MQA Discussion Group

0 Members and 1 Guest are viewing this topic. Read 18820 times.

James Tanner

  • Facilitator
  • Posts: 16591
  • The Demo is Everything!
    • http://www.bryston.com
Re: MQA Discussion Group
« Reply #80 on: 9 Sep 2017, 03:35 pm »
OK, but the same can be said for any company that makes speakers, amps, DAC's, etc and makes claims about their products. A "vested interest group" is called a company right? Every company has a vested interest in peddling their wares. Why single out codecs, the same can be said for hardware. How many companies claim that their proprietary wave guide, power supply, circuits or whatever provides blah, blah, blah. If we are going to discuss marketing claims you gotta look at the industry (new thread would be needed if we open that can of worms  :o)

Hi

I agree and that's why I am glad i am involved in a company - Bryston - that builds honest scientifically based products at honest prices and is open to all rational thought.

james

witchdoctor

Re: MQA Discussion Group
« Reply #81 on: 9 Sep 2017, 04:20 pm »
Hi

I agree and that's why I am glad i am involved in a company - Bryston - that builds honest scientifically based products at honest prices and is open to all rational thought.

james

IMO transitioning from a passive speaker to a good active speaker is MUCH more important than a codec :D

zoom25

  • Full Member
  • Posts: 586
Re: MQA Discussion Group
« Reply #82 on: 9 Sep 2017, 06:09 pm »
MQA. Kill it with fire. We all gotta pay for it one way or another even if we don't use MQA.

Reminds me of the first two minutes of this Wilt Chamberlain interview: https://www.youtube.com/watch?v=8vnQ7c1y-vM

Every time I've spoken with Lucey, he's been straight forward. If you need MQA to achieve a particular sound, you simply aren't cut out for being an engineer in the first place. Learn how to make it sound good. Keep it lossless and deliver it lossless. MQA makes zero sense in 2017. We have FLAC and WAV in both 16 and 24 bits. You want to save space, then go for FLAC. You want uncompressed, there is WAV and AIFF. You really want to save space, there is MP3.

If Bryston pays for MQA, we'll end up having to pay extra for that garbage. Same reason why the BDA-3 costs more because of HDMI licensing. (Correct me if I'm wrong here.)

MQA. Kill it with fire.

CanadianMaestro

  • Full Member
  • Posts: 1117
  • Skepticism is the engine of true progress
    • Hearing Everything That Nothing Can Measure
Re: MQA Discussion Group
« Reply #83 on: 9 Sep 2017, 06:49 pm »
MQA. Kill it with fire. We all gotta pay for it one way or another even if we don't use MQA.

Reminds me of the first two minutes of this Wilt Chamberlain interview: https://www.youtube.com/watch?v=8vnQ7c1y-vM


Good video.

Money and disrespect. Next to death and taxes.....

AmpDesigner333

  • Industry Participant
  • Posts: 1893
    • Digital Amplifier Company
Re: MQA Discussion Group
« Reply #84 on: 9 Sep 2017, 07:08 pm »
"MQA: No need to worry" thread:
http://www.audiocircle.com/index.php?topic=147977.0
There is a HUGE misconception about what MQA really is.  Here's an excerpt from the thread above:

"MQA is just a CODEC (code-decode).  It's an algorithm, a.k.a. "software".  It's a "lossy" format, meaning compressed, like MP3.  It's purpose is to conserve transmission bandwidth - which is less of an issue every day.

The device (PC, streamer, transport, etc.) that drives your DAC runs software that decodes various formats and passes them along or converts them to some other format.  Ultimately, there's a hardware connection to transport the data to your DAC.  The stream being transported is typically PCM or DSD.  The "transport connection" is typically USB or SPDIF.

In the case of a PC streaming MQA, the data can be passed along to the DAC or decoded by the PC and converted to a "usable format" for your DAC.  This means your DAC doesn't need to support MQA for you to listen to MQA.

Let's take JRiver for example.  JRiver plays files (or streams), and can convert just about anything to PCM.

So, whatever you use to playback audio files or stream audio should provide a way to convert to PCM.

Here's the "controversial stuff"....  MQA is yet another "new" standard in the search of license fees, and many manufacturers are rejecting it.  There's just a lot of hype surrounding it currently, possibly due to lack of other industry news.

IMHO, Lossless high rate, high resolution PCM (not DSD, as I have discussed in other threads) remains the highest quality audio."

Krutsch

  • Full Member
  • Posts: 472
Re: MQA Discussion Group
« Reply #85 on: 9 Sep 2017, 11:01 pm »
WTF is HDCD??  :scratch:

 :sleep:

A format that was competing with DVD-Audio and SACD that didn't survive.

witchdoctor

Re: MQA Discussion Group
« Reply #86 on: 10 Sep 2017, 01:33 am »
As far as the MQA format goes it is a LOT younger than PCM or DSD. All of the major labels are on board (woo hoo) and I see new MQA albums every week (every day?) on Tidal. Yesterday it was Pet Sounds, today the R.E.M. catalog, etc. While Bryston is passing on implementation for 2018 there are some very interesting vendors getting on board. Right now NAD/Bluesound and Mytek appear to be the most popular players in the MQA space. At CES we will likely see a lot of announcements from new hardware manufacturers, more releases from the labels, and more streaming services. I can't wait :thumb:

http://www.mqa.co.uk/customer/our-partners

Xinon

Re: MQA Discussion Group
« Reply #87 on: 10 Sep 2017, 04:17 am »
Mastering Engineer Brian Lucey on MQA:

http://magicgardenmastering.com/

MQA is, in all of my tests so far, brighter and thinner with more distortion, and a sense of excitement (remember Aphex?) and even volume from the artifacts. Sure many lay people will be fooled. I don't like to see pros fooled, but oh well, that's why I'm busy I suppose and they're posturing next to a new product so I'm free to tell the truth here.

If there is a market for MQA sure that's great, all good. Yet I don't think we should let Audiophiles (who are a small market in the first place) set definitions on good sound quality. They're not even remotely objective as they don't have the source files and do not know the intentions of the creative team. They're also easily fooled by all sorts of things, because they are essentially experiencing creativity, by tweaking the playback.

That's fine, and I love them for paying money for great systems, but MQA is what it is - a form of creativity and control.

Deblurring

The bottom line for me, this whole "bettering" the source file and "deblurring" malarkey has to be stopped. It's less data and a new filter ... not better or equal to the source files, that are not that much larger, and are easily downloaded by most people in the world with 2017 tech.
To say otherwise is either greed, a dreamers imagination or just bad I helped Mytek redesign the Manhattan II, and everything in the room was fine. Previously I also heard a flaw in the Bricasti M1 and helped them to make it greater, after years on the market. My room set up is working very well, thank you.

Yet the fact is MQA is not better and not neutral, anyone who has done the tests that has endorsement integrity I respect, hears the same thing. Corporations are getting on board to make money, and people are lining up to get on that gravy train.

I'm fine with any new product clients want to pay for, I just think we need to live in reality. Less data and a filter is not better, and should seem a skeptical claim even without hearing anything. Better than the source, and any kind of "correcting PCM" claims are just absurd.

100% agree!
Thats my point exactly, give us the "real" masters, bandwith is no longer an issue(in most part of the world).

gbaby

  • Jr. Member
  • Posts: 134
Re: MQA Discussion Group
« Reply #88 on: 10 Sep 2017, 04:23 am »
As far as the MQA format goes it is a LOT younger than PCM or DSD. All of the major labels are on board (woo hoo) and I see new MQA albums every week (every day?) on Tidal. Yesterday it was Pet Sounds, today the R.E.M. catalog, etc.

If you owned the major labels and could resale by recycle your products to the same people, wouldn't you be on board?  :nono: Its marketing but it takes a lot of discernment to understand this fact.

gbaby

  • Jr. Member
  • Posts: 134
Re: MQA Discussion Group
« Reply #89 on: 10 Sep 2017, 04:24 am »
Delete

witchdoctor

Re: MQA Discussion Group
« Reply #90 on: 10 Sep 2017, 04:48 am »
If you owned the major labels and could resale by recycle your products to the same people, wouldn't you be on board?  :nono: Its marketing but it takes a lot of discernment to understand this fact.

If I owned the major labels I would start drinking, have you seen this chart on music sales?

https://blog.thecurrent.org/2014/02/40-years-of-album-sales-data-in-one-handy-chart/

It is NOT marketing, it is survival because of pirating, it takes 0 discernment to understand that fact.

R. Daneel

  • Full Member
  • Posts: 683
Re: MQA Discussion Group
« Reply #91 on: 10 Sep 2017, 06:45 am »
A format that was competing with DVD-Audio and SACD that didn't survive.

Completely untrue. HDCD stands for "high definition compatible digital" and was developed by Pacific Microsonics to complement the CD format, not to replace anything. HDCD discs are compatible with any CD player but to uncover the extra 4 bits, you need a HDCD decoder.

Considering HDCD discs are usually well recorded anyway, the additional decoding doesn't make a whole lot of difference.

R. Daneel

  • Full Member
  • Posts: 683
Re: MQA Discussion Group
« Reply #92 on: 10 Sep 2017, 06:49 am »
If you owned the major labels and could resale by recycle your products to the same people, wouldn't you be on board?  :nono: Its marketing but it takes a lot of discernment to understand this fact.

I couldn't have said it better myself.

James Tanner

  • Facilitator
  • Posts: 16591
  • The Demo is Everything!
    • http://www.bryston.com
Re: MQA Discussion Group
« Reply #93 on: 10 Sep 2017, 10:53 am »
Hi Folks,

This is really long but will give you a good technical overview of what MQA does under the cover.





Technical Questions Regarding MQA

MQA stands for Master Quality Authenticated and has been widely promoted as a file format that achieves two separate goals:

1. The ability to deliver high-resolution digital audio files in a limited bandwidth “pipeline”.

2. The ability to correct faults in the original digital recording for improved overall fidelity.

Both goals are laudable. From what we can find MQA has not has been explained clearly and openly enough to answer all technical questions, that we feel are relevant and timely, about the process. After careful study of the information regarding MQA, including publicly released material, presentations, the patent application, and journal articles, we believe we have gained insight into the technical operation of MQA. Even after this research, questions still remain.

We would like to summarize what we have learned, and ask further questions, so that we can better understand MQA.

The questions we ask are those that we feel need to be asked, and answered. This paper looks to do both, to the best of our understanding. It is offered in the spirit of open inquiry. We have prepared this document to assist all interested parties – customers, manufacturers, journalists, and digital designers.

These are our questions:

1. Is MQA Lossless?
2. Why Isn’t It Clear Whether MQA Is Lossy Or Not?
3. How Lossy Is MQA?
4. Can MQA Improve The Sound Quality Of Existing Digital Recordings?
5. Does MQA use DRM? (Digital Rights Management, or “copy protection”)

1.Is MQA Lossless?

In general, MQA has been promoted as a high-resolution, and therefore presumably lossless technology. The question naturally arises, “How is it possible to fit a full highresolution digital audio file into a smaller container than previously thought possible?” The short answer is that according to all technical information reviewed by us, MQA is a lossy process. Even when fully decoded by an MQA decoder, the original high-resolution audio file irretrievably loses resolution when run through the MQA encode/decode process. To understand this more clearly, we will adopt some of the terminology used by MQA:

Baseband: Audio information from 0 - 24kHz, the Nyquist limit of 48kHz sampling.
Dual-rate: Audio information from 24kHz - 48kHz, the Nyquist limit of 96kHz sampling.
Quad-rate: Audio information from 48kHz - 96kHz, the Nyquist limit of 192kHz sampling

(Naturally these numbers can easily be adjusted for multiples of the CD audio sampling rate of 44.1 kHz.)

There are two distinct ways in which MQA irretrievably loses resolution:

1. Using a technique called “encapsulation”, MQA fits the quad-rate audio (48kHz –96kHz) information using lossy compression algorithms similar to those used in MP3. All lossy compression schemes (eg, MP3, AAC) irretrievably lose resolution. (New evidence suggests that MQA’s “encapsulation” scheme actually discards all audio information above 48kHz.) In contrast, dual-rate audio is compressed using a lossless compression scheme. All existing lossless audio compression schemes to date (eg, FLAC, ALAC, Dolby True HD, and others) are based on predictive algorithms developed by the late Michael Gerzon1. Although not specified, it seems likely that the lossless compression used by MQA for the dual-rate audio is also based on Gerzon's predictive algorithms, likely even FLAC itself.

2. The high-resolution audio is delivered by MQA in a baseband (44.1kHz or 48kHz) FLAC container that is 24 bits deep. MQA's “audio origami” process “folds” the lossless dual-rate and “encapsulated” quad rate audio data into the lower bits of the 24-bit FLAC container. MQA's literature states that the number of Least Significant Bits reserved for the “folded” dual and quad-rate audio is dynamically allocated, depending on the spectral content of the music encoded. All of the presentations from MQA to date show that the musical example is a Ravel string quartet. In the MQA presentation to the Japan Audio Society, published in November 2015, after the first “fold” from 192kHz to 96kHz, the resolution has been decreased from 24 bits to 17.2 bits. It appears, based on all known information supplied by MQA, that a 24-bit audio master file will be reduced to no more than 17bits of resolution of average resolution across the audio band, and furthermore that all quad-rate audio data will either be compressed using lossy techniques or discarded entirely. The bottom line is that MQA is unquestionably a lossy format, and that the original highresolution master file can never be recovered from the MQA file.

2.Why Isn’t It Clear Whether MQA Is Lossy Or Not?

The earliest technical information we can find regarding MQA was the JAES paper “A Hierarchical Approach to Archiving and Distribution”, released in October 2014:

http://www.aes.org/e-lib/download.cfm/17501.pdf?ID=17501

At the time of the paper’s release, authors Bob Stuart and Peter Craven chose to present their audio graphs (see Figure 8) in “Noise Spectral Density [NSD] dBFS [dB Full Scale] per-root Hz” with a 1 Hz bandwidth. In this paper the noise floor for 16-bit, 192 kHz audio signal dithered with a TPDF (Triangular Probability Distribution Function) was stated as -144 dBFS per-root Hz, 1http://www.audiosignal.co.uk/Resources/Digital_squashing_A4.pdf while a 20-bit, 192 kHz audio signal dithered with a TPDF (Triangular Probability Distribution Function) was stated as -168 dBFS per-root Hz. This is a full change of 48 dB over standardly-accepted digital audio practice where it widely agreed that 16 bits yields roughly -96 dB S/N and 24 bits yields roughly -144 dB S/N:

https://en.wikipedia.org/wiki/Signal-to-noise_ratio#Digital_signals

Additional detailed technical information on the quantization noise that results from digitizing an analog audio signal is available from Analog Devices:

http://www.analog.com/media/en/training-seminars/tutorials/MT-001.pdf

Later presentations from MQA still use an unconventional noise scale (sometimes labeled “Spectral Level”, sometimes labeled “Bin Level”) but now change the S/N ratio by 24 dB. This yields -120 dBFS for 16 bits and -168 dBFS for 24 bits. We feel that the way in which this information is presented obscures the underlying performance limits of MQA, making it very difficult for all but the most technically proficient to grasp the full implications of the information. No justification of the unorthodox approach is offered by MQA other than the fact that the authors themselves have used it in previously published papers.

3. How Lossy Is MQA?

Using the example of the Ravel string quartet from MQA's paper presented at the Japan Audio Society, we know that the maximum baseband audio resolution is reduced by at least 6.8 bits. We wonder if it is possible that there are even higher losses of resolution? To answer this question, we refer to the classic paper by Caltech Professor James T. Boyk, “There's Life Above 20 Kilohertz!”, published in 1992 and available at: http://www.cco.caltech.edu/~boyk/spectra/spectra.htm In this paper, Prof. Boyk examines the spectral content of a wide variety of musical instruments, including those that produce definite pitches, and also those that do not. Prof. Boyk's measurement equipment limited the measurements to 102.4 kHz. Printed below are two charts from this paper, the first from musical instruments with definite pitches and the second from musical instruments and other sound sources without definite pitches.

Table 1: Musical instruments with definite pitch.

Instrument SPL (dB) Upper Limit of Harmonics (KHz) Power Above 20 kHz (%)
Trumpet (Harmon mute) 96 >50 0.5
Trumpet (Harmon mute) 76 >80 2.0
Trumpet (straight mute) 83 >85 0.7
French horn (bell up) 113 >90 0.03
French horn (mute) 99 >65 0.05
French horn 105 >55 0.1
Piano 111 >70 0.02
Violin (double-stop) 87 >50 0.04
Violin (sul ponticello) 77 >35 0.02
Oboe 84 >40 0.01

Table 2: Musical instruments and other sources without definite pitch.

Instrument SPL (dB) Upper Limit of Harmonics (KHz) Power Above 20 kHz (%)
Speech Sibilant 72 >40 1.7
Claves 104 >102 3.8
Rimshot 73 >90 6.0
Crash Cymbal 108 >102 40.0
Triangle 96 >90 1.0
Keys jangling 71 >60 68.0

Prof. Boyk's analysis of the high-frequency content of various musical instruments is important to understanding MQA's presentations. This is because the choice of string quartet seems an odd one. One can readily see that the violin has the lowest level of highfrequency energy of all musical instruments tested, except for a solo oboe. (Not listed in Prof. Boyk's paper are the other two members of a string quartet, the viola and the cello. It seems likely they have even less high-frequency content than a violin. In contrast, most popular music and much classical music (eg, symphonic works) employ cymbals extensively. We can see that a cymbal has 1000x (+30dB of power) as much energy above 20 kHz as does a violin. It is reasonable to assume that using “audio origami” to “fold” musical content with a much higher level of energy above 20 kHz than a string quartet into the lowest bits of MQA's 24-bit FLAC container will require more than the 6.8 bits required for the first “fold” of a string quartet.

Unfortunately we can only speculate at this point. MQA has released no data regarding the amount of bits required to “fold” the dual- and quad-rate audio data for musical content with high levels of audio energy above 20 kHz. Is it possible that an MQA encoded file reduces the baseband audio resolution below 16 bits? As strange as it may seem for a self-proclaimed “high-resolution” audio format, it seems possible. At this point we have no way of knowing, as MQA has released no data from which we can draw meaningful conclusions. What we do know at this point is that the resolution from the first “fold” of a string quartet (with extremely low levels of musical energy above 20 kHz) reduces the baseband audio resolution to 17.2 bits. We would like to know what the resolution in the baseband of other types of music is when it is processed by MQA.

4. Can MQA Improve The Sound Quality Of Existing Digital Recordings?

There have been several anecdotal reports from well-respected journalists and recording engineers that MQA can improve the subjective sound quality of existing digital recordings. (Some of these reports are from parties alleged to have received financial support from MQA, and are therefore highly suspect.) MQA claims that their process can correct for errors in the original digital audio file created by the specific A/D converter used to make the recording. There are two distinct cases where this concept can theoretically be applied. It is useful to examine each case separately, in order to gain the greatest insight into this question:

First case:

A single known converter is used to create the original digital audio file.

Examples of this would include:

a) Transfers of analog sources (typically tape) using a single known A/D converter.
b) “Purist” digital recordings made with a single known A/D converter and captured directly to a digital tape or hard drive storage system.

In cases such as these the only tools available (other than complete remastering, which could include EQ, compression, reverb, and multiple other effects) are to use various digital filters to “compensate” for the “shortcomings” of the original recording. To the best of our knowledge, the first company to use custom digital filters was Wadia in the late 1980s. Wadia was founded by a group of digital engineers who left 3M. Perhaps Wadia's major contribution to the understanding of digital audio is the insight that there is an inextricable link between the steepness of a filter (digital or analog), the sharpness of the “knee”, and the associated amount of distortion in the time domain (MQA’s so-called “time blur”).

Prior to Wadia's custom digital filters based upon DSP (Digital Signal Processing) chips, all commercial digital filters used as steep a slope as possible, with a sharp a knee as possible (given the limitations of the then-current technology). The reason for this is to produce as close an approximation as possible (given the real-world constraints of both price and silicon “real estate” in the integrated circuits) to the theoretical ideal of a sinc (sin(x)/x) filter that was assumed to be the “correct” filter for digital audio, according to the original ground-breaking work in the theory of digital audio by Shannon and Nyquist. When an impulse is passed through a steep filter with a sharp knee, it will spread that impulse over time. (A sinc filter would spread the impulse over an infinite amount of time.) MQA refers to this spreading of an impulse as “time blur”. Perhaps the ultimate expression of this is the theoretical sinc filter, which has an infinitely steep cutoff slope and an infinitely sharp knee. The reason this filter is theoretical is because it would require an infinite number of taps. Digital filters, which sample the audio signal in discrete-time increments, multiply each sample of the audio data by a coefficient to create the desired filter. As the digital audio samples are sent through the time-delayed sequential pipeline, the audio samples are diverted to be multiplied by the filter coefficients via “taps”. To reduce the number of taps from infinity (for a sinc filter) to a physically realizable number, a “window” is applied to the sinc function, resulting in a filter with a finite number of taps. Common windows include rectangular, triangular, Hamming, Hanning, Kaiser, cosine, Blackman, Lanczos, cosine, Dolph-Chebyschev, and many, many more, including an infinite number of arbitrary ones.

If all else is equal, a greater number of taps will yield a digital filter with a sharper knee, a steeper cut-off slope, a greater stop-band attenuation, or some combination of these three parameters. Again, assuming all else is equal, a greater number of taps will create more spreading of an audio signal impulse – what MQA refers to as “time blur”. When digital audio was introduced, virtually all digital filters were of the type known as linear phase. A linear-phase filter is still the standard digital audio filter used to this day for both record (A/D conversion) and replay (D/A conversion) and has several advantages:

a) There is no phase shift through the filter, regardless of frequency.
b) The coefficients of the filter are symmetrical, requiring only half as many coefficients for a given filter. This markedly reduces the cost of the filter, in terms of both money and also real estate in the silicon chip’s memory required to store the coefficients.

The disadvantage of a linear-phase filter is that there will be energy released before the onset of an impulse. This is completely unknown in nature (acausal, meaning the event occurs before the cuase). In contrast, a minimum-phase filter has no energy released before the onset of an impulse, but only after. This mimics the real-life situation whereby all sounds create echoes that follow the original impulse. Nowhere in real life is there an a causal situation whereby there are “pre-echoes” that precede the original sound.

The disadvantages of a minimum-phase filter are that it produces phase shift (in the case of digital audio, this is generally confined to low levels of phase shift and only at high frequencies close to the cut-off frequency of the low-pass filter) and that it requires twice as many coefficients as an equivalent linear-phase filter, increasing both the monetary cost and the associated silicon area to store twice as many filter coefficients. One of the earliest attempts to correct for errors in the A/D converter known to us was the “apodizing” filter described by Peter Craven in his JAES paper from March 2004, “Antialias Filters and System Transient Response at High Sample Rates”. The principle used in Craven's apodizing filter placed a single minimum-phase low-pass filter at the end of the playback chain, with a cutoff frequency just below the Nyquist frequency of all of the “upstream” linear-phase digital filters.

The idea was to filter out all of the pre-ringing from the linear-phase filters used in the A/D converter and replace it with only post-ringing from the minimum-phase apodizing filter. Subjective studies have indeed found that Craven's apodizing filter did provide improved subjective sonic results compared to the typically used linear-phase filters of that era. However, these studies also found that the post-ringing created by the steep, sharp-knee apodizing filter also produced unwanted sonic artifacts. In short apodizing filters as such may be helpful in some instances, but are not the “answer” to the problems of digital. It appears that the filters used in the MQA process are slow-rolloff apodizing filters, and these may provide sonic benefits over traditional “brickwall” linear phase filters. However there are many commercial products available using these types of filters, and there is no need to create a proprietary, closed format with licensing fees and royalties to benefit from improved digital filters.

Second case:

This is much more complex…
Virtually all modern recordings are made in the digital domain. It is not unusual for multiple brands and models of A/D converters to be used for different tracks of a multitrack recording. These A/D converters are often different brands and models, and may even be located in separate recording facilities and recorded at different times using different sample rates. In addition, the digital mixing console (DAW, or Digital Audio Workstation) will have its own digital filters. The end result is far from the first case of a single known A/D converter. It is unclear to us how any particular compensation could be applied to such a situation.

Back in the late '80s, Wadia showed the world the audible sonic impact digital filters had. Since that time more and more digital audio design and manufacturing companies have begun using custom digital filters, both for A/D conversion as well as D/A conversion. We know from the JAES paper and the MQA patent filing that MQA uses a slow roll-off digital filter. The complete MQA A/D + D/A chain running at 192 kHz has a frequency response that reaches -3 dBFS at ~37 kHz and -10 dBFS at ~51 kHz. It is this slow rolloff that allows for improved response in the time domain - the reduction of what MQA refers to as “time blur”, but in reality this idea has been used in other commercially available products for nearly three decades.

Does MQA Use Digital Rights Management?

MQA has stated many times that there is no DRM (Digital Rights Management, or “copy protection”) used in its files. This raises many questions, as it is quite clear that there is DRM embedded in the system. This is easily seen by visiting the website of Utimaco of Germany, one of the premier companies in the world of of professional cybersecurity solutions. Their work with MQA was used as a case study by Utimaco, and this brochure can be downloaded from the Utimaco website:

https://hsm.utimaco.com/wp-content/themes/utimaco-hsm/PDFMerger/mergedpdfs/utimaco_hsm_download_files_59a2261413eec4.29482009_20170827.pdf

Clearly there is some type of DRM built into MQA. As to what it is (or will be) used for, one can only speculate, as MQA denies this even exists. Some possibilities include tracing the lineage of potentially “pirated” files, but also extend to removing the rights of customers, as Sonos has begun to do with their service:

https://www.techdirt.com/articles/20170822/10030038059/sonos-users-forced-tochoose-
between-privacy-working-hardware.shtml

Conclusion

MQA's own documents reveal that the highest resolution of which MQA is capable is only 17.2 bits – a loss of 6.8 bits from the 24-bit high-resolution source files. We don't know if the losses extend past that with recordings containing high-frequency energy (such as cymbals), nor do we know if the losses are limited – say to even 8 bits (16-bit resolution).

As time goes on and we learns more about MQA we will continue to unravel the questions surrounding it. At this point in time it appears that the MQA digital filter offers superior time-domain performance (reduced “time blur”) compared to the digital filters built into most DAC chips. Yet for those companies who choose to create their own custom digital filters, it is trivially easy to create filters that are at least the equal of their MQA equivalents, and potentially superior. MQA presents a road-block for further development of digital filters. This applies to both the professional community (using A/D converters) and the consumer community (using D/A converters) as the filters in both converters are fixed. This limits innovation and the fruits of continuous research, development, and improvement.

Finally, the fact that DRM is embedded into MQA files raises a specter of potential abuse of customer’s legitimate rights to make backup copies or to transfer purchased media to other devices, as well as to limit the future rights of customers who choose to rent (“stream”) musical content using MQA.

THROWBACK

  • Jr. Member
  • Posts: 247
Re: MQA Discussion Group
« Reply #94 on: 10 Sep 2017, 01:21 pm »
Thanks, James. Most comprehensive treatment yet, beautifully presented.

CanadianMaestro

  • Full Member
  • Posts: 1117
  • Skepticism is the engine of true progress
    • Hearing Everything That Nothing Can Measure
Re: MQA Discussion Group
« Reply #95 on: 10 Sep 2017, 02:22 pm »

Conclusion

MQA's own documents reveal that the highest resolution of which MQA is capable is only 17.2 bits – a loss of 6.8 bits from the 24-bit high-resolution source files. We don't know if the losses extend past that with recordings containing high-frequency energy (such as cymbals), nor do we know if the losses are limited – say to even 8 bits (16-bit resolution).

As time goes on and we learns more about MQA we will continue to unravel the questions surrounding it. ....it is trivially easy to create filters that are at least the equal of their MQA equivalents, and potentially superior. MQA presents a road-block for further development of digital filters. This applies to both the professional community (using A/D converters) and the consumer community (using D/A converters) as the filters in both converters are fixed. This limits innovation and the fruits of continuous research, development, and improvement.

Finally, the fact that DRM is embedded into MQA files raises a specter of potential abuse of customer’s legitimate rights to make backup copies or to transfer purchased media to other devices, as well as to limit the future rights of customers who choose to rent (“stream”) musical content using MQA.


 :o :thumb:

witchdoctor

Re: MQA Discussion Group
« Reply #96 on: 10 Sep 2017, 02:29 pm »
I can find a whitepaper to support PCM over DSD, that DSD is better than PCM, that there is no difference between the two or that it is really the server that is more important than the DAC/CODEC. Take your pick.
What I do know is that the majority of hirez DSD and PCM files cost $20+ a pop and you need to buy a dac/processor/player capable of decoding them.
MQA is available for $20 a month with new titles being released constantly. No special player/DAC needed.
The people making the most noise protesting MQA are freaking out that their products are going the way of the betamax. They have already planted themselves in a chosen formats corner, no new contenders needed thank you. Especially a contender that costs significantly less and is easier to implement. That includes any manufacturer making DSD related products or files or hi rez PCM whether it is hardware or software related.
For all you hirez PSM/DSD fans why hasn't it been widely adopted by consumers or the major labels? If you look at the FACTS the only sector that has widely adopted hirez PCM/DSD are the hardware manufacturers, the same people who are freaking out about MQA.
For me all it took was listening to an MQA playlist for 30 minutes or more and then try going back to FLAC. FLAC lacks in comparison to my ears. If I had thousands of hirez titles to stream in comparison I would but oops, I would have to BUY them (at least the ones I could find) for $$$$$, good luck with that.







CanadianMaestro

  • Full Member
  • Posts: 1117
  • Skepticism is the engine of true progress
    • Hearing Everything That Nothing Can Measure
Re: MQA Discussion Group
« Reply #97 on: 10 Sep 2017, 02:38 pm »
Rational science and objective measurements reveal facts.
That's not "freaking out".

Nobody's "freaking out" here, other than the MQA proponents elsewhere, who support false claims from the MQA "engineers" like Stuart et al., by buying into that format.

I will bet the same people who deny climate change (!) are also the ones likely to buy into MQA.


jseymour

Re: MQA Discussion Group
« Reply #98 on: 10 Sep 2017, 02:40 pm »
Witchdoctor, address the technical and DRM issues of MQA and we might take you seriously.

witchdoctor

Re: MQA Discussion Group
« Reply #99 on: 10 Sep 2017, 02:47 pm »
Rational science and objective measurements reveal facts.
That's not "freaking out".

Nobody's "freaking out" here, other than the MQA proponents elsewhere, who support false claims from the MQA "engineers" like Stuart et al., by buying into that format.

I will bet the same people who deny climate change (!) are also the ones likely to buy into MQA.

If you feel spending $25 for Marvin Gayes "What's Going On" in 24/96 PCM is rational when I can stream the MQA version instantly be my guest. I am sure there are "rational" download sites lined up to take your money, have fun :o