Chords New M -Scaler

[barrows]

I think it would be very interesting if Chord decided to sell a software MScaler package.  This could be, for example, sold as a plug in to ROON and/or Audirvana Plus.  I believe, as a software package it could be copyrighted, and would likely provide a nice additional income stream for Chord.  The customer for the hardware based MScaler would be very different than the one which would want the software version, and I doubt that they would lose sales of the hardware version.

This would make comparisons with other oversampling approaches easier and more interesting.  But there is still the fact that as far as i can tell from Rob Watts' explanations, the Chord oversampling approach is highly optimized to work best with the Chord DAC approach (Pulse Array).  

The DAC still really matters, of course.  It is cool that Chord has made MScaler oversampling available at a lower price point, after all, HQPlayer plus a computer capable of using its full capabilities is still more affordable than MScaler. 

[mansr]

9 minutes ago, adamdea said:

Could you expand on your thoughts on the best choice. -assuming we are sticking with linear phase is this just a question of the right windowing function or do you mean that we shouldn’t bother with an actual sinc?  Am I right in thinking that you can make something with more attenuation for a given filter length by not using a windowed sinc?

A windowed sinc function approximates an ideal low-pass filter. A sufficiently close approximation will of course provide a reconstruction that is accurate to whatever precision you desire. However, with limited resources (and that is always the case), there may well be a different transfer function producing a result that is better in the ways that actually matter. Take CD audio with 44.1 kHz sample rate. If we accept that the upper limit of human hearing is 20 kHz, the goal of the interpolation filter should be to preserve lower frequencies as accurately as possible while blocking images above the Nyquist frequency. To do this, we don't actually need a perfect low-pass filter since there's a 2 kHz margin between the upper limit we care about (20 kHz) and the Nyquist frequency (22.05 kHz). It's not a lot, but it does provide for some flexibility in the filter design. If some steepness can be traded for reduced passband ripple or improved stopband attenuation, that is a good thing. Higher sample rates allow filters with wider transition bands, further increasing the options. Of course, this all hinges on the premise that frequencies above some limit are unimportant to human hearing. If you refuse to believe this, you may as well give up on reproduced audio entirely.

 

9 minutes ago, adamdea said:

there is a plausible sounding theory out there that (to be on the safe side)  linear phase filters should not be longer than something to do with the ears’ filter bins possibly the time constant. I think that Jj over at HA and Fokus have both mentioned this point, but I have never quite grasped it. 

That sounds a little sketchy. A lot of plausible-sounding ideas turn out to be complete BS on closer examination. I'm not familiar with this one, so I can't say for sure, but I am sceptical.

[mansr]

10 minutes ago, barrows said:

This would make comparisons with other oversampling approaches easier and more interesting.

Perhaps they don't want that.

[barrows]

19 minutes ago, mansr said:

Perhaps they don't want that.

I do not think it would really matter in the end from a sales perspective.  Those who are going to prefer their Chord gear will still do so.  I have worked in the high end for long enough to know that manufacturers are not sitting around and scheming about ways to dupe people, manufacturers are trying really hard to improve system performances and sound quality, and to do so also by making products which can distinguish themselves int he market place: that Chord has a somewhat unique approach which accomplishes that.

[rayl1234]

35 minutes ago, barrows said:

I do not think it would really matter in the end from a sales perspective.  Those who are going to prefer their Chord gear will still do so.  I have worked in the high end for long enough to know that manufacturers are not sitting around and scheming about ways to dupe people, manufacturers are trying really hard to improve system performances and sound quality, and to do so also by making products which can distinguish themselves int he market place: that Chord has a somewhat unique approach which accomplishes that.

 

I am the innocent type of person... the more innocent possible explanations I see are:

 

1. RW is a silicon guy.  Software is not his expertise. Chord can't step in bec Chord does not own RW's designs.  Chord manufactures/sells them and pays RW a royalty, but as has been publicly disclosed, the arrangement is subject to unilateral termination for convenience by either party....

 

2. There is more to the story than upsampling though, as replacing traditional DACs with a hybrid DAC/amp combination (DX amps in Chord speak) is very much the product direction for RW, so, together with Davina ADC project, creating a software version would be a huge distraction that does not fit the roadmap.

[Jud]

Is another reason not to produce a version tied to hardware that the pure software version may be easier to copy, or nothing much to choose between the two?

[Jud]

6 hours ago, ecwl said:

Just like nobody who charges for their software upsampler so far has been willing to open source their code so that we can view them.

 

The SoX upsampling and modulators available in Audirvana Plus are AFAIK those available open source in SoX. 

[The Computer Audiophile]

9 minutes ago, Jud said:

Is another reason not to produce a version tied to hardware that the pure software version may be easier to copy, or nothing much to choose between the two?

Software support for a HiFi company is a nightmare. Ask every one of them. 

[ecwl]

7 minutes ago, Jud said:

 

The SoX upsampling and modulators available in Audirvana Plus are AFAIK those available open source in SoX. 

Great. And I think @mansr is a SoX maintainer so I was hoping that he would provide a recommended setting that he thinks would approximate or beat the Hugo M-Scaler/Chord Blu2/Mojo. I'm downloading the Audirvana plus trial version now as I type this... Would also be interesting to see if iZotope 64-bit SRC is also open source.

[ecwl]

Hmmm... Running Audirvana+ is interesting. I only used Mojo to compare. As it was that obvious. And yes, HQPlayer is clearly better than Audirvana+ SoX or iZoTope 64-bit SRC.

 

First of all, when running SoX, the filter max length was 30,000 samples. Does that mean if I"m doing 1fs to 8fs, I'm getting 8 x 30,000 so 240,000 taps? Not sure. Anyway, on the maximum setting of 99.5% Nyquist, 30,000 samples, antialiasing 100 and pre-ringing 100 (linear phase), Mojo sounds way better without upsampling. IZoTope 64-bit SRC sounds slightly better than SoX. Once again for 1fs to 8fs, Steepness 200dB, filter max length 2 million, cutoff freq 1x Nyquist, anti-aliasing 200 and pre-ringing 100 (linear phase), Mojo still sounds clearly better to me. So I didn't even bother hooking my Blu2 M-Scaler to Mojo/DAVE to compare, whereas I did with HQPlayer in Roon.

 

For example, with Dave Brubeck's Unsquared Dance, the piano timbre simply sounds unnatural and the piano key strikes sound soft when it's not supposed to with SoX/iZoTope compared to Mojo. The hand clapping in the track sounds much more realistic with Mojo than iZoTope which is a bit better than SoX. When playing Janelle Monae's Pynk, the finger snapping doesn't sound like snapping of the fingers at all with SoX or iZoTope whereas they do with Mojo.

 

I guess of the software filters I have listened to so far, HQPlayer is the way to go. But I am having trouble finding 1fs to 8fs/16fs software upsampler that sound better than Mojo's WTA filter to me. That said, I totally get that I'm not even using HQPlayer as it's intended to be since HQPlayer performs best with 1fs to DSD512 into a DSD DAC. So apples vs oranges.

 

[Jud]

2 hours ago, ecwl said:

Great. And I think @mansr is a SoX maintainer so I was hoping that he would provide a recommended setting that he thinks would approximate or beat the Hugo M-Scaler/Chord Blu2/Mojo. I'm downloading the Audirvana plus trial version now as I type this... Would also be interesting to see if iZotope 64-bit SRC is also open source.

 

iZotope is not open source, though there is a free trial.

 

Do you hear a difference among 10,000 taps, 100,000 taps, and 2 million taps with iZotope in Audirvana Plus?

[ecwl]

15 minutes ago, Jud said:

Do you hear a difference among 10,000 taps, 100,000 taps, and 2 million taps with iZotope in Audirvana Plus?

Did not try it. Just went for 2 million with IZotope because my time is precious. 

 

I do hear a difference between 1 million taps from Blu2 vs DAVE’s 164,000 taps.

[Adyc]

4 hours ago, The Computer Audiophile said:

Software support for a HiFi company is a nightmare. Ask every one of them. 

 

I totally agree when one consider multi platforms support. There are myriad computer configurations.  Chord is a very small company. They really struggle to meet the demand. I only got my Blu2 after one year I placed the order. I consider the demand of MScaler is even much higher. There is really no reason for Chord to produce any software version of MScaler. 

 

I don’t think it is true that software + computer is cheaper than MScaler. Consider one may need to address the noise problems from computers. Adding all the costs, Mscaler maybe cheaper and it definitely looks much better. 

[mav52]

12 hours ago, ecwl said:

Did not try it. Just went for 2 million with IZotope because my time is precious. 

 

I do hear a difference between 1 million taps from Blu2 vs DAVE’s 164,000 taps.

 

So there is another software choice besides HQP with  IZotope, Audirvana Plus any more ?

[ecwl]

1 minute ago, mav52 said:

 

So there is another software choice besides HQP with  IZotope, Audirvana Plus any more ?

I didn’t try iZotope but Audirvana Plus has an SoX filter and an iZotope filter.

i believe xxHighEnd can also upsample with lots of computational prowess but I did not try it because it looks a bit daunting but I think I’ve reached a point where I no longer feel it is my onus to try to find a software filter that beats M-Scaler. But if a Chord DAC owner can identify a 1fs to 16fs software filter that improves on the Chord DAC’s own WTA filter, I’d be more than happy to give it a listen for myself and to compare that to the M-Scaler in Blu2.

[Whitigir]

Does anyone use Audirvana Plus for Win10 64 bits ? Please show me how to do Izotope filter ?

[ElviaCaprice]

5 hours ago, ecwl said:

 I think I’ve reached a point where I no longer feel it is my onus to try to find a software filter that beats M-Scaler. But if a Chord DAC owner can identify a 1fs to 16fs software filter that improves on the Chord DAC’s own WTA filter, I’d be more than happy to give it a listen for myself and to compare that to the M-Scaler in Blu2.

 

I said about the same thing.  There won't be any because there isn't any software solution that works even close to Chord's hardware.  But plenty of FUD.

[Whitigir]

27 minutes ago, ElviaCaprice said:

 

I said about the same thing.  There won't be any because there isn't any software solution that works even close to Chord's hardware.  But plenty of FUD.

And how would you know ? Guessing ? Speculating ? Biasing ?

[rayl1234]

5 minutes ago, Whitigir said:

And how would you know ? Guessing ? Speculating ? Biasing ?

 

I will readily concur with the "isn't" (present tense) portion based on having spent an intense 1.5 months following RMAF 2017 trying almost everything under the sun, both software and hardware (up to around USD 30k), before pulling the trigger on swapping out the Direct Stream / Huron for a Chord stack.

 

I wish we could hear a comparison -- instead of arguing over software vs hardware -- to what might potentially be a real contender and is shipping (reports of some delays due to metal work notwithstanding) today: namely the MSBs.... At the time, there was only the Select 2, which at USD 90k to start, I didn't bother to evaluate.  But now there are the Reference and the Premier, the latter quite within in the price range.  Honestly, it is the only competitor on my radar for the next upgrade cycle (say in 5-7 yrs) at this time...

[Whitigir]

Well, there is no Software VS hardware.

 

They are both parts of your devices.  Your PC can’t run without an OS.  The same as MScaler.  People are interested because for whatever Chord is saying on the MScaler, it look achievable by a simple software on a PC (without using MScaler).

 

the thing is that, we don’t know exactly what WAT filters are.  But from whatever being told, it is a perfected Filters for upscaling as it would even down convert Native Studio DSD into PCM to get processed by Chord Own DAC.

 

it is another saying is that, you all suck! Only Chord knows how to properly convert digital sound to analog sound in a perfect way

 

Yes, some experts has already being saying that WAT filters is just another term By Chord for Linear Filters.  Anyways, no one knows as no one can look into those sources codes and figure out what it exactly is yet.

 

It is just wrong to say FPGA can do something that a PC can’t

[Miska]

On 8/4/2018 at 5:14 PM, mansr said:

The point is that it's just another windowed sinc (JAWS) filter. It's not even clear that this is the best choice for audio applications.

Last time I checked they had just 120 dB stop-band attenuation, which in my opinion is not enough, or in other words I don't see reason why not do more. That's why I use 192 dB so that it matches at least 32-bit resolution... But anyway, I think there are so many aspects in a filter design, that single parameter like number of taps is pointless without all the other information...

 

[Miska]

On 8/4/2018 at 5:16 PM, ElviaCaprice said:

Which only confirms the findings of Romaz and others that have tried with DAVE using HQP.

 

You cannot make best out of HQP with something like DAVE, especially if you feed it DSD. Not that it would be necessarily bad combination, just not the best one either. I don't think DAVE can operate as a bit-perfect DAC...

 

[mansr]

24 minutes ago, Whitigir said:

Yes, some experts has already being saying that WAT filters is just another term By Chord for Linear Filters.  Anyways, no one knows as no one can look into those sources codes and figure out what it exactly is yet.

If you send me one, I'll figure it out.

[rayl1234]

1 minute ago, Whitigir said:

Well, there is no Software VS hardware.

 

They are both parts of your devices.  Your PC can’t run without an OS.  The same as MScaler.  People are interested because for whatever Chord is saying on the MScaler, it look achievable by a simple software on a PC (without using MScaler).

 

the thing is that, we don’t know exactly what WAT filters are.  But from whatever being told, it is a perfected Filters for upscaling as it would even down convert Native Studio DSD into PCM to get processed by Chord Own DAC.

 

it is another saying is that, you all suck! Only Chord knows how to properly convert digital sound to analog sound in a perfect way

 

 

I believe many have already conceded that it is 100% achievable on a PC with some amount of either CPU or GPU resources.  I certainly believe so.  Just that there is not yet an off-the-shelf solution for doing so that has tested to my satisfaction, and this includes being able to work with streamed audio content to a Windows web browser as a requirement for me.  I stand by that statement.

 

I am not here to defend Chord per se -- I am always looking for something better.  Software doing the same thing -- it might be cheaper -- but I am looking for something better. 

 

I feel this thread has turned into one of attacking Chord bec it is "just" filtering, which everyone "knows" can be done in software... This seems to have touched upon a nerve.  "What audacity to claim an improvement if it can be 'simply' done in software?" Is that the rationale?

 

Isn't the FPGA code also software?  Of course it is. Is it that some are offended that "software" costs USD 4.5k?  Heck, I paid USD 9.5k for the same "software" 7 months ago bec it works well, it is packaged, and it is supported.

 

Chord/RW have every right to be proud of their product -- why is that suddenly interpreted as a big FU to the rest of the world instead of simply being creators' pride?  Would we believe a product is better if its creators were NOT proud to speak highly of their product????  That's crazy. Or is the only acceptable way to express pride to provide full disclosure?  I don't see GitHub repos for any of the software packages being discussed here.... So why the double standard? 

 

 

[Whitigir]

Everybody has the right to be proud of their products.  However, the way the M-Scaler is doing and RW marketing, it offend other people, including the engineers, the artists, the studios, and even enthusiast 

 

see ? SACD from a Native Studio meant that it is a perfected tracks in everything that the studio want the listener to listen to, without tinkering and alternating the sound signatures.  That is what Native DSD is all about

 

Now, let’s not get into why DSD is better or worse than PCM.  I enjoy both under different circumstances.  However, if you claim that you rather sell people equipments that Down Convert Native DSD from a Studio to PCM regardless of rate, because you believe that your other equipments will do a more accurate jobs ? 

 

Are you kidding me ? Do you know how much a Studio invest in their resources ? To be outdone by your mere Chord Stack ? Give me a break

 

I am not flaming on Chord, but I am not a blind sheep either.  I used to respect Chord, but losing it every day.

 

I wouldn’t feel so bad if Chord simply states “by using M-Scaler, the consumers can have their own Chord house sound even from Native DSD tracks”.  This statements is different than “recovering the most accurate Informations at ADC stages and reconstruct it more accurately”.  Huge different