MQA pt II

[mansr]

1) Yes, the defocussing allows one to go sub-pixel. Yes, the pixels have a finite area, ideally as much as possible, since their job is to catch as many photons as they can. And yes, the 'missing pulses' problem is no problem provided anti-aliasing filtering is in place. BS then has this idea that you can omit anti-aliasing filtering, and sample with a triangular or spline or Mickey Mouse kernel (which is still boils down to standard, instantaneous sampling followed with a crude filter).

The star field can be treated as a collection of point sources, making it sort of a 2D equivalent of a series of Dirac pulses. The imaging you described is an example of sparse sampling made possible by constraints other than the (spatial) bandwidth, in this case the density of stars. You obviously know all this. Since music isn't made up of Dirac pulses (or anything remotely similar), no useful analogy can be made here.

 

[Mr_Sukebe]

More pages of guff, and yet the comments I’ve read by MQA users were typically very positive. Funny that

 

[Jim Audiomisc]

In principle, given sufficient accuracy about the telescope antenna pattern and a high enough SNR you can also get some idea of the actual angular width of a star. However this gets to be hard work in most practical cases. That's why Hanbury-Brown dreamed up his interferometer... which I think many 'experts' said couldn't work... until it did.

The obverse of the above is more useful. e.g. for the work that went into my sickening thud I use Jupiter to measure the antenna pattern of UKIRT. Needed to correct for the actual angular size of Jupiter as seen from Mauna Kea as it isn't quite a point source, just much smaller than a beamwidth.

Yes, these things matter because they do lead into the reasons why assuming impulse functions of ADCs must never pre-ring is, erm, dubious. And most telescopes have a nicer pattern than many microphones! 8-]

The ADC does not *create* the information, it just converts its format. (Ideally)

It's understandable that most audiophiles won't have any interest in this sort of thing. But for those working on digital audio it *is* important, and not knowing can lead to mistakes.

FWIW I also tried many years ago to find out details of disc cutting. Little info was available, but what I did find was quite remarkable, and explained a few things.

 

[Werner]

Since music isn't made up of Dirac pulses (or anything remotely similar), no useful analogy can be made here.

I disagree. The analogy is valid in that it shows that low-bandwidth sampling can yield a very high location accuracy, contrary to intuition.

 

[mansr]

I disagree. The analogy is valid in that it shows that low-bandwidth sampling can yield a very high location accuracy, contrary to intuition.

Only because other constraints are met, which isn't the case for audio signals.

 

[Werner]

Some very minimal music might be similar, though

 

[mansr]

Some very minimal music might be similar, though

Steve Reich's Music for Pieces of Wood, maybe. And the 4'33", but that's cheating.

 

[Werner]

And the 4'33", but that's cheating.

After 18 years of cumulative radiation damage, or during a solar storm, you are far from the average performance of 4'33".

At that time the focal plane array is filled with junk. Junk from which the stars are to be dredged. But this is OT.

 

[Jim Audiomisc]

Only because other constraints are met, which isn't the case for audio signals.

I'm not clear what you have in mind above. They key requirement is that the input to the ADC has to be fully constrained to Nyquist and isn't clipping the linear system.

Once that is done the accuracy of the ability to specify features like inter-sample peaks, say, depends on the available SNR/resolution for the pattern of information sampled.

Similarly, you need to filter, etc anyway to avoid unwanted distortions.

Is your concern simply about Werner's inclusion of Dirac? if so, the filtering requirement presumably covers that.

 

[boneman]

I can't claim to know anything really about MQA other than the basic premise and others may already have said this but it seems to me it's all in the name: Master Quality "Authenticated" whichs says to me DRM. Authenticated according to who? To me this is nothing more than a way to control access under the guise of giving you 'the best possible sound'. You then pay for it twice. Once when you buy the 'MQA' capable DAC and again when you but the 'MQA' encoded file to play.

 

[Werner]

Mans' objection is that the starfield has the known property of existing of relatively few impulse-like bright points on a background of blackness, contrary to music, which is a signal that can move near-anywhere at any time. He has a point (pun intended), but in reality it is not that black and white (pun also intended). But again, this is OT and not really contributing to the discussion beyond the message that seemingly crude techniques can and do provide highly-accurate results.


(Oh, I got #1000 and #1111!)

 

[UliBru]

Genuine question. I have seen demonstrations of the pre-ringing being removed, but not the other way round. My intuitions produce contradictory answers but I know that intuition is not a good guide. Is the answer the same for an all pass filter as it is for a band limiting filter?
Anyone?

A given all pass filter convolved with its time-reversed all-pass filter results in a Dirac pulse. Thus a removed pre-ringing can be recovered.

I guess that the all-pass filter is simply applied despite nobody knows if the "master" track really needs it. As long as it does not really harm it does not matter.
Anyway it is no problem to create filters like described by the patent and to apply it even without MQA to verify its usefulness. Example: EQ Cookbook all-pass filter with f=20kHz and Q=0.7

Similar topic: aren't there several DACs out there where it is possible to switch the characteristic of the brickwall filter?

 

[Mr_Sukebe]

Yep, there are a number of CDPs and DACs that have filter options.
For that matter, ALL DACs include a filter (or so I understand), so all this talk about "exact" data is an interesting point, as every day is going to do it's own thing at the other end regardless.

As for the licensing and similar, do we have a similar threat about Dolby, as they've been earning money from licensing of their solutions for decades.

 

[DimitryZ]

A given all pass filter convolved with its time-reversed all-pass filter results in a Dirac pulse. Thus a removed pre-ringing can be recovered.

I guess that the all-pass filter is simply applied despite nobody knows if the "master" track really needs it. As long as it does not really harm it does not matter.
Anyway it is no problem to create filters like described by the patent and to apply it even without MQA to verify its usefulness. Example: EQ Cookbook all-pass filter with f=20kHz and Q=0.7

Similar topic: aren't there several DACs out there where it is possible to switch the characteristic of the brickwall filter?

All modern DACs have many variable filters that can be switched. Contrary to popular belief, many are audible, but not dramatically so.

 

[DimitryZ]

Yep, there are a number of CDPs and DACs that have filter options.
For that matter, ALL DACs include a filter (or so I understand), so all this talk about "exact" data is an interesting point, as every day is going to do it's own thing at the other end regardless.

As for the licensing and similar, do we have a similar threat about Dolby, as they've been earning money from licensing of their solutions for decades.

And now they have all the major streaming services, including the big ones - embracing Dolby Atmos, which is perceptually lossy and often plain bad when downmixed to two channels.

Yet, billions of electrons are used on audiophile forums complaining of eminent MQA takeover.

Isn't it ironic?

 

[DimitryZ]

IMO the little kernels and splines always have been a red herring.

At worst they have absolutely nothing to do with this.

At best they refer to the highest-rate anti-alias filters that MQA uses. But given how crude these filters are, their actual shapes are of little importance.

'If you can't dazzle them with brilliance, baffle them with "science".'



The pulses are stand-ins for real signals. And of course we want to locate real signals temporaly with whatever accuracy is required. But this does not pose any exotic demands on the detecting apparatus, be it a machine or your auditory system.

An example from my daily line of business: a spacecrafts's attitude is determined relative to the fixed stars. This is done with a camera and a slightly defocused lens. The camera may be coarse, e.g. a measly 512 x 512 pixels, and yet star positions are determined to deep sub-pixel accuracy, translating in mere arc-seconds attitude error. Why? It is not the sampling that is the limiting factor (the pixel array), but rather the system's noise floor. Sounds familiar?



Dirac pulses are mathematical abstractions: infinitely short, unit energy, hence infinite amplitude. With some hyperbolic analogy they are quickly likened to nuclear explosions or the big bang, both being wrong, but they convey the idea that a Dirac has no place in music.

For a truly elegant target tracking system, nothing beats the precessing focal point and circle of IR detectors of the original Sidewinder missile. Implemented with eight vacuum tubes. Gyro, spin control and primary telescope mirror was implemented with a single part made out of a magnet and polished into an optical surface.

 

[adamdea]

I guess that the all-pass filter is simply applied despite nobody knows if the "master" track really needs it. As long as it does not really harm it does not matter.

yes that is my question, especially if a minimum phase anti imaging filter is then applied.

Similar topic: aren't there several DACs out there where it is possible to switch the characteristic of the brickwall filter?

Not sure where you are going with this but it seemed to become fashionable around the time of the craven paper which mentioned the idea of removing pre -ringing. Difficult to argue with having a choice

 

[Mr_Sukebe]

That’s a very strange link to have suggested, as variable filtered CDPs/Dacs have been introduced occasionally for a long time

 

[adamdea]

They have been around but I don’t think they were common before 2004 were they?

 

[Mr_Sukebe]

Hardly as though they’re common now. Struggling to see your point