[qboogie]

I have no personal experience with Pixel 4a, but I seem to recall a statement from the UAPP dev that there is a problem with Google's latest libraries that prevents UAPP from playing bit-perfect on it, so they (the UAPP dev) pulled it from the Pixel 4a compatibility list until it has been fixed.

I cannot seem to locate where I read that, I'll keep looking. If I find it I'll link it here.

Edit: OK, it was just yesterday in this thread, and it is the Pixel 4 series:

https://www.head-fi.org/threads/usb...ort-for-android.704065/page-301#post-16221008

And it isn't just about bit-perfect, UAPP (and other apps using their own driver) actually crash the phone when connecting a USB DAC. I can see the dilemma if someone wants to use UAPP without a USB DAC. You should email the Dev about this.

Thanks for the reply. I'll wait until the problem is fixed.

 

[Bobbetybob]

Wow, the Google situation is a bit of a mess, I have a 4a (regular 4G version) and it seems to work perfectly with UAPP.

 

[mw7485]

@Davy Wentzler - just reaching out to see if this is by design, and if not, can it be fixed.

When I access a DLNA source (in my case Plex), and go into folder view, I can see the scroll bar on the right, but try as I may I can't touch on the bar and fast scroll through the folder list. I have to continually and manually flick up the screen to scroll.

Can this be fixed?

I pointed this out to Davy a couple of weeks ago but didn't hear back. The Network view has a similarly skinny and useless vertical scroll bar, so there is a lot of finger swiping for my Samba shares.

 

[kickss]

Please help me! Can anybody show me how to get rid of black screen popup (rendered with Title, Next Pause Previous button) when I press Next / Previous on my DAP (screen off). Thanks in advance

 

[gLer]

I pointed this out to Davy a couple of weeks ago but didn't hear back. The Network view has a similarly skinny and useless vertical scroll bar, so there is a lot of finger swiping for my Samba shares.

Ah yes, in Network View too. Hopefully he'll come up with a fix.

 

[alleroy]

After reading last pages of this thread I am a little confused: has any sense to use UAPP with a DAC which has not hardware support for MQA, or does the software unfold provided by UAPP still a good job in this sense?

 

[Dannemand]

After reading last pages of this thread I am a little confused: has any sense to use UAPP with a DAC which has not hardware support for MQA, or does the software unfold provided by UAPP still a good job in this sense?

UAPP is fantastic for plain USB DACs and internal HiRes DACs: It plays both 16/44 PCM (CD quality) and HiRes bit-perfect to the DAC, where most other players will go through the Android mixer; It also plays DSD in whichever format your DAC accepts (native, DoP or converted to PCM); AND it decodes (unfolds) MQA and plays it on the USB DAC, internal DAC or through Bluetooth.

And it does all this playing local files, from network servers (NAS or DLNA) or streamed from Qobuz or Tidal.

The additional unfolding of MQA performed by MQA DACs (so-called rendering) is just a trifle, which you aren't likely to hear anyway. (Some would say you don't even want it, but that's another discussion.)

I've used UAPP for years on various devices, and recommend it highly. It should be in any audiophile Android users arsenal.

Trial can be downloaded on their website.

 

[Currawong]

IOW if "real" Master tracks (i.e. ones with additional samples to unfold) play correctly on your DAC after UAPP has performed first unfold, but those 44.1 and 48KHz tracks DON'T, then I think it's a sign that UAPP's upsampling of them is what's breaking MQA.

Whether a track is MQA or not is indicated by certain bits set in the output. This has been proven by removing the HF content from an MQA file, but keeping the indicator bits, causing an MQA DAC to indicate that the file is MQA. So UAPP could be removing those bits. Going by how Roon works though, those bits can be passed on after the initial unfold is done in software. However, note, most importantly, there are NO MQA tracks with 2nd or 3rd or any other unfolds on TIDAL. None!

After reading last pages of this thread I am a little confused: has any sense to use UAPP with a DAC which has not hardware support for MQA, or does the software unfold provided by UAPP still a good job in this sense?

The best option is, you're a TIDAL user is: Cancel your subscription and switch to Qobuz (unless Qobuz doesn't have all the music you like). Basically, MQA, at least in all the files I've seen, works like this:

1. The MQA files contain high-res content, consisting of data relating to frequencies from 22-48 kHz, encoded and compressed into the higher frequencies up to 24 kHz (or sometimes lower, depending), resulting in a lossy 24/48 file (or so, as there are other options possible) with other data stating what the supposed "original" resolution was.
2. UAPP can decompress (unfold) this into the original 17/48, 17/88 or 17/96 file*, assuming the original had that resolution or higher. There may be some variation here depending on the original.
3. An MQA capable DAC will up-sample (render) this with an MQA filter (one of a second of very short, minimum-phase filters, from memory) if it sees the MQA bits set on the incoming data. A non-MQA DAC will just do whatever it does normally with the high-res data it receives, so it's best to let UAPP unfold (decompress) the MQA.

*17, not 24, as a number of bits are used to hold the high-res data, and are thus lost with the origami compression.

 

[rlw6534]

Whether a track is MQA or not is indicated by certain bits set in the output. This has been proven by removing the HF content from an MQA file, but keeping the indicator bits, causing an MQA DAC to indicate that the file is MQA. So UAPP could be removing those bits. Going by how Roon works though, those bits can be passed on after the initial unfold is done in software. However, note, most importantly, there are NO MQA tracks with 2nd or 3rd or any other unfolds on TIDAL. None!


The best option is, you're a TIDAL user is: Cancel your subscription and switch to Qobuz (unless Qobuz doesn't have all the music you like). Basically, MQA, at least in all the files I've seen, works like this:

1. The MQA files contain high-res content, consisting of data relating to frequencies from 22-48 kHz, encoded and compressed into the higher frequencies up to 24 kHz (or sometimes lower, depending), resulting in a lossy 24/48 file (or so, as there are other options possible) with other data stating what the supposed "original" resolution was.
2. UAPP can decompress (unfold) this into the original 17/48, 17/88 or 17/96 file*, assuming the original had that resolution or higher. There may be some variation here depending on the original.
3. An MQA capable DAC will up-sample (render) this with an MQA filter (one of a second of very short, minimum-phase filters, from memory) if it sees the MQA bits set on the incoming data. A non-MQA DAC will just do whatever it does normally with the high-res data it receives, so it's best to let UAPP unfold (decompress) the MQA.

*17, not 24, as a number of bits are used to hold the high-res data, and are thus lost with the origami compression.

MQA makes some sense when you understand that the information above 48 kHz PCM (24 kHz audio) is being encoded and stored in the 0-48 kHz band below the noise level of the track (effectively using some the 24 bits of resolution and unutilized dynamic range). While it technically can reduce dynamic range, it doesn't really matter (I guess) if it's below the noise level and the track didn't have the dynamic range to start with.

It's the 16/44.1 and 24/48 MQA tracks that are really confusing as the above doesn't apply... Is that just upsampling and filtering?

 

[Dannemand]

It's the 16/44.1 and 24/48 MQA tracks that are really confusing as the above doesn't apply... Is that just upsampling and filtering?

Yes and no.

1) Most (if not all) 16/44 MQAs are created from old masters that only exist in 16/44. This is the bulk of the millions of new Warner MQA tracks. They really do not contain any origami because there ARE no additional samples to encode. I've been whining about those for some time, and I still wish Tidal would offer the non-MQA variants of them. (Note: Some have speculated that changing Tidal Quality setting from Master to HiFi will deliver a non-MQA version, but that is NOT correct: You get exactly the same 16/44 MQA track. I compared them, they match bit-for-bit, and the HiFi versions DO authenticate as MQA by my DAC.)

Bob Stuart claims that by re-encoding these CD releases using the MQA encoder -- and knowing which ADC was used to encode them originally -- flaws are being corrected, and often the re-encoded versions are significantly clearer, even on non-MQA hardware. See links below. Don't kill the messenger

https://bobtalks.co.uk/blog/science-mqa/16b-mqa-what-is-it/#
https://audioxpress.com/news/tidal-expands-mqa-catalog-available-for-tidal-masters-streaming

2) Masters recorded in 24-bit at 44.1, 88.2, 176.4 or 352.8 KHz are of course encoded in 24-bit and delivered as 24/44 MQA. The first link above explains how they can actually be truncated to 16-bit and still authenticate -- and claimed to still preserve some HiRes detail. This explains a mystery (to me) that playing such tracks from BubbleUPnP (which can only access the 16/44 tier) actually authenticate by my MQA DAC and display as fully unfolding (even up to 352.8KHz in the case of 2L tracks).

MQA haters might argue that this proves there is nothing in the additional 8 bits of MQA tracks -- which can easily be true, as many masters have limited dynamic range and precision. But comparing the actual files, I see plenty of content in those additional bits. IOW, there is significant bit-difference between 24-bit samples and their 16-bit truncated counterparts -- not just zeroes. And both do indeed authenticate and unfold on my MQA DAC.

3) Of course some Master tracks are just 24/44 with no origami to unfold (such Billie Eilish No Time To Die, and Simon Rattle's BPO Brahms cycle). They are usually 2-3x the bitrate of 16/44 tracks -- about the same as tracks that unfold 2X or 4X. This makes me wonder if they actually DO contain origami after all, and it's just meta-information causing my DAC to misinterpret them.

4) Finally there are a number of Tidal Master tracks with 48KHz base sample rate, of which some unfold to 96 or 192 KHz. Most new encodings of quality analog recordings seem to be released this way (such as the Everest Records Master Transfers, like Copland's Symphony No 3 directed by himself, and Tchaikovsky's violin concerto with Spivakovsky/LSO). Others are released as 24/48 and unfold to 48 KHz (such as the Barenboim Staatskapelle Berlin Bruckner cycle).

Anyways, those 16/44, 24/44 and 24/48 tracks that contain no origami will STILL be upsampled 2X by UAPP's MQA decoder according to the Dev. I would prefer if they weren't and just let the DAC perform its upsampling (unless you're a NOS fan). But I don't know if this is Davy's decision or one made by MQA.

Edit: Fixed link to Rattle Brahms cycle.

 

[blotmouse]

Must be a UAPP thing. Audirvana doesn't do this with MQA. 44.1 is unfolded to 44.1 then hits the HW DAC for full encode at... 44.1.And answers what I thought was happening with my M8. UAPP is upsampling the first unfold and sending to the DAC, but doesn't support the M8 HW and the file is stuck at the first unfold.

 

[Dannemand]

Whether a track is MQA or not is indicated by certain bits set in the output. This has been proven by removing the HF content from an MQA file, but keeping the indicator bits, causing an MQA DAC to indicate that the file is MQA. So UAPP could be removing those bits. Going by how Roon works though, those bits can be passed on after the initial unfold is done in software. However, note, most importantly, there are NO MQA tracks with 2nd or 3rd or any other unfolds on TIDAL. None!

Thank you. Interesting. This may be an intentional result of the design described in my post above, by which samples can be truncated and still preserve partial origami and authentication. Still puzzling that MQA DACs display the full original sample rate in those cases. I was certainly scratching my head when I saw that.

That WOULD be explained if it's all in the metadata (the indicator bits, as you're saying) and the whole origami story is just a Big Lie (not sure you're also saying that). I am not ruling that out, but it would be outright fraud, and a fantastic legal risk taken by Meridian, MQA and Stuart. Anybody who is involved in MQA implementations (studios, labels, DAC/XMOS manufacturers, app developers like Davy etc) would know, and must be in on the fraud. If it really is so. That seems improbable to me, but I guess not unheard of.

 

[rlw6534]

Yes and no.

1) Most (if not all) 16/44 MQAs are created from old masters that only exist in 16/44. This is the bulk of the "millions of new Warner MQA tracks". They really do not contain any origami because there ARE no additional samples to encode. I've been whining about those for some time, and I still wish Tidal would offer the non-MQA variants of them. (Note: Some have speculated that changing Tidal Quality setting from Master to HiFi will deliver a non-MQA version, but that is NOT correct: You get exactly the same 16/44 MQA track. I compared them, they match bit-for-bit.)

Bob Stuart claims that by re-encoding these CD releases using the MQA encoder -- and knowing which ADC was used to encode them originally -- flaws are being corrected, and often the re-encoded versions are significantly clearer, even on non-MQA hardware. See links below. Don't kill the messenger

https://bobtalks.co.uk/blog/science-mqa/16b-mqa-what-is-it/#
https://audioxpress.com/news/tidal-expands-mqa-catalog-available-for-tidal-masters-streaming

2) Masters recorded in 24-bit at 44.1, 176.4 or 352.8 KHz are of course encoded in 24-bit and delivered as 24/44 MQA. The first link above explains how they can actually be truncated to 16-bit and still authenticate -- and claimed to still preserve some HiRes detail. This explains a mystery (to me) that playing such tracks from BubbleUPnP (which can only access the 16/44 tier) actually authenticate by my MQA DAC and display as fully unfolding (even up to 352.8KHz in the case of 2L tracks).

MQA haters might argue that this proves there is nothing in the additional 8 bits of MQA tracks -- which can easily be true, as many masters have limited dynamic range and precision. But comparing the actual files, I see plenty of content in those additional bits. IOW, there is significant bit-difference between 24-bit samples and their 16-bit truncated counterparts -- not just zeroes. And both do indeed authenticate and unfold on my MQA DAC.

3) Of course some Master tracks are just 24/44 with no origami to unfold (such Billie Eilish No Time To Die, and Simon Rattle's BPO Brahms cycle). They are usually 2-3x the bitrate of 16/44 tracks -- about the same as tracks that unfold 2X or 4X. This makes me wonder if they actually DO contain origami after all, and it's just meta-information causing my DAC to misinterpret them.

4) Finally there are a number of Tidal Master tracks with 48KHz base sample rate, of which some unfold to 96 or 192 KHz. Most new encodings of quality analog recordings seem to be released this way (such as the Everest Records Master Transfer, like Copland's Symphony No 3 directed by himself, and Tchaikovsky's violin concerto with Spivakovsky's / LSO). Others are released as 24/48 and unfold to 48 KHz (such as the Barenboim Staatskapelle Berlin Bruckner cycle).

Anyways, those 16/44, 24/44 and 24/48 tracks that contain no origami will STILL be upsampled 2X by UAPP's MQA decoder according to the Dev. I would prefer if they weren't and just let the DAC perform its upsampling (unless you're a NOS fan). But I don't know if this is Davy's decision or one made by MQA.

Wow. Thanks for the detailed response! FYI, my Hiby R6 2020 with 16x MQA actually plays the 16/44.1 tracks at 705.6 kHz which makes no sense.

 

[Dannemand]

Wow. Thanks for the detailed response! FYI, my Hiby R6 2020 with 16x MQA actually plays the 16/44.1 tracks at 705.6 kHz which makes no sense.

You're welcome. And no, that makes no sense. Strange that different DACs interpret these differently. Of course their marketing departments have a hand in this ("just make our numbers bigger"). But each of these products must be certified by MQA -- as Gustard learned the hard way when they had to freeze their new X16 DAC for months after release, waiting for certification.

Just to make sure we're talking about the same thing: What does your DAC display for Jaqueline du Pre's Dvorak? That one is definitely a 16/44 recording and MQA release with no additional samples.

 

[Dannemand]

Must be a UAPP thing. Audirvana doesn't do this with MQA. 44.1 is unfolded to 44.1 then hits the HW DAC for full encode at... 44.1.And answers what I thought was happening with my M8. UAPP is upsampling the first unfold and sending to the DAC, but doesn't support the M8 HW and the file is stuck at the first unfold.

Thank you. Yes, I think it's a UAPP thing. I have bugged Davy enough recently, but for anybody who is up for it, this is easy to document: If it plays as 44 or 48 KHz on a hardware MQA DAC, that's what it should play as from UAPP on a non-MQA DAC.