MQA is in dire straits, the company announced today that it is in receivership (essentially a bankruptcy filing) and diligently seeking an investor.
MQA was once marketed as an audio format that would make “studio quality” achievable at home. Soon, however, questionable claims were found to have been made by its inventors. For example, the format turned out not to be – as first claimed – lossless: it does compress lossy (in a different way). Admittedly much less than something like the challenged mp3 of course. But still: a lie is a lie. It also quickly became apparent that MQA has DRM capabilities. It made the whole format a bit hazy. However, it was practical to save a lot of bandwidth when streaming audio. Tidal also saw this potential and adopted MQA as its own streaming format. The disadvantage was that you needed certified MQA DACs – whether or not integrated into an amplifier – to unlock the full potential of Tidal streams.
Not known to the general public
MQA never got through to the general public. Sure, the just slightly better mainstream hi-fi sets sometimes supported it. But the vast majority of buyers didn’t care, didn’t even know what it was. We’re also talking here about a group of users who stream primarily via Spotify and to whom things like lossy and lossless mean little to nothing. From the critical audiophile market, revenues were apparently lagging as well. It remains hard to sell that a bit perfect lossless FLAC or even PCM file would sound worse than an MQA. Meanwhile, MQA’s investor apparently also sees things a bit bleak in terms of the format’s future. And so the company behind the format announced that it is diligently looking for a new investor. And while doing so they had no other option than going into receivership; the UK equivalent of (the first stage of) bankruptcy.
Uncertain future
In theory it will be possibly find an investor who sees something in the not entirely uncontroversial file format. The question then arises: what do they intend to do with it? And what if this investor turns out to be Chinese? Given the amount of equipment that supports MQA and is connected to a network, that scenario could raise serious security and privacy issues. Then there are the tight agreements with manufacturers, will they have to pay more to continue supporting the feature? Or will the reins actually be loosened? Many questions and an uncertain future lie ahead. With the main question, asked from day one: is the world really waiting for yet another new – closed – audio format?
I think two things are being confused. Compression algorithms, like being used for ZIP archives, or for FLAC and ALAC file, or Column Store compression in analytical databases, all use the same mathematical principles in which data can be stored by (cleverly) rearranging recurring data and keeping an index of how this rearrangement is done, so the order can be restored once the data is retrieved.
This is why these algorithms are lossless, they do not throw away any information value, but store it in an efficient way. For more information on how that works, this article gives some explanation: https://www.howtogeek.com/362339/how-does-file-compression-work/
The fax example is an example of using this class of mathematical solutions.
Lossy compression, like the JPEG algorithm for images, or MP3 for music, does throw away the data. When the data is retrieved, it uses the knowledge that our human brains fill in the blanks of nearby information. Our brain is quite a powerful calculator when it comes to contextual reconstruction (so powerful that we can dream up stuff from incomplete context, hence conspiracy theories and all that). The data is lost forever once thrown away though, as compared with the source the data is not consistent and complete anymore.
If you notice a difference in the reconstruction of the data, comparing lossy or lossless compression, that’s up to the use case. Some people do see differences when comparing a TIFF file of a digital image and the JPEG version with moderate compression, other’s don’t. Our human capability to fill in the blanks differs per person, or the context in which this capability is used for the same person. Same is true for hearing and MP3.
Remember that Apple’s claim for using 256Mb AAC files in the iTunes Store when it was introduced was “that people won’t hear the difference with a CD”. That was a load of baloney, it was all about the storage size of iPods and the marketing you could store thousands of tracks on an iPod.
Over time, Apple added the capability to rip CDs in ALAC and AIFF instead of AAC only to iTunes and even when played back over the lousy speakers of a MacBook, the differences are audible.
The ‘big debate’ about MQA has always been about the mathematics used and reconstruction filters. Bob Stuart himself has been the instigator of this debate when the MQA format was introduced with all kinds of claims and his refusal to detail how it works and that debate is going on like a peat fire to this day.
Once people started to investigate it was clear that MQA uses a lossy compression algorithm. If that is bad or not is for everyone to discuss, just as the different Bluetooth codecs use variable lossy compression depending on the quality of the connection. For its intended use, that’s probably a smart decision. There isn’t a big debate about AptX or any of the Bluetooth specifications since those specs have been transparant from the beginning. I have the gut feeling that if the specs of MQA would have been transparant from the start, the discussion would never have derailed like it did.
ps: databases indexes are not an example of compression technique, the index is what it says it is : a vector to point where a certain value can be found in the data files, to speed up data search and retrieval in the database. As such, it doesn’t play a role in the consistency of the data stored, like lossy and lossless mathematical techniques do in storing the data.
Comments like yours dear Martijn are the reason why I keep on correcting wrong statements. You mourn about “confusion” and then mix up everything in the discussion to a cocktail of big misunderstandings. It is clear and obvious that methods like MP3 , AAC and similar codecs reduce the information “because signals covered by a strong signal before can not be detected usually” or ” most people wont hear a difference”. But that is not what MQA is about. If you read e.g. this explanation (https://bobtalks.co.uk/a-deeper-look/deeper-look-mqa-16b-in-the-last-mile/) you can see that there is nothing lost or replaced by guesses. And if you show me any instrument or voice with an overtone above 48 kHz and -20dB level I would agree with you that this is “lost”. But that does not exist, how creative makers of fuzz pedals or auto-destroying filters ever will get. So we talk about empty pages here. Can you show me ANY serious investigation which proved MQA to be lossy ? And pls spare me the totally broken trials of “Golden something” . They have been debunked years ago.
You are correcting something I haven’t written. I’m sorry, it ends here.
A better definition of lossless is that, after all the processing is over, the output digital signal is exactly the same as the original signal. Easily tested by a computer algorithm. Reinhard’s assertion that MQA throws away bits above 48KHz is therefore incompatible with his claim that it is lossless.
I am concerned that Pulse Code Modulated encoding is the default in these discussions. Think about Direct Stream Digital where each bit reflects whether the sound pressure level is above or below the previous level. DSD is bit-dense which would appear to make lossless compression difficult. For example, the sound of silence is a repeated 0101010101 pattern. However this can be run-length encoded as 5(01). The best example of run-length encoding that I can think of is the Group 4 fax standard which digitizes and run-length encodes a scanned line of paper. For written content, most of any line will be white with a few bits of black. This encoding trick is also used the build indexes in some database systems.
MQA is a system to help transporting music signals from source to end. I agree with your definition, if we add ” is exactly the same for music-, room-, and all other signals which are related to the content of the source”. Let me give you another example. If we have a pdf with 10 pages of scientific information and at the end 10 empty pages, have we lost something if we only send the first ten pages and skip the last 10?. Definitely not. Or only in eyes of sophistic “technical pure” information judgement. MQA does not transport empty content , because it is quite easy to understand that it does not make sense to do it. Your example of the fax machine is great, because already more than 50 years ago technicians – and I mean once with a brain and knowledge how to make use of it in best sensev- decided to “skip” unnecessary “empty information” . Thx for your comment.
You seem to have no clue about MQA, or less than a primary school kid. Why do you then dare to write about it? To call MQA not lossless, shows you never even read about how it works. Not acceptable for a technical oriented HiFi-blog. Meanwhile even the least informed could read a statement from Bob Stuart, that it is about selling of a part of MQA, which requires going into receivership. Are you now a section of the ill informed ASR forum, or have you founded a new MQA haters club?
MQA is not lossless. This has been investigated multiple times. It throws away bits, so it is not lossless by definition of the word.
Your aggressive comments are not ok. This isn’t the first time you use a tone which is bordering on uncivil. Either stop reading this website if it offends you so much, stop commenting, or adjust your tone of voice. We don’t like to remove comments, but there is a limit. Be warned.
Dear Reinhard,
Martijn already commented on your tone of voice. Why the anger? I interviewed Bob during the High End Munich. Fact is: MQA is NOT lossless. Especially so with high-res audio. It has also been proven several times. The definition of lossless is that no data is being altered or thrown away. MQA does both… even with 24 bit / 48 kHz the signal is being manipulated. We are not a MQA haters club… Why should we hate MQA? If Bob was being transparent from day-1 there wouldn’t be an issue at all. But if you claim stuff like “Studio Quality”, Lossless Studio Quality and “Master Quality Authenticated”… well… That is just not true…
And please, just keep it civil. We do as well. If you don’t like Alpha Audio, please just delete your account and ignore us. No hard feelings. At all.
Dear Jaap and Martin. I wonder if you really have spoken with Bob Stuart about the MQA process, because I definitely have. And the truth is, that MQA is not changing anything up to 48 kHz bandwidth. Above 48 kHz MQA is “throwing away” bits which can not contain musical signal because there are no music signals with very high level in the region between 48 kHz and 384 kHz . Overtones and reflections are never high level. So instead of transporting 0s they are ignored. I don’t call that lossy , but clever as a technician. Only people like Golden something , who tried to “measure” completely idiotic test-signals would call that lossy. AND yes, I will always shout out loud against power line promotion or wrong technical statements for my whole HiFi life. But maybe not here again, because after stopping my patreonship I may continue to stop reading. Your blog seemed promising, but looses a lot over time.
Like I said: I had an interview with Bob Stuart himself. He never gave a clear answer to my question: is MQA losless. In stead he avoided the question and started a new topic about cleaning windows.
But hey: opinions about products and technology may differ. If you like MQA… use it. It’s a free world.
Here is our interview with Bob Stuart, it can easily be found on YouTube: https://www.youtube.com/watch?v=3r_VRxcwODI. We have had several talks with him and his staff of MQA.
At Alpha-Audio, we strive to be respectful to the brands and its people. We appreciate that you as one of our readers to do the same. Your words of us having ‘no clue’ do not fit into these principles which we kindly ask you to comply with. Thank you.