I'm listening to MQA from Tidal right now, using a Audioquest Dragonfly Black DAC. The software decodes up to 24/96, which is sent to the DAC. The time details are still lost with this, but I'll have to say it sounds pretty good. Whether we needed this format is another question. What is the point of compression when we have internet speeds now that support 4k video with ease?
My count came to 514 titles on Tidal with "Master" MQA quality. That will get old really fast when you consider that this is for all genres. It seems heavy on classic rock remasters, some newer pop/rock, and some well done classical, but very little jazz.
My count came to 514 titles on Tidal with "Master" MQA quality. That will get old really fast when you consider that this is for all genres. It seems heavy on classic rock remasters, some newer pop/rock, and some well done classical, but very little jazz.
I have no inside information regarding MQA, nor seen a licensee specification. However, from what I've been able to glean from reading public info. about MQA, it seems rather clever. Here are the primary objectives of MQA, as they appear to me from the outside looking in.
1. Minimization of temporal blurring, or smearing. This seems the overriding performance concern. Of course, the audibility of such smearing is controversial. It is, however, easy to see the smearing mechanism occuring within digital anti-alias and anti-image filters.
For the sake of illustration how, let's take 10 cycles of an Fs/4 tone burst to represent a transient and pass it through an 255 stage linear phase FIR filter. The 10 cycle burst will, on average, have a duration of, 4 samples/cycle = 40 sample periods. After passing through a 255 stage filter the 40 sample tone burst will be stretched from 40 sample periods to, 40 + 255 = 295 sample periods. This stretching, however, does not have uniform power, instead giving an profile reflective of the filter coefficients. This is no issue for infinitely continuous signals, which is what the sampling theorem assumes. The issue with audio is the signals are finite in duration. They are transient. The subjective audibility is, as I said above, disputed.
Bob Stuart and co. believe that the temporal stretching/blurring/smearing effect that's caused by use of sharply band limited filtering is subjectively audible with music. They believe it to be the cause of what has pejoratively come to be known as, digital sound. Their proposed solution is to utilize filters which minimally stretch the signal in time. Which necessarily means utilizing filters having relatively soft transition slopes. However, softly sloping anti-alias filters allow aliasing if they don't reach their stop band before hitting the Nyquist frequency. Meridian AES papers have referred to 3rd order quasi-Gaussian filters as being ideal in the required respects.
In my view, MQA is intended to set a transient optimized standard for anti-alias and anti-image filters from the recording end through the playback end of the chain. The purpose for including the ultrasonic band information, as I see it, is only to provide a wide transition band for the soft sloping time optimized filters to operate without incurring aliasing. As such, the information captured in the ultrasnic region can be made lossy, as it's only needed as spectrum to prevent aliasing, not for the fidelity of any otherwise humanly inaudible musical information contained there.
Of course, the very same could be done via DVD audio, for example, except that there is no set required filter standard for the objective of minimizing temporal blurring, which MQA master authentication does require, as I understand it.
2. To provide a singular content container which simultaneously holds low-rez, medium-rez and high-rez versions of the same program content, thus simplifying and minimizing the cost of what would otherwise require multiple distribution formats. Resolution in this context meaning, temporal resolution. The singular content container would inherently provide some copy protection for the hidden higher rez versions, as they would require MQA decoding in order to be accessed, and otherwise probably appear as dither-like psudo-noise. That's just a guess on my part.
3. Most technically intriguing to me is the de-blurring feature. I'm uncertain what is meant, exactly, but I take it to mean somehow removing the temporal spreading already encoded on non-MQA music recorded utilizing typically sharp anti-alias filters. There is a DSP process known as de-convolution, which can reverse the temporal stretching/blurring process of convolution utilized by FIR filters, but the actual process MQA utilizes remains an intriguing mystery to me. If anyone knows how MQA does this, I'd love to read how.
1. Minimization of temporal blurring, or smearing. This seems the overriding performance concern. Of course, the audibility of such smearing is controversial. It is, however, easy to see the smearing mechanism occuring within digital anti-alias and anti-image filters.
For the sake of illustration how, let's take 10 cycles of an Fs/4 tone burst to represent a transient and pass it through an 255 stage linear phase FIR filter. The 10 cycle burst will, on average, have a duration of, 4 samples/cycle = 40 sample periods. After passing through a 255 stage filter the 40 sample tone burst will be stretched from 40 sample periods to, 40 + 255 = 295 sample periods. This stretching, however, does not have uniform power, instead giving an profile reflective of the filter coefficients. This is no issue for infinitely continuous signals, which is what the sampling theorem assumes. The issue with audio is the signals are finite in duration. They are transient. The subjective audibility is, as I said above, disputed.
Bob Stuart and co. believe that the temporal stretching/blurring/smearing effect that's caused by use of sharply band limited filtering is subjectively audible with music. They believe it to be the cause of what has pejoratively come to be known as, digital sound. Their proposed solution is to utilize filters which minimally stretch the signal in time. Which necessarily means utilizing filters having relatively soft transition slopes. However, softly sloping anti-alias filters allow aliasing if they don't reach their stop band before hitting the Nyquist frequency. Meridian AES papers have referred to 3rd order quasi-Gaussian filters as being ideal in the required respects.
In my view, MQA is intended to set a transient optimized standard for anti-alias and anti-image filters from the recording end through the playback end of the chain. The purpose for including the ultrasonic band information, as I see it, is only to provide a wide transition band for the soft sloping time optimized filters to operate without incurring aliasing. As such, the information captured in the ultrasnic region can be made lossy, as it's only needed as spectrum to prevent aliasing, not for the fidelity of any otherwise humanly inaudible musical information contained there.
Of course, the very same could be done via DVD audio, for example, except that there is no set required filter standard for the objective of minimizing temporal blurring, which MQA master authentication does require, as I understand it.
2. To provide a singular content container which simultaneously holds low-rez, medium-rez and high-rez versions of the same program content, thus simplifying and minimizing the cost of what would otherwise require multiple distribution formats. Resolution in this context meaning, temporal resolution. The singular content container would inherently provide some copy protection for the hidden higher rez versions, as they would require MQA decoding in order to be accessed, and otherwise probably appear as dither-like psudo-noise. That's just a guess on my part.
3. Most technically intriguing to me is the de-blurring feature. I'm uncertain what is meant, exactly, but I take it to mean somehow removing the temporal spreading already encoded on non-MQA music recorded utilizing typically sharp anti-alias filters. There is a DSP process known as de-convolution, which can reverse the temporal stretching/blurring process of convolution utilized by FIR filters, but the actual process MQA utilizes remains an intriguing mystery to me. If anyone knows how MQA does this, I'd love to read how.
As I understood it, Meridian have spent a few years making a database of common and not so common A/D convertors used in studios and the temporal stretching/blurring of each one's filters.
When the audio is encoded by an MQA licence holder, the software/hardware (?) works out the ADC used in the recording of the material by a kind of fingerprint of this temporal stretching and encodes the audio in such a way that the standard MQA decoding will then compensate .
So I guess a studio that claims to be able to master in a MQA friendly way simply uses ADCs that are on the database and probably in a standardised resolution too.
I did though think that the main advantage of recording in high res like 384 or DXD (352?) was to avoid this very thing, using shallow filters at both ends whilst the aliasing is well out of the audible range? I guess, having 384kHz as a standard just isn't marketable yet and people take advantage via price hikes instead, incurring the usual "high res = a con" reaction
When the audio is encoded by an MQA licence holder, the software/hardware (?) works out the ADC used in the recording of the material by a kind of fingerprint of this temporal stretching and encodes the audio in such a way that the standard MQA decoding will then compensate .
So I guess a studio that claims to be able to master in a MQA friendly way simply uses ADCs that are on the database and probably in a standardised resolution too.
I did though think that the main advantage of recording in high res like 384 or DXD (352?) was to avoid this very thing, using shallow filters at both ends whilst the aliasing is well out of the audible range? I guess, having 384kHz as a standard just isn't marketable yet and people take advantage via price hikes instead, incurring the usual "high res = a con" reaction
Apparently Warner, Sony and Universal will have their entire back catalogues converted to MQA by the spring (only Warner have gone public so far) and several other MQA streaming services should come online later this year.
MQA and Warner: the Real Scoop | Stereophile.com
MQA and Warner: the Real Scoop | Stereophile.com
The problem I'd guess Meridian would have is that whether they like to or not, they have to get together with major labels in order to roll out the technology on any kind of scale. If it is ignored by a major label because they don't get a big enough slice of middlemen profits, they can kill any tech - such is their power, like a Mafia model.
Still, MQA has a limited life span determined by streaming bandwidths and media storage costs. Hardly worth the end user investing much in I would say unless they are a prolific consumer of new music?
Still, MQA has a limited life span determined by streaming bandwidths and media storage costs. Hardly worth the end user investing much in I would say unless they are a prolific consumer of new music?
Some more oil o the fire:
Archimago's Musings: COMPARISON: MQA "Authentication" & Sound Quality? (Mytek Brooklyn & Meridian Explorer2)
"And yes, no need to mince words. MQA is a form of DRM which not only insults end users, but also pisses off manufacturers. Linn is right IMO to call it an "outright land grab".".
It is a serious read, so if you only want to spend 30 seconds to get the lowdown about MQA, don't bother.
Jan
Archimago's Musings: COMPARISON: MQA "Authentication" & Sound Quality? (Mytek Brooklyn & Meridian Explorer2)
"And yes, no need to mince words. MQA is a form of DRM which not only insults end users, but also pisses off manufacturers. Linn is right IMO to call it an "outright land grab".".
It is a serious read, so if you only want to spend 30 seconds to get the lowdown about MQA, don't bother.
Jan
Is it much different to HDCD? Licenced technology and a little HDCD light would come on, inferring it was better if one thought that.. Ultimately it's in the head of the consumer.
I don't think many people will care because they're streaming and people these days are happy with Spotify over bluetooth 4.
Have to remember that Linn are a label AND rival hifi manufacturer.
I guess this:
MQA-Encoded CDs? Yes! | Stereophile.com
Although I'm stopping myself from replying "its a CD of mushy, probably cheesy music" ...
Well I guess we've had dither that crams 20 or 19 bits onto CD since the mid 90's.
I'm surprised they can find enough bandwidth in the noise to encode MQA into it.
Once the analogue stream, at its source, is sampled in time to create a digital representation, the harm is done that can not be corrected any more, no matter the measures undertaken later in a digital audio reproduction chain. Measurements tell us it's all good & perfect; our ears tell a different story altogether.
Many people are finally giving up on digital reproduction (after spending hundreds of thousands of dollars to get to a point where they are going to enjoy the music in the same way they enjoyed analogue turntable), and going back to reel-to-reel / turntables..... reel-to-reel is the latest buzzzz
It's not that much different than HDCD, no. It serves even less of a purpose with the availability of actual high-res PCM material. Tying it to a lossy codec is even more hilarious. It's a pure money grab that does not improve sound quality and is promoted by pseudoscience and lies.
This is not the first time Meridian has tried to pull this crap. They hurt DVD-A with their proprietary MLP codec as well.
The goal is similar to Sony's with SACD, too. To get paid from everyone in the chain. It will fail just like SACD because it is not actually audibly better than the same recording encoded in standard lossless PCM.
Last edited:
A form of torture. The fact that they're still wanking this BS is amazing.
Last edited:
Yeah, I mean just look at the title of that CD! Bleaghhh same old pap sung by again and again..
Oh, you didn't mean the music?
I'm not into MQA - neither an advocate, nor do I dismiss it. The full, uncompressed music is out there already and easy to find so no need for it in my world.
But I'm interested to see the same posters who also enjoy participating in a lot of other threads that involve saying things are a con/BS/money grab/conspiracy. And all arriving at the same time.
hehehe
That's only an observation, not a critism.
I don't think it is as black and white as being BS/money grab/conspiracy theory etc. If you run a business and tech firm that revolves around the reproduction of music then you'll still have a desire to work on new products and research and then have to justify that with a profit plan.. it's the same for any tech company and the same for any individual who has a mindset of invention.
To call it out as a con though needs some evidence presented to the discussion otherwise it loses any intellectual interaction and becomes a moaning match (like many other threads). It would be much more interesting to see a thread that systematically and logically, with evidence pulls MQA apart.
But I'm interested to see the same posters who also enjoy participating in a lot of other threads that involve saying things are a con/BS/money grab/conspiracy. And all arriving at the same time.
hehehe That's only an observation, not a critism.
I don't think it is as black and white as being BS/money grab/conspiracy theory etc. If you run a business and tech firm that revolves around the reproduction of music then you'll still have a desire to work on new products and research and then have to justify that with a profit plan.. it's the same for any tech company and the same for any individual who has a mindset of invention.
To call it out as a con though needs some evidence presented to the discussion otherwise it loses any intellectual interaction and becomes a moaning match (like many other threads). It would be much more interesting to see a thread that systematically and logically, with evidence pulls MQA apart.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.