Chord Electronics Owners & Discussion Thread

[davewantsmoore]

1 hour ago, Music2496 said:

Not at all, you've jumped the gun on that one there. I've discussed some of the maths directly with him

So you don't think RW disagrees with what is in the article???

 

   (ie. that using "sinc" is "good", if it were possible .... and that a computer can use a much better approximation of sinc, than a more limited device can).

 

That is what he says in the video ..... "I made up a really close approximation to sinc" .....   I can get my computer to compute audio using a much closer approximation to sinc, I could leave it running overnight if I wanted.     Given that my computer could be 100 million times faster, and I am leaving it 50 thousand times longer ...... that's ummm, a lot.

 

There's of course a lot of other issues.  Selection of bandstop filter....  how you actually go about approximating sinc, etc.

 

1 hour ago, Sime V2 said:

Pfft, I’ll be more original than that ………………………………

If you used this:

 

 

People might mistake it for a middle finger.

[rand129678]

33 minutes ago, davewantsmoore said:

So you don't think RW disagrees with what is in the article???

Of course not. Never said he disagrees with anything technical there. Of course he agrees with a lot of the maths related content of the article!  Except he obviously doesn't want you to use software, he wants you to use an M-Scaler, for obvious reasons...

 

The point that appears funny to me, is from a marketing prospecting, to have a link to an article talking about software upsampling being better than M-Scaler, on Rob's M-Scaler page, when we all know Rob is so vocal (in writing) against using software up-sampling with his DACs (again for obvious commercial reasons)! I must be the only one that gets a chuckle out of it.

You're getting way too deep into technical arguments, when I'm just chuckling about the high level optics - nothing more than just a chuckle. From a technical perspective (for the fourth time now) I agree with a lot of the article  and haven't disagreed with anything technical you have said...

 

If you haven't noticed, none of my comments have anything to do with technical aspects of the article... and again, I use HQPlayer (the really loooong linear phase super fast roll-off Dave-like poly-sinc filter too) !

 

Respectfully, it feels like banging my head against a brickwall (hehe), just trying to explain something that made me chuckle. Nothing more than that. Nothing serious in a technical sense.

 

I'd like to tap out (1M times) again please lol.

 

 

 

 

Edited January 24, 2019 by Music2496

[Ittaku]

41 minutes ago, sandrews888 said:

However, in theory if we send the signal from the Daves digital outs (bnc) into the deqxs digital in, in theory, would this sound better with Tikandis vs analog in? Also if we used a mscaler would the deqx downscale the upscaled 768khz signal back down to 24/96? This might be an experiment we should try when I bring the Dave down as the Dave has digital outs whereas the Qutest doesn't. Food for thought

The deqx won't be able to take 768 in though. I'm not sure what its limits are, but the 768 used between Chord components uses their proprietary dual coaxial mechanism. I've only ever seen single coax support up to 384, and many devices don't accept past 192. Pretty much any device out there that supports higher is using their own proprietary communication mechanism, or are doing it over network connections and relying on local rendering. Someone with a deqx could chime in and say what resolution it supports. My DSP only supports up to 192 in.

[aechmea]

768k has been mentioned.  It is too much for the DEQX digital inputs and won't work.  192k is max input, which would then be resampled to 96k anyway.

 

That's why any M-scaling and DAC would seem (to me) to be best positioned after the DEQX.

[davewantsmoore]

 

39 minutes ago, Music2496 said:

The point that appears funny to me, is from a marketing prospecting, to have a link to an article talking about software upsampling being better than M-Scaler, on Rob's M-Scaler page, when we all know Rob is so vocal (in writing) against using software up-sampling with his DACs (again for obvious commercial reasons)! I must be the only one that gets a chuckle out of it.

 

Right...  I'm not sure it's "funny" ....  but it's a good example of the marketing department not being so out of touch with reality as they sometimes are.

 

You can't very well have a premise which says "get really close to sinc" .....  and then deny that ways to get even closer to sinc than you have, would not also be good.

 

Of course, brute forcing sinc requires long(er) delays, and a computer, and/or offline processing ....  this doesn't sound (at all) like it would eat Chord's target market ..... so I don't really think they'd be concerned about the admission.

 

39 minutes ago, Music2496 said:

You're getting way too deep into technical arguments

It was only the idea that RW wouldn't agree that using a computer could closer approximate sinc  (ie. that he would disagree with what was in the article).     Given that this is objectively true..... then he would obviously admit that (perhaps, with some relevant caveats).    That's all I was saying.

 

39 minutes ago, Music2496 said:

and haven't disagreed with anything technical you have said...

Yes, I know.... You've just said that RW wouldn't agree.

 

He would, because it's just a plain fact...... he'd also give some caveats about why an M scaler would be a much more practical and reliable choise, of course.

39 minutes ago, Music2496 said:

If you haven't noticed, none of my comments have anything to do with technical aspects of the article

Yes, I did notice.    What makes you think I didn't !?

[bhobba]

5 hours ago, Sime V2 said:

And let’s face it, MQA does not have a future imo. 

We need to see how this plays out.   As a standard for transmitting 96k (bit stacking etc) that's backward compatible with 48k it has a future IMHO.

 

Thanks

Bill

 

[rand129678]

1 hour ago, bhobba said:

We need to see how this plays out.   As a standard for transmitting 96k (bit stacking etc) that's backward compatible with 48k it has a future IMHO.

 

Thanks

Bill

 

I don't mind the 1st unfold (up to 96kHz) for Tidal streaming. 

 

Regarding the 1st unfold only, from one of MQA's biggest critics, Archimago:

 

"Objectively with the songs I examined, the software decoder works well to reconstruct what looks like the equivalent 24/96 download."

 

and

 

"Bottom line: TIDAL/MQA streaming does sound like the equivalent 24/96 downloads based on what I have heard and the test results"

 

https://archimago.blogspot.hk/2017/01/comparison-tidal-mqa-music-high.html

 

Again, this applies to the 1st unfold only (up to 96kHz)… he’s done plenty of analysis on the stuff after the 1st unfold, which doesn’t need repeating here of course.

 

 

Edited January 24, 2019 by Music2496

[agisthos]

Dave if you want to constantly bang on about how RW is wrong, how any PC can do his upscaling, how its nothing special, how its just marketing, and on and on, why don't you go start another thread for that and leave these guys in peace to discuss their M-Scaler?

[davewantsmoore]

5 hours ago, agisthos said:

Dave if you want to constantly bang on about how RW is wrong

'cos it's directly relevant to the M Scaler "story" of "1 million taps"     Chord have an article published on their website.   People are saying RW himself would disagree with the facts of the article.....   and I do not think he would (evidenced by the direct quotes from him supporting said facts).

 

Anyways, I think we're finished now.... as it seems some positions have been relaxed.

[sir sanders zingmore]

23 hours ago, bhobba said:

The MQA guys do not believe small amounts of aliasing are audible, Rob Watts believes it is very audible. 

Should be pretty simple to determine who is right, right?

[rand129678]

19 minutes ago, davewantsmoore said:

People are saying RW himself would disagree with the facts of the article..... 

Nobody has said that anywhere...

 

If you really think Rob Watts will agree with the following words in that article (which is what I said - never commented about the theory discussed in the article...), then I really can’t help you.

 

“in which respect it’s even better than the M Scaler."

 

This is a really painful back and forth. Apologies to all that I engaged lol.

 

No more from me on it. Back to M-Scaler.

Edited January 24, 2019 by Music2496

[davewantsmoore]

7 minutes ago, Music2496 said:

“in which respect it’s even better than the M Scaler."

 

".... but, if you put that digital data into something that's called a sinc function filter, which has an infinite amount of processing, it will perfectly reconstruct the original"

 

Then he goes on to discuss how his algorithm goes about an approximation of sinc, and how the degree to which you do that, has a very big impact on sound quality.

 

ie.  He is saying that .....  using a computer to orders of magnitude better approximate sinc ....  does a better job than what is used in the M scaler  (with the previously mentioned caveats).

 

 

[rand129678]

On 06/01/2019 at 10:19 PM, Sime V2 said:

Now that’s some shielding

 

Nice! Hadn’t seen the internals before.

 

A heap of metal for both FPGA heat dissipation and shielding (as it's designed to be close to a DAC and other analogue components) I assume.

 

Edited January 24, 2019 by Music2496

[bhobba]

35 minutes ago, Sir Sanders Zingmore said:

Should be pretty simple to determine who is right, right?

Of course not.  What people prefer is very individual.

 

Thanks

Bill

 

 

[Ittaku]

4 minutes ago, bhobba said:

Of course not.  What people prefer is very individual.

In this case it is, if done correctly. The assertion is whether aliasing is audible or not. If two virtually identical filters are used in a comparison, and one allows aliasing whilst the other doesn't, then the comparison is valid, provided it's properly tested with blind testing yadayada... If they're indistinguishable, then it's inaudible. There's no comparison about which sounds better here. Comparing MQA to Chord's upscaling is comparing thousands of other things as well.

Edited January 24, 2019 by Ittaku

[sir sanders zingmore]

3 minutes ago, bhobba said:

Of course not.  What people prefer is very individual.

 

Thanks

Bill

 

 

Who said anything about preference?

Your comments referred to audibility 

[bhobba]

18 minutes ago, davewantsmoore said:

He is saying that .....  using a computer to orders of magnitude better approximate sinc ....  does a better job than what is used in the M scaler  (with the previously mentioned caveats).

 

The M-Scaler is a computer - the most powerful currently available on a single chip with 74 cores.   It is way beyond what can be done on a PC or Mac.

 

What he is saying is he has done calculations that shows to perfectly reconstruct a sync (ie brick-wall) filtered signal to 16 bit accuracy you need 1 millon taps.  What shocked his ears was that when doing that it made a big jump in sound quality.

 

Thanks

Bill

[bhobba]

2 minutes ago, Sir Sanders Zingmore said:

Who said anything about preference?

Your comments referred to audibility 

And you think what is audibly better does not vary from person to person?  I like MQA, but others do not.  There is no right or wrong.  I own an M-Scaler but my TT2 DAC had a fault that needs to be fixed before I can check out how it compares to MQA.  Right now however I have broken my Distal Femur and am to a large extent bed ridden so can not do it at the moment - I am reliant on others in the mean time.  Unfortunately those others all have other things to do so checking it out will proceed slowly.  Distal Femur breaks take a year or more to heal especially mine which was severely broken.

 

Thanks

Bill

[sir sanders zingmore]

6 minutes ago, bhobba said:

And you think what is audibly better does not vary from person to person?

Who said anything about “ better”?

your comments referred to audibility. 

 

Sorry to hear about your break. Sounds painful.

[legend]

19 hours ago, sandrews888 said:

 Hi Rod @legend

We will be able to answer this questions with our ears when I get to your place in a few weeks. However, in theory if we send the signal from the Daves digital outs (bnc) into the deqxs digital in, in theory, would this sound better with Tikandis vs analog in? Also if we used a mscaler would the deqx downscale the upscaled 768khz signal back down to 24/96? This might be an experiment we should try when I bring the Dave down as the Dave has digital outs whereas the Qutest doesn't. Food for thought

DEQX only processes signals up to 192k - though from memory I have played some 384 files which it accepts, presumably downsamples to 192?  And its dual Sharc processor chips can 'only' handle 96k  - there is a lot of DSP processing to do, with digital Xovers, driver correction. room correction etc - so it downsamples everything to 96k before upsampling the DSP output again to 192k for the DACs,

 

There also may be an aural difference between the SPDIF inputs (one of which is BNC) and the USB input on the DEQX,

 

Slightly off topic (though related to timing effects) I have just used DEQX to correct my passive loudspeakers for time alignment (and frequency linearity) at all frequencies above 100 Hz (limited by non-anechoic measurements)  - measuring the speaker as though it was a single driver.  The result was a sound much more like that from the Qutest (without M-Scaler!) - perhaps the Qutest was slightly more laid back but both had similar detail, PRAT, imaging etc, -  more like just changing the contrast button slightly on a TV.

Edited January 25, 2019 by legend

[bhobba]

1 hour ago, Ittaku said:

In this case it is, if done correctly. The assertion is whether aliasing is audible or not.

That's not the issue ie if you can tell a difference or not.  The issue is, is it worrisome to the listener or not - there is a difference.  I do not know if the reason is aliasing or not but some like MQA and some do not.  But knowing how MQA works my guess is its associated with the filters used, and hence aliasing comes into it.   In MQA when encoded the best filter to down-sample it to 96k is decided from analysing the music.  At the other end the best filter to upsample it to its original sample rate is decided based on the down-sampling filter used, which is watermarked in the MQA stream. That would not be done unless the differences in filters were audible.  In fact it is claimed that part is getting better and better:

https://www.stereophile.com/content/mqa-tested-part-2-fold-manufacturers-comment

 

Added later:

After reading some of the replies the issue may be use of the word audible.  Differences in filters clearly are and I believe that has been demonstrated many times.   The issue is separating aliasing from temporal smear to identify which part is creating the audible difference.  But wheter Rob or the MQA people are right about the cause of the difference its personal preference that will decide what each listener likes.

 

Thanks

Bill

Edited January 25, 2019 by bhobba

[sir sanders zingmore]

On 24/01/2019 at 10:25 AM, bhobba said:

The MQA guys do not believe small amounts of aliasing are audible, Rob Watts believes it is very audible

 

5 minutes ago, bhobba said:

That's not the assertion.  The issue is not if its audible - the issue is, is it worrisome to the listener or not - there is a difference

 

Your two statements above seem contradictory to me

[bhobba]

27 minutes ago, Sir Sanders Zingmore said:

your comments referred to audibility. 

And you don't think audibility varies from person to person?  As mentioned in another post the differences in the filters MQA uses is audible - at least they claim it is.  But how people react to it ie if its worrisome is individual preference.  John Darko has posted it sounds better to him (as it does to me), but some very experienced audiophiles I know do not like it describing as having a metallic sheen etc.  Who is right and who is wrong - there is no correct answer.

 

Thanks

Bill

[bhobba]

8 minutes ago, Sir Sanders Zingmore said:

Your two statements above seem contradictory to me

There is no contradiction.  There are two different views and only hearing it can decide.   And once human listeners become involved human preference comes into it.  But I will be more concise.   MQA chooses a filter to down-sample and upsample it later - there are audible difference between the filters.   MQA people claim the aliasing is not audible - Rob claims it is.  Exactly how do you propose to check what aspect of the filter creates the audible difference bearing in mind there is a trade-off - increase aliasing - decrease temporal smear - how to separate the two - beats me.

 

Thanks

Bill

[Ittaku]

43 minutes ago, bhobba said:

That's not the assertion. 

Sorry then, I took your comment on face value alone. There were no other qualifiers in your original statement about aliasing. The rest of the debate is relevant sure, but audible or not is a pure binary assertion which isn't what you meant. The test for audibility of just that one aspect IS doable, as I said, but that's not what you're after after all.

Edited January 25, 2019 by Ittaku