase filter further away from the audible band (than at 192KHz). Looks like it is Benchmark’s strategy to get what they regard as the best performance in the audible pass band region out of the ESS 9028 chip.

 

No, I'm kind of bored of all the ESS DACs out there... And DAC3 doesn't support the kind of inputs I'm looking for, so it is not interesting to me.

 

250 GHz? Certainly it is not doing that. I'm not aware of any FPGA that would be capable of switching at those rates. 250 MHz sounds more like it. It could be virtual rate of ASRC. Like for example HQPlayer has virtual intermediate of 3.6 GHz when going from 44.1 kHz to 24.576 MHz in a single step. But it would be incredibly stupid to actually produce output at such rate.

 

But why 211 kHz and not for example 800 kHz?

 

Signalyst - Developer of HQPlayer

Pulse & Fidelity - Software Defined Amplifiers

Link to comment

In @Shadorne‘s defence, Benchmark did have on their webpage “Conceptual Oversampling” to 250GHz as a feature of their DAC3s just yesterday except all references to this has been removed today. What it said (since none of us can read it now), is that by oversampling to 211kHz, it is effectively upsampling to 250GHz and then downsampling to 211kHz even though it never actually goes through the upsampling computation. That’s why they called it Conceptual Oversampling.

 

My guess it’s some crazy term the marketing people cooked up. Because by that definition, you can argue that ASRC from 44.1kHz to 96kHz is effectively conceptual oversampling at 4233.6kHz (44.1 x 96). Maybe these recent posts made the Benchmark people realize how ridiculous they sound and took the reference to conceptual oversampling off. Anyway, so yes, I think @Miska guess is correct.

 

as usual, we are getting off topic so I’ll mark my own post. Haha. 

Link to comment
2 hours ago, ecwl said:

In @Shadorne‘s defence, Benchmark did have on their webpage “Conceptual Oversampling” to 250GHz as a feature of their DAC3s just yesterday except all references to this has been removed today. What it said (since none of us can read it now), is that by oversampling to 211kHz, it is effectively upsampling to 250GHz and then downsampling to 211kHz even though it never actually goes through the upsampling computation. That’s why they called it Conceptual Oversampling.

 

My guess it’s some crazy term the marketing people cooked up. Because by that definition, you can argue that ASRC from 44.1kHz to 96kHz is effectively conceptual oversampling at 4233.6kHz (44.1 x 96). Maybe these recent posts made the Benchmark people realize how ridiculous they sound and took the reference to conceptual oversampling off. Anyway, so yes, I think @Miska guess is correct.

 

as usual, we are getting off topic so I’ll mark my own post. Haha. 

I don’t know about the marketing. It sounded like they just padded with zeros up to 250 GHz in order to allow small timing adjustments at 4 picosec (by simple register shift at 250 GHz). I think it was Siau who wrote the article so it was more science based than marketing. The term conceptual is because they don’t feed it to the DAC - this oversamped 250GHz data stays in the digital domain - it an intermediate step - after register shift for timing adjustment  the 250 GHz is down sampled to 211KHz before being fed to the ESS chip. The term conceptual applies because it is digital stream of data that never gets converted into the analog real world - so calling it “GHz” is a conceptual term as the data is never clocked at that sample rate.

 

@Miska

 

211KHz was chosen as the optimal rate for the ESS chip. They did measurements on it and found what worked best. It slightly defeats the built in brick wall filters on the chip by pushing the filter corner up slightly above the audible band but still maintains adequate stop band.

 

250GHz is clearly “conceptual”. Mathematically correct but computationally no devices are operating at that speed.

Link to comment
2 hours ago, Shadorne said:

I don’t know about the marketing. It sounded like they just padded with zeros up to 250 GHz in order to allow small timing adjustments at 4 picosec (by simple register shift at 250 GHz).

 

 

This is whacky and off topic. The only circuits that operate at these speeds are research. There is no consumer logic that operates at 250 GHz 

Custom room treatments for headphone users.

Link to comment
23 hours ago, Miska said:

 

No, I'm kind of bored of all the ESS DACs out there... And DAC3 doesn't support the kind of inputs I'm looking for, so it is not interesting to me.

 

250 GHz? Certainly it is not doing that. I'm not aware of any FPGA that would be capable of switching at those rates. 250 MHz sounds more like it. It could be virtual rate of ASRC. Like for example HQPlayer has virtual intermediate of 3.6 GHz when going from 44.1 kHz to 24.576 MHz in a single step. But it would be incredibly stupid to actually produce output at such rate.

 

But why 211 kHz and not for example 800 kHz?

 

 

Here is what Benchmark claim,

 

Quote

 

CONCEPTUAL OVERSAMPLING

The digital filters in the DAC2 operate at a conceptual sample rate of about 250 GHz. Incoming audio is conceptually upsampled to 250 GHz and then down sampled to 211 kHz using a filter that mathematically behaves as if it is operating at a 250 Giga-sample-per-second rate. We use the word "conceptual" because the calculations and internal clocks are not actually running at 250 GHz. Due to the mathematics of upsampling, most of the filter calculations require a multiply by zero operation. These unnecessary zero-product calculations are eliminated while all of the non-zero calculations are executed. The net result is mathematically equal to the results that would have been produced by executing every calculation at a 250 GHz sample rate. Eliminating the unnecessary calculations reduces the DSP and processing rates to a manageable load.

 

end quote

 

It has been a while since I did all the physics and mathematics related to time series analysis, convolution deconvolution etc. but their claim sounds totally plausible to me. Mathematically it is correct and could be easily implemented.

Link to comment
3 hours ago, Miska said:

Yeah, there's nothing special or extraordinary about that. I don't know why they bother to write about such. And why they invented term "conceptual oversampling" for something others call "polyphase filter".

Because then they can appear to have something others do not. My respect for Benchmark just dropped by half.

Link to comment
5 hours ago, Miska said:

 

Yeah, there's nothing special or extraordinary about that. I don't know why they bother to write about such. And why they invented term "conceptual oversampling" for something others call "polyphase filter".

 

 

Wow. A bit condescending. First you say oversampling to 250 GHz is impossible and now it is nothing special or extraordinary. So which is it? My BS detector has gone off - you talk like a leading authority on D to A conversion  (bored by ESS chips) but I think I can see behind the mask someone is struggling. I notice that almost all of your posts tend to have a marketing angle intended to promote what you hawk. Perhaps Benchmark are simply guilty of the same - hardly surprising. Kettle calling the pot black me thinks?

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...