I don't encourage anyone to read anything from Brian Lucey on any listening tests results. I know because I have discussed it with him. See this starting post for example on what he has written:
https://gearspace.com/board/mastering-forum/1171365-mqa-discussion-denver-rmaf-10.html#post12952542
"
Where as mastered for iTunes is harmonically cold and loses some low volume/low end information, actually altering the groove to make everything sound like a nerdy white wedding band, MQA brightens the high-mids in the Mid section while thinning the low-mids on the Sides. "
You believe this?
Harmonically cold? MQA brightens the high-mids? Thinning low-mids? You really believe this kind of subjectivist nonsense? I sure hope not.
He needs to start learning to use blind controlled test before he speaks on any topic like this.
This is the problem with MQA. People become completely different when it comes to this topic. If I had brought up those quotes in any other context, people would throw stones at Brian Lucey. Have him fight with MQA and all of a sudden he is a beacon of audio science when it comes to matters related to MQA. Don't do that please.
Before telling you what I got by listening to MQA, I've read your post, and I'm sorry to not be able to tell anything on iTunes masters as I don't use it, so you may be right on this one.
Now, I've also read the two posts above, and think that when Brian Lucey made a mistake by saying "
Let’s just sell the 24 bit files at the mastering session sample rate " which can lead us to an error thinking it all depends on mastering, which is wrong, because mastering itself can depends on recording and mixing sessions.
So when you asked why 24bit and not 16bit, it's not a problem of dynamic of the stereo file while mastering, it's because 24bit was needed for recording/mixing, and by that, it's kept while mastering (and can also be needed as some gears in mastering will only work at 24bit).
By doing use of 24/44.1, he respects the orginal bit depth and sample rate, and it's more easy to do it now than before because some plugins that were working better with higher SR are now working better at lower SR while internal oversampling themselves.
Unless I didn't understand your question, you can't tell him 16bit would be enough while the fact to keep 24bit goes in the sense of what everybody is asking for : the real file created in the studio.
What I get is that you both started to battle with your "extreme" experiences that don't share something on this case, instead of trying to find what you have in common that could help. We can't forget that we need both of your works to be all able to listen music
If I supposed that it was a battle, it's because you only respond to my post quoting this part, but not my question, and not the part where I said that Lucey have said somewhere that used in the whole chain from recording sessions, MQA may be an intresting technology.
I don't remember where he say that, but the way I read what's come from your link, he's not talking about the encoding they choose by itself, but about what comes out and you listen, and the difference may not come from the encoding but more from the processing while decoding and up to the filter in the DAC.
I'm sorry to say this, as I know your experience and have all the respect for your review, but you were not more perfect than him by assuming he would not tell the difference between ACC 256 and CD. It may be right, it may be wrong, but I can confirm you that on tracks I know perfectly, but on some kind of music and not all, I get a 100% result in ABX even after 100 tests between AAC 320 and FLAC, so I would think he may be able to.
And on the same track, doing the test between FLAC and MQA, I can find which one it is.
So back to MQA listening :
I may not have use the same words, but I will tell you what I got and I'm waiting for what some other sound engineers think (I've been lucky to work in recording studios even if worked more on live events) :
- in the first listening experience, I was kind of excited, but I couldn't tell if it was an improvement because it was pleasant on some tracks, not on others
- after more and more listening (and I don't only use Tidal, but also Qobuz and Amazon HD), the main thing I heard, still on some tracks only, is the pre-ringing change.
A track part with mainly a bass and drums is clearly giving me the impression of : more dry, bigger attack, and I started to imagine the use of a transient exciter, which actually is not needded if you can kill some pre-ringing as it gives a bit the same impression than increasing transient by deleting what can "cover" it (it's not really covering it, but the fact that hearing another thing before any transient will make you hear less clearly the transient itself coming after that).
But, I started to think that you can do a bit of that while creating the master (and it made me remember that the company that created Soundscope had also a software to erase the pre-ringing of any track), and so it was not really comparable to a PCM file because MQA looks more like encoding+processing and not encoding only, and by that,
I started to ask myslef if we can call it "what the artist want you to hear"
But I don't want to take a full conclusion based only on what I listen, I want the truth in measurement also.
Now, regarding test, if it can help, I found a way to record in the digital domain a MQA stream, and get a FLAC file that contains the MQA tag (by recording, no piracy of files). I need to check that recording method is right by doing the same with 2L free tracks and check if the recorded file is the same than the original one. Do you think that it's they are the same, it can confirm that the recording method is right ?
It could help to get parts of any song that would be ideal to test, and not be limited on 2L tracks.
PS : I'm still waiting for you to confirm or not if the OP file that got a blue light, while it was a "not acceptable file" with error in encoder log shows that the Blue light can be lying to the customers 
