[Danjen22]

Does the box say MQA on it? Is it a UK model?

Either way, this could be just their attempt to explain away why something isn't working.

It is unfortunate that LG never provided any confirmation of MQA authorization: We're all just guessing whether it picks up the MQA and how it unfolds it.

Yes, the box states MQA on the back and it's a UK model LMV500EM. You'd think their technical team would know these things, lol.

 

[Beatlelarry]

Hello. I bought an LG G7 to use it mainly as a music player. It is currently with Android 8, but the update to Android 10 appears to download. If I upgrade, does it lose the ability to play mqa files? Basically, I use the Tidal app, both streaming and offline. In streaming, I use Tidal through the USB Player Pro application. In short: should I update to Android 10, as well as the updates that appear to me from Tidal and USB Player Pro, or do I stay on Android 8? Thanks.

 

[Dannemand]

Hello. I bought an LG G7 to use it mainly as a music player. It is currently with Android 8, but the update to Android 10 appears to download. If I upgrade, does it lose the ability to play mqa files? Basically, I use the Tidal app, both streaming and offline. In streaming, I use Tidal through the USB Player Pro application. In short: should I update to Android 10, as well as the updates that appear to me from Tidal and USB Player Pro, or do I stay on Android 8? Thanks.

I believe MQA survives the update to Android 10 as long as you use UAPP as player. So that's the best answer.

Tidal app as player has not worked with MQA since version 2.21.0 (over a year ago). It sends the music through the Android mixer, which garbles it -- and loses MQA in the process. If you roll back to Tidal app 2.21.0 (as described earlier in this thread) it should avoid the mixer (and thus still work with MQA) but ONLY as long as you avoid Android 10.

I don't have firsthand experience with G7, though. We have reports that V35 DOES work correctly even after updating to Android 10. It's possible that is also true for G7. But you still have to use Tidal app 2.21.0 -- or UAPP.

 

[Beatlelarry]

I believe MQA survives the update to Android 10 as long as you use UAPP as player. So that's the best answer.

Tidal app as player has not worked with MQA since version 2.21.0 (over a year ago). It sends the music through the Android mixer, which garbles it -- and loses MQA in the process. If you roll back to Tidal app 2.21.0 (as described earlier in this thread) it should avoid the mixer (and thus still work with MQA) but ONLY as long as you avoid Android 10.

I don't have firsthand experience with G7, though. We have reports that V35 DOES work correctly even after updating to Android 10. It's possible that is also true for G7. But you still have to use Tidal app 2.21.0 -- or UAPP.

Perfect, thank you very much. As I use Tidal offline a lot, I think the best solution will be to stay on Android 8 and install version 2.21.0.

 

[Dannemand]

Perfect, thank you very much. As I use Tidal offline a lot, I think the best solution will be to stay on Android 8 and install version 2.21.0.

I agree!

 

[Danjen22]

Yes, the box states MQA on the back and it's a UK model LMV500EM. You'd think their technical team would know these things, lol.

Last reply from LG regarding MQA.

Good Morning, Mr



Thank you for your email.


We apologise if you have experienced miscommunication in regards to the MQA audio format. We can confirm, that it was on the device at launch. When the phones were updated to Android 10 this has caused MQA to stop working.

This is still being looked into on by us if anything can be done to resolve it, but it does look like some limitations have been put in place with the new android version. We would have to work with ‘google’ to try and resolve if possible. Due to this we are not able to provide timeframe when this issue will be resolved.


Should you need to contact us please call our Customer Support Team on 0344 8475454 and quote reference number



Kind regards,


Guess that's that then, no surprise.

 

[Dannemand]

@Danjen22: Thank you for sharing this. That sucks. But at least we now have official acknowledgement from LG that it's broken as of Android 10. Maybe that's also why V60 doesn't have MQA: They accidentally broke it in Android 10.

 

[Dannemand]

I realized I hadn't performed audio_flinger tests on the last few Tidal app versions. Since I don't plan to extend my subscription later this month, I thought I'd better test it now. (I only listen to Tidal 1-2 hours a week, so cannot justify the cost.)

But there is nothing new as of Tidal app version 2.39.0: Everything goes through the mixer: 16/44 (CD), 16/44 (MQA), 24/44 (MQA), 24/48 (MQA), 24/192 (MQA) 24/352 (MQA), it's all resampled to 16/48. Artifacts introduced, MQA lost.

Tidal app version 2.21.0 still uses Direct path for all MQA tracks (including the millions of 16/44 Warner albums) so I recommend using that when you need offline. As noted a few posts up, this too is broken once you go to Android 10 (V50, V60, G8). But for V30 with Oreo or Pie you're still good. And V35 too, even on Android 10 (tested by @csglinux). We need reports whether G7 still works on Android 10 with Tidal app 2.21.0; it might, as it is of same generation as V35.

Of course UAPP still just works: It uses Direct path with anything I throw at it, and thus remains my recommended player when Tidal offline isn't required.

I also tested HiBy Player this time, since I had previously installed it to test with an external USB DAC (where it DOES bypass the mixer). But it too fails on Quad DAC and plays through the mixer. (One note for those using Bluetooth on DACs or headphones/IEMs: HiBy implements the new UAT codec. I didn't test it, as I no longer have a DAC that supports UAT. But worthy of mention.)

I haven't attached the audio_flinger dump files, as there is nothing to see beyond what I reported here. I am happy to upload them if anybody is interested

 

[csglinux]

For any that still doubt paying for MQA is like hiring someone to rob you:

 

[Dannemand]

For any that still doubt paying for MQA is like hiring someone to rob you:

I saw that one, and it is majorly flawed. It is almost as misleading as the claims from MQA themselves. If not more.

The first clue is early in the video where he admits that the MQA encoder refused to encode the synthetic sinusoids and square waves, which contain signals far outside the envelope which MQA was designed to encode. He had to bury them in the middle of music tracks and thus trick the encoder into performing at its worst. Which I am sure was his intention.

If you read the response from MQA, they state that 11 of his 14 tracks displayed warnings from the encoder for this very reason, although the author claims he never saw those warnings. Maybe Tidal's implementation didn't display them, which is a flaw on their part.

The MQA encoder was obviously written with the assumption that studios use it to create "best effort" output. But with this "public" version of the encoder now on Tidal, it appears they will have to add further tests to catch malicious attempts to create "worst effort" output.

Before GoldenOne had posted his video, he had advertised its coming in numerous threads across the internet (remember he runs a YouTube business). And the outcome was obvious already then: MQA assumes that music is recorded with microphones and digitized with ADCs, both of which introduce low-level noise that takes up a significant space in a FLAC file. MQA appropriates that space to store higher rate sample info (its much ballyhooed lossy part). That is obviously going to fail if you feed it a synthetic signal then expect to find those bits intact in the output. I understand that the encoder has a specific setting for synthetic music, which would have been the one to use in these tests, if one were looking for "best effort" output.

Some of the analysis performed by GoldenOne was interesting, for example the downwards aliasing of ultrasonics. But he should have performed his analysis on "best effort" content, such as the 2L samples which provide many DXD and MQA tracks for public scrutiny, instead of creating these "gotcha" examples. You did that in the past, finding little difference between the MQA and the HiRes FLAC. I am sure that never occurred to him...

Another legitimate issue he raises is that 24-bit MQAs can be truncated to 16-bit and still authenticate. I've been scratching my head over that as well, although it is documented by MQA as a special permitted case (I've linked to their explanation of that in the past on head-fi). But again, GoldenOne ignored that and instead concluded that MQA files can be molested to your hearts content and still authenticate, and thus the authentication claim is just fake.

GoldenOne created a large thread about this on ASR -- where he is now the hero, despite his otherwise subjectivist credentials. The moderators even banned the few members who dared challenge his findings with very specific technical counterpoints. So that thread is now a 500-post echo chamber of MQA venom.

As I've posted several times on this forum, I wished MQA didn't exist and that studios would willingly share their HiRes FLAC crown jewels. But that is not the world in which we live. As it is, music consumers would be better served if MQA would be evaluated on its actual merits, not on claims of technical deficiency fabricated by crusaders who hope they can make it go away.

 

[csglinux]

I saw that one, and it is majorly flawed. It is almost as misleading as the claims from MQA themselves. If not more.

The first clue is early in the video where he admits that the MQA encoder refused to encode the synthetic sinusoids and square waves, which contain signals far outside the envelope which MQA was designed to encode. He had to bury them in the middle of music tracks and thus trick the encoder into performing at its worst. Which I am sure was his intention.

If you read the response from MQA, they state that 11 of his 14 tracks displayed warnings from the encoder for this very reason, although the author claims he never saw those warnings. Maybe Tidal's implementation didn't display them, which is a flaw on their part.

The MQA encoder was obviously written with the assumption that studios use it to create "best effort" output. But with this "public" version of the encoder now on Tidal, it appears they will have to add further tests to catch malicious attempts to create "worst effort" output.

Before GoldenOne had posted his video, he had advertised its coming in numerous threads across the internet (remember he runs a YouTube business). And the outcome was obvious already then: MQA assumes that music is recorded with microphones and digitized with ADCs, both of which introduce low-level noise that takes up a significant space in a FLAC file. MQA appropriates that space to store higher rate sample info (its much ballyhooed lossy part). That is obviously going to fail if you feed it a synthetic signal then expect to find those bits intact in the output. I understand that the encoder has a specific setting for synthetic music, which would have been the one to use in these tests, if one were looking for "best effort" output.

Some of the analysis performed by GoldenOne was interesting, for example the downwards aliasing of ultrasonics. But he should have performed his analysis on "best effort" content, such as the 2L samples which provide many DXD and MQA tracks for public scrutiny, instead of creating these "gotcha" examples. You did that in the past, finding little difference between the MQA and the HiRes FLAC. I am sure that never occurred to him...

Another legitimate issue he raises is that 24-bit MQAs can be truncated to 16-bit and still authenticate. I've been scratching my head over that as well, although it is documented by MQA as a special permitted case (I've linked to their explanation of that in the past on head-fi). But again, GoldenOne ignored that and instead concluded that MQA files can be molested to your hearts content and still authenticate, and thus the authentication claim is just fake.

GoldenOne created a large thread about this on ASR -- where he is now the hero, despite his otherwise subjectivist credentials. The moderators even banned the few members who dared challenge his findings with very specific technical counterpoints. So that thread is now a 500-post echo chamber of MQA venom.

As I've posted several times on this forum, I wished MQA didn't exist and that studios would willingly share their HiRes FLAC crown jewels. But that is not the world in which we live. As it is, music consumers would be better served if MQA would be evaluated on its actual merits, not on claims of technical deficiency fabricated by crusaders who hope they can make it go away.

I've never found myself persuaded by the argument that we shouldn't and/or don't need to consider synthetic test signals because they don't exist in real life. What if I want to make a symphonic-electro-pop song? Besides, worst-case scenarios are what all good engineers should worry about. Then we know we're covered for every other scenario. Even if this guy's intention was indeed to trick MQA, he still makes many valid points. Think about the flip side - how would you 'trick' a standard FLAC encoding? The concept of separate encoders for bandwidth-limited 'real' music signals and electronic music/t-signals should be a red flag. Which encoder would it use for my symphonic-electro-pop track?

I need to confess that the tests I ran were only to generate a spectrogram from a very short history to confirm that UAPP was unfolding. The extent of my testing was only that the spectrograms looked reasonably similar to the eye. (And we all acknowledge MQA can unfold content above 22 kHz.) But imagine I toggle on/off some signal 20 dB below an 80 dB fundamental. That part of the graph now reads 80.04 dB and you wouldn't see that on a spectrogram, so nobody should assume the tests I ran in any way corroborate the accuracy of MQA encoding. The fact that this guy was able to sneak in an MQA encoding of a complete set of RMAA test signals is freaking awesome. I don't know the reasons why MQA would refuse to encode t-signals that weren't wrapped in something to disguise it as conventional music, but I can't presume their motives are scientifically or ethically pure. Davy and I both had an interest in getting sweep files MQA encoded. I can personally attest to MQA's modus operandi with any technical questions - wait three weeks and then show the middle finger. I hope this guy still has his original and MQA encoded files for others to play with? If so, this looks like a slam dunk.

Even if you're a believer in the teachings of Bob Stuart, I don't see there's much of a compelling reason for LG-V-series owners to stick with Tidal anymore. If we have to live with a music streaming service who's offline content will only play back through the Android mixer at 48 kHz, Qobuz can be had for $5 less per month (actually, I believe there's currently an offer where it's $7.50 less/month for the first year).

P.S. You have to love those survey result that showed 66% of people polled thought MQA's sound quality was as good as, or better than, 192 kHz FLAC

P.P.S. My mistake - if you have a V30, you still have @Dannemand's cool roll-back trick for Tidal. Newer Android 10/11 devices such as the V60 seem to have run out of luck with Tidal though as everything, including Masters, is now up-sampled.

 

[Dannemand]

Once again, as I understand, the MQA encoder will accept signals outside the envelope if you select that mode. I assume in that case it will work more like regular FLAC. And that would be the appropriate mode to use when testing its performance on synthetic signals.

But by ignoring the characteristics that govern the remaining 99.99% of music, we're wasting space in FLAC to bit-perfectly preserve noise. MQA chose to use that space for information which they claim contributes more to sound quality. One can disagree with that trade-off, but that was always the official premise of the codec.

Hence testing it by tricking it into performing one type of encoding while feeding it a different kind of content, and publishing the results as an example of MQA's inferiority, does not reveal malice nor deficiency on the part of MQA, but rather malice and dishonesty on the part of the tester.

This impression is further strengthened when the tester gets paid for every view of his video, and has promoted it as heavily as he has. He certainly knows how to stroke the pleasure center of his MQA hating audience. I bet it is by far the most profitable video yet on his YouTube channel.

I agree that Qobuz has become more competitive against Tidal price wise and does entirely avoid the complexity of MQA. Does their app work now to bypass the mixer on our LG phones -- which it didn't last time I tested it? But then, Tidal has since lost their LG their compatibility (as discussed earlier in this thread). So it's UAPP regardless of service.

Personally I have no strong attachment to MQA, nor against it. I think people should choose the service that has the features and catalog they like. Of course if one decides to crusade for or against MQA, that will determine one's choice. But I don't see anything in this video convincingly demonstrating inferior sound quality on MQA's part for properly encoded content. If the author would do an equally "thorough" test on the 2L tracks or other "best effort" content -- showing the best MQA can deliver, instead of only showing the worst it can be tricked to deliver -- that would convince me a whole lot more.

And that is as much time as I will spend on that debate. I had already decided a few weeks ago to stay out of it, when GoldenOne first began spamming the threads to promote his upcoming video. And yet I couldn't help myself when it spread to this thread, too. So done for my part

Edit: Just saw the updates to @csglinux' post. Let us be clear that there is a lot of poorly recorded junk albums on Tidal that are now in MQA, which doesn't make them any less junky. But there are also some excellent recordings, which are available as HiRes MQA. I have previously mentioned the Everest Records 3-track 35mm tape transfers, which you can buy on HDtracks as 24/192, or you can play them on Tidal as double-folded 24/192 MQA. I am completely convinced that they are not fake upsamples. They are 60 years old, and some are showing their age, while other are still excellent. That said, Tidal used to have plain 16/44 variants as well, which they no longer do. Choosing the HiFi quality setting will just give a downsampled version, which contains MQA leftovers, but no longer decodes or authenticates. I do find that most regrettable -- and foolish on Tidal's part -- as it prevents the user from getting a clean non-MQA version.

 

[monsieurfromag3]

That said, Tidal used to have plain 16/44 variants as well, which they no longer do. Choosing the HiFi quality setting will just give a downsampled version, which contains MQA leftovers, but no longer decodes or authenticates. I do find that most regrettable -- and foolish on Tidal's part -- as it prevents the user from getting a clean non-MQA version.

That is increasingly my problem with Tidal. While I do have MQA-compatible sources, up to 16x, and bought the MQA add-on for my phone, I still don’t care about high resolution - too many complications for something that can’t properly be experienced outside of a live setting. Having to contend with the dubious commercial and technical practices of MQA, and even when opting out getting the truncated, next-best version of an MQA file, makes me feel trapped. MQA feels essentially like a trap, trying to take over playback selling at least a measure of snake oil. I’d say roll on Spotify Hi-Fi, and hope they don’t use MQA, but the truth is they will never work with UAPP, and I don’t much care about FLAC vs mp3 as long as it’s mangled by Android’s crapshoot of an upsampling.

 

[Irish1966]

Confused newbie with a couple of questions please. I have just received a V30+. My only music sources will be Amazon Music app (not hd) and my old CD collection which I recently digitised mostly to FLAC and some mp3 246kbps. Will I benefit from something like Neutron or UAPP (easier interface), or poweramp. All of this is new to me and while I've read through the thread, terms like bit rates, sampling, DSD, MQA etc. are foreign to me and I'm just getting more confused the more I read ! So, just in relation to my specific and limited needs, could someone please let me know what they would recommend. Thanks in advance.

 

[csglinux]

Hi @Irish1966 - welcome and congrats on your nice new phone!

Your FLAC and mp3 collections will both benefit from the use of Neutron or UAPP. That's not to say you'll alway hear the errors if you don't use these apps, but both the apps mentioned can play back your tracks without putting them through the standard shredder/mangler otherwise known as the 'Android mixer'. In layman's terms - pretty much every modern Android phone slightly destroys every 44 kHz FLAC or mp3 file - unless you use Neutron or UAPP.

Amazon Music is a strange sort of half-way house. It seems that (probably because of Android's stupid 48 kHz default sample rate) Amazon decided that they may as well put all their music out at 48 kHz. That's not great, but it's slightly better than going through Android's mixer, because Amazon use a higher-precision, i.e., better quality, up-sampler. (At one point the UAPP developer was in talks with Amazon to see about integrating it into UAPP. My guess is nothing happened because the integration would have been pointless, i.e., once Amazon have up-sampled everything, there's no way back to the original, unadulterated bit-perfect 44 kHz version.)

Worst case scenario, your V30+ should still sound better than all other non-LG phones