Technically, it is lossy, considered the file size is compressed almost by 12:1 ratio (compared to 1.7:1 in FLAC) when compared to a DXD 24/384k FLAC. Moreover the bit rate is capped at 1.5Mbps and not higher or match original bit rate of DXD.
Bob Stuart’s claim of ‘lossless’ is only applied to subjective listening and not the technical aspect of the MQA. He also claimed that ‘lossless’ is difficult to achieve when convert, mix down and when DSP is applied during mastering.
Not quite. MP3 aims to remove stuff that the typical listener cannot detect or isn’t bothered by. MQA is based on an analysis of parts of the frequency/amplitude space in which there is no signal, such as very high amplitude at high frequency or very low amplitude everywhere. This is an objective analysis.
So it is lossy in the data transfer sense that you cannot recover an arbitrary signal after MQA transfer, but it is lossless in the music transfer sense that you can recover a music signal. Thus, you cannot use MQA to transfer the binary files that make up Microsoft Office, which is not the purpose. This has been well discussed for years.
Looked at the Audiogon debate.
“While I haven’t been able to do this comparison myself, reading a highly regarded golden ear admit this in print is warning enough for me.”
“I haven’t heard MQA so I shouldn’t comment, but of course that won’t stop me… My present setup sounds good enough to me so I won’t be checking out the Emperor’s New Clothes”
“I’ve yet to hear it but will at the LA Audio Show next month but some very respected innovators…”
“I have not heard MQA but it is…”
Sigh.
Bob Stuart nor MQA (the company) claim the MQA process is lossless.
From the very beginning, they’ve acknowledged it’s a lossy process that can be delivered in FLAC / ALAC / WAF / AIFF (i.e., “lossless”) file formats.
Some people apparently misunderstood the use of lossless file formats to mean that the process itself was lossless, which it clearly is not - though the parts of the signal that are discarded are obviously completely inaudible; i.e., “empty space” above 22k and below the noise floor below 22k.
Looked at rocknerd:
“Before anything else, we must stress yet again: high-resolution audio makes zero difference at best and causes intermodulation distortion at worst, and even the golden ears advocating this stuff can’t tell the difference from CD quality in A/B/X testing. Hi-res audio is completely made-up ■■■■■■■■■ and anyone trying to sell it to end consumers is committing fraud. The only advantage of hi-res audio is that it tends not to have LOUDNESS WARS mixes.”
Since higher-than-CD-frequencies are without value, it doesn’t really matter if they are transmitted in a lossy way, does it? That’s a relief, we can stop all this debate.
I’m not sure any bringing in any debate from other forums is helpful. There is no right of reply, and it’s not as if there’s any shortage of discussion here.
Exactly, the process of encoding and decoding of MQA is lossy but the process of delivering via FLAC or ALAC wrappers are lossless. So there’s are generally some misunderstandings that MQA itself is lossless.
Just more about seeing what other audio lovers in a different community have to say about it. It’s interesting to see the same undercurrents of pro- vs. anti- MQA going on there.
Guy you keep saying that MQA is not lossless (which is true), but do you acknowledge that the data which is lost cannot be heard and therefore the file contains the same audible content as a lossless file? I think people have been trying to explain this to you when you were first excited about MQA being available on Tidal back in December 2015.
Cheers
Tom
The same could be argued for numerous other formats, such as AAC, MP3, AC-3, DTS, etc.
AJ
No, it couldn’t - at least, not honestly.
Those are all audibly lossy processes that punch large swaths of data right from the middle of the audible frequency range. When properly implemented, MQA does not change anything below 22k, and this can be shown in spectra analysis.
Au contraire. For many people, many so called “lossy” codecs are audibly lossless at certain bit rate thresholds in properly controlled tests. And if you have not participated in any controlled testing, then you do not know what is audibly “lossy” or lossless for you.
AJ
The are also many people out there who claimed and have misleaded it otherwise. I myself have listening to fully decoded MQA and I must said it does sound better than Redbook. I’m not anti-MQA, over the years I’ve softened my stand and come to accept benefits and limitations of this format, mainly through listening and many read up.
What I don’t like is the hyper marketing tactics behind the format; it have brought a lot of confusion, heated arguments and many unsustainable claims not just here but out there. Limitations of software decoding that requires to purchase a new equipment plus licensing deals from the very start of the recording, editing, distribution, manufacturers of new equipment and finally, we consumers will probably need to absorb that additional cost.
As a Tidal Hi-Fi subscriber, I’m looking forward for Roon to implement MQA decoding (hopefully more than 2x), this will open up a lot opportunity for comparison and listening tests instead of relying on unsustainable claims out there.
I was talking about the method of compression, which is completely different between the formats you listed (which all remove data from the audible spectrum) and MQA (which does not).
I should not have included the word “audibly” in my response, as whether or not the compression is audible is irrelevant to the actual method of compression.
MQA truncates 24 bit inputs to 17 bit in order to append its own coding within those LSBs. Now, some say that they can hear audible benefits from 24 bit word length. Regardless of audibility, truncation does remove data and cause some loss throughout the passband, including below 22 kHz. In that sense, MQA is “lossy” and similar to the aforementioned codecs.
What makes MQA different is its “audio origami” bandwidth extension – while maintaining backward compatibility. In that sense, it is somewhat comparable to the DTS 96/24 codec.
AJ
Is MQA lossless? Read this for a starter.
What is the bit depth of MQA? Read this and see the examples that follow to see the bit depth of MQA with and without a decoder.
I don’t buy this. It’s just sloppy journalese. As a statement it’s like saying “if you haven’t participated in any controlled testing, then you don’t know if you can distringuish red and blue”. While it may be true, it does not have to be true. Many people believe they can distinguish between red and blue, and are correct in this belief, without controlled testing. It can be the same in audio.
Of course if someone states they can tell the difference, you don’t have to believe them. But you can’t say they can’t be right just because they haven’t submitted themselves to “controlled testing”. It’s not logical.
Thanks Guy I. Starting to believe that you’re not a industry insider and that your are a politician:slight_smile:
You managed to avoid then question altogether.
I have no problem with lossy. If good choices are made then data can be removed without affecting the perceived audio. For most listeners somewhere between 174kbps and 245 kbps is transparent for the LAME mp3 encoder using VBR (see https://en.wikipedia.org/wiki/Transparency_(data_compression) ).
The answer given to the question © MQA prevents us changing the signal before playback is the one that gives me the most concern. However good (or bad) individuals judge the MQA treatment of audio, being able to do room correction (and any other desired signal processing) will yield far larger benefit than anything MQA can do. Of course MQA cannot admit that so they give the cop-out answer that they do. It has been the over-hyping and marketing more than anything which has got my hackles up about MQA. If they had come across more openly about the technology I may have applauded them for its benefits rather than feeling distaste over the way they deflect from the compromises they inevitably had to take to devise this technology.
Having read the MQA threads on the hydrogenaudio forum where hardly a positive word has been said about MQA, I find myself leaning much further into the anti-MQA camp.