More on Lenbrook's plans for MQA

[kemmler3D]

It will have taken more than 3 emails to integrate the FOQUS architecture.

Yes, between the legal and marketing teams for sure.

One email to the engineering department would probably take care of it. "Please implement an apodozing filter with the following parameters (attached) for the 1.0 firmware on the new chip, thanks in advance"

 

[pogo]

Firmware is not enough. If you look at the data sheet, there is a separate digital signal path for FOQUS including the new hardware architecture.

 

[pogo]

What's so special about another version of minimum phase filter?
Inspira impulse response decay tail looks very uneven.
It's just another effect plugin for content creators.

There are other mechanisms behind Inspira, e.g.:
If the FOQUS by MQA Labs ADC has been used upstream of the Inspira plugin, it will automatically configure itself to avoid “overcorrection” to the signal.

 

[Nutul]

Firmware is not enough. If you look at the data sheet, there is a separate digital signal path for FOQUS including the new hardware architecture.

First things first: Demonstrate. De.mon.strate. Facts, not blablabla. People doing blind tests, first; and then architecture. There is no magic in patenting, uh? So, please, make the inner things known and you'll get credibility.

 

[Nutul]

There are other mechanisms behind Inspira, e.g.:
If the FOQUS by MQA Labs ADC has been used upstream of the Inspira plugin, it will automatically configure itself to avoid “overcorrection” to the signal.

Hmm, is this a "buy all from us" thingy...?
And then... what is "overcorrection"? Transparency should be enough for this to have no effect on the "corrected" signal...

 

[boxerfan88]

There are other mechanisms behind Inspira, e.g.:
If the FOQUS by MQA Labs ADC has been used upstream of the Inspira plugin, it will automatically configure itself to avoid “overcorrection” to the signal.

Errm…
1. How does it know F was used upstream? They encoded something/marker/flag into the recorded signal that should not be there? (corrupted) Or is the marker placed at some metadata area therefore not corrupting the recorded signal?


2. If the F marker was not detected, let’s say because of some other plugins upstream or the signal went through some analog chain causes the marker to be lost, then Inspira will over correct?

 

[boxerfan88]

If I've understood correctly, this is how it works - watermarking (but I'm only just reading up on it and it will probably take me another 10 man-years to get through all the published and related MQA ones): Link

Thus, the audio signal is subject to some pre-processing, producing a signal 104 that is clipped to known bounds. The Data Burier 114 then adds data-dependent noise of known peak magnitude to produce the output signal 102 on a quantisation grid O2. The noise is dependent on the data 143 to be buried, which comprises watermark data and additional data 141 produced by the Preprocessing.

:

Clipping full level sine waves causes visible distortion products on test equipment and to avoid criticism of the system fidelity there is a need to minimise the level of these distortion products.


Eeeeeeek .... watermark by corrupting original audio!!

So, let me try to recap my understanding. If any studio inadvertently uses Foqus during tracking, all the tracks shall be corrupted (watermarked). The only way to get back the original pristine audio is to use Inspira in the mixing/mastering chain. If Inspira is not used, the corrupted (watermarked) audio is then used for mixing/mastering. What this means is once Foqus is used during tracking, the original audio capture is permanently corrupted (watermarked), and they will pretty much hold you by the b#lls and force you to use their plugin to get back the original audio.

Is my understanding wrong? Sounds like MQA all over again .... but much much worse by poisoning the audio capture!
.

 

[Snoopy]

NAD / Bluesound / ESS , easy enough to ignore these companies.

 

[boxerfan88]

Yes, see NAD M33V2.

MQA FOQUS (ES9823MPRO only)

The ES9823MPRO is the world's 1st ADC to incorporate FOQUS by MQA Labs®. FOQUS introduces an innovative
approach to decimation in ADC chips, featuring a compact, minimum-phase response that stays below the human
sensitivity threshold. Its unique architecture is linear and low complexity, performing just a single dithered quantisation step.
This makes FOQUS the most transparent decimation process available today.

Using FOQUS with ESS' MQA DAC renderer enabled devices, such as the ES9039PRO, allows for an MQA analog to
analog connection. For more information, visit MQA's website: mgalabs.com.

Source: ES9823PRO/MPRO datasheet.

Thank God ESS is so smart! They produce ES9823MPRO and ES9823PRO (without M) ADC chips. ESS won't be suckered (locked-in) into the MQA nonsense.

If the main use case of ES9823MPRO is to provide "MQA analog to analog connection" in NAD devices, then all is good. Hope MQA just stays within the Lenbrook ecosystem. And I sincerely hope they don't try to poison the general music production chain.

I'm pretty confident most studios are smart enough to avoid having their content corrupted (watermarked) and having their b#lls held by MQA.
Similarly, audio interface manufacturers should be smart enough to avoid MPRO ADCs.

.

 

[Snoopy]

MQA FOQUS (ES9823MPRO only)

The ES9823MPRO is the world's 1st ADC to incorporate FOQUS by MQA Labs®. FOQUS introduces an innovative
approach to decimation in ADC chips, featuring a compact, minimum-phase response that stays below the human
sensitivity threshold. Its unique architecture is linear and low complexity, performing just a single dithered quantisation step.
This makes FOQUS the most transparent decimation process available today.

Using FOQUS with ESS' MQA DAC renderer enabled devices, such as the ES9039PRO, allows for an MQA analog to
analog connection. For more information, visit MQA's website: mgalabs.com.

Source: ES9823PRO/MPRO datasheet.

Thank God ESS is so smart! They produce ES9823MPRO and ES9823PRO (without M) ADC chips. ESS won't be suckered (locked-in) into the MQA nonsense.

If the main use case of ES9823MPRO is to provide "MQA analog to analog connection" in NAD devices, then all is good. Hope MQA just stays within the Lenbrook ecosystem. And I sincerely hope they don't try to poison the general music production chain.

I'm pretty confident most studios are smart enough to avoid having their content corrupted (watermarked) and having their b#lls held by MQA.
Similarly, audio interface manufacturers should be smart enough to avoid MPRO ADCs.

.

I would even be fine with the Nonsense if it were optional to use :-/ just add it as additional filter.

Makes Lenbrook look pretty unattraktiv as brand.

 

[Zgrado1970]

There are other mechanisms behind Inspira, e.g.:
If the FOQUS by MQA Labs ADC has been used upstream of the Inspira plugin, it will automatically configure itself to avoid “overcorrection” to the signal.

What do you gain by continuing to advocate for this business?

 

[boxerfan88]

I would even be fine with the Nonsense if it were optional to use :-/ just add it as additional filter.

Heh heh ... ESS made MQA Foqus sh*t defeatable via register 28 .

.

 

[pjug]

Extract from FOQUS White Paper:
FOQUS comes to market in partnership with semiconductor companies who specialise in audio chips.
A FOQUS ADC chip can be employed in any device where an analogue source is supported. Expect to see FOQUS in leading professional studio gear, prosumer devices, and high-end playback equipment. FOQUS currently offers output sample rates from 768kHz to 44.1kHz.

I took a little time to try to understand what is going on with FOQUS. The B Spline filters are very simple filters. These are just a cascade of moving average filters. Lots of ADCs use these, and the chip manufacturers usually use the terminology sinc2, sinc3, sinc4 to indicate the order number. They don't require much computational power so can run on just about anything. I use sinc4 filters myself for ADCs in my work; I had no idea I was having anything to do with B-splines! What it seems they are doing is all described in one of the FOQUS white papers. I can't seem to link to it right now but I uploaded it as an attachment.

The issues are pretty much along the same lines as problems with QRONO. These weak filters are probably fine for high sample rate recording that can then be downsampled with proper filters. But they want to use these filters to downsample all the way to 44K and 48K and so there is aliasing. In the paper they try to downplay this. But here you have artifacts in the audible band (worse than QRONO in that way). The aliasing may or may not be audible, but it seems like a terrible price to pay for the sake of solving "ringing".

 

[Nutul]

... and they will pretty much hold you by the b#lls and force you to use their plugin to get back the original audio.

Here it is: I tried to say it in a more polite way, but you nailed it. It is what it is. And somebody has to say it.

 

[GXAlan]

The aliasing may or may not be audible, but it seems like a terrible price to pay for the sake of solving "ringing".

It’s nice to have that AES paper.

They write: “the expected worst-case audio is already 80 dB below full scale, implying that in-band alias components would be below –125 dB and, significantly, 70 dB below the music content.

It’s not clearly a “terrible” price. But the question is if you get anything at all. They talk about “temporal blurring” without taking the time to prove that this is audible. Now that we have the AES papers, we can pull:

AES Journal Forum » The Gentle Art of Dithering

The AES Forum is where our members get to interact with each other in between conferences and section meetings.

secure.aes.org

PDF publicly available here.

They then reference

But haven’t tried to see what those papers say…

If MQA is inaudible, then the argument that it’s a poison is countered by the idea that it enables marketing budgets toward creating new recordings. It’s a bit like PlayStation, Xbox, Switch, or macOS first party exclusives.

If MQA aliasing is audible, in rare cases (which does NOT seem to be the argument here) then we have to ask if temporal blurring is audible, in rare cases. Then, it becomes a production design choice in trade-offs…

 

[kemmler3D]

Thus, the audio signal is subject to some pre-processing, producing a signal 104 that is clipped to known bounds. The Data Burier 114 then adds data-dependent noise of known peak magnitude to produce the output signal 102 on a quantisation grid O2. The noise is dependent on the data 143 to be buried, which comprises watermark data and additional data 141 produced by the Preprocessing.

:

Clipping full level sine waves causes visible distortion products on test equipment and to avoid criticism of the system fidelity there is a need to minimise the level of these distortion products.


Eeeeeeek .... watermark by corrupting original audio!!

So, let me try to recap my understanding. If any studio inadvertently uses Foqus during tracking, all the tracks shall be corrupted (watermarked). The only way to get back the original pristine audio is to use Inspira in the mixing/mastering chain. If Inspira is not used, the corrupted (watermarked) audio is then used for mixing/mastering. What this means is once Foqus is used during tracking, the original audio capture is permanently corrupted (watermarked), and they will pretty much hold you by the b#lls and force you to use their plugin to get back the original audio.

Is my understanding wrong? Sounds like MQA all over again .... but much much worse by poisoning the audio capture!
.

To me this sounds like steganography which uses a very low level and very low frequency signal to hide a few bits in an audio track. It's been used for years (it was / is considered a way to catch pirates) and afaik hasn't been considered a threat to fidelity.

 

[pjug]

They write: “the expected worst-case audio is already 80 dB below full scale, implying that in-band alias components would be below –125 dB and, significantly, 70 dB below the music content.

That is their analysis at 96KHz, not 44KHz or 48KHz. But even so, they point out that their analysis is only for music signals and that "Of
course, there can be tonal signals in that frequencyrange from monitors or power supplies". But then they don't say anything more about potential problems from pickup of those kinds if signals.

 

[GXAlan]

Here’s a question.

If I took two recordings, one processed with and without MQA and people failed to discern a difference in blind testing, would that convince you that MQA was not harmful? One might argue that there would need to be more test tracks.

If I took two recordings and the MQA track was universally preferred in blind testing, would that convince you that MQA was not harmful? One might argue that there are other remastering tweaks that result in two different recordings…

But if I point out that there are some MQA exclusive content, implying marketing dollars being directed, how is MQA different from iMovie/Final Cut Pro as MacOS exclusives or Mario Kart as a Nintendo exclusive?

 

[kemmler3D]

Here’s a question.

If I took two recordings, one processed with and without MQA and people failed to discern a difference in blind testing, would that convince you that MQA was not harmful? One might argue that there would need to be more test tracks.

If I took two recordings and the MQA track was universally preferred in blind testing, would that convince you that MQA was not harmful? One might argue that there are other remastering tweaks that result in two different recordings…

But if I point out that there are some MQA exclusive content, implying marketing dollars being directed, how is MQA different from iMovie/Final Cut Pro as MacOS exclusives or Mario Kart as a Nintendo exclusive?

I'll bite.

As far as I know MQA didn't have much/any audible effect on audio and I expect FOQUS will be the same.

So in that sense it's not harmful to the recording in practice even if it's giving up a few bits of quality in theory.

You could even construct an argument that it's beneficial to listening, in that it gets people to listen more closely, trying to hear its effect, which then helps their listening experience.

But I do think it's harmful to the consumer and industry, though. If we assume it has no audible effect, then this is a lot of engineering and marketing going towards nothing. And maybe worse, it's driving consumer perception toward believing BS about audio. Which is something that is already out of control.

I don't know if "throwing gasoline on a raging fire" is too dramatic, but it's not what anyone needed or asked for.

 

[boxerfan88]

But I do think it's harmful to the consumer and industry, though. If we assume it has no audible effect, then this is a lot of engineering and marketing going towards nothing. And maybe worse, it's driving consumer perception toward believing BS about audio. Which is something that is already out of control.

Hear! Hear!

.