[FiiO UTWS5 2025/UTWS5] True Wireless Bluetooth Amplifier Is Officially Released!

Feb 13, 2023 at 7:16 PM Post #1,906 of 3,144
Just tested it and the M17 + UTWS5 can handle 96/24 at 900bps. 192/24 won’t play ball and gets downsampled.
You can watch the screen recording HERE
Again, it will go to balanced as soon as the setting page gets closed.
 
Last edited:
Feb 13, 2023 at 7:29 PM Post #1,907 of 3,144
Hi everyone.
Finally cant take it anymore. bought this thing on stellar reviews, and its quite a compelling product with great potential - great SoC + great amp + flashable - in a nice TWS-ish form factor, in a well refined case.

FiiO wasn't wrong, technically. They put a piece of solid hardware indeed capable of doing 24/96, with decent battery life, in a small form factor. Hardware wise - this thing is the bees knees. Good stuff.

However, FiiO has managed to build something worth less than the sum of its parts. What was promised to be 24/96 - was quickly discovered to be limited to lossy codecs with annoying clipoffs at 15khz frequency range, no matter the codec used.

And the promised LHDC was only just released, and but quite buggy and lack of host support, tbh. Might as well just give us a good open source codec - more utility.

Since FiiO has managed to deliver on strong hardware, yet lacking firmware - i think its time to make our own.

Able to serial into the device and issue commands to change codec, tweak bitpool settings, negotiate higher bitrate on an old codec which delivers aptx HD-like quality. (go google SBC XQ). My personal best was 640kbps SBC - dual channel, Loudness, 12 blocks, depth 8 @ bitpool of 55. Have given up on aptx and LHDC. aptx adaptive havent tried because my samsung exynos phone didnt provide it. But thats another story.

Have also bricked my left earpiece in this trial and error process, but thank goodness for dfu-mode.

Anyway. FiiO - if you are able to post publicly on a git - or send me a decompiled version of the latest *.bin file, that would be great.

--------------
Also, i noticed that majority of user pain points are centric around 2 things (besides giving crappy codecs and calling it a Hi-Fi unit) -

Priority fix for these bugs please -
1. Sticky DVC disable mode - since the DVC settings i set on android (in developer settings) are sticking to the next device i connect to, which results in wavering volume upon hitting some gain ceiling and dithering. Factory reset makes the symptom go away. but doesn't solve the problem cos it'll come back.

2. Not sure why, but this thing somehow auto downsamples/upsamples the stuff it plays sometimes, contributing to a peculiar muddy, slightly modulated metallic tone in the music. This is what people have experienced after upgrading to 1.5, and went back to 1.42, but actually not required. Just flash twice and factory reset. After that make sure it stays on one frequency - 44.1 or 48. - and stick to it all the time. (This is quite impractical for me personally, but i've made everything 44.1 because of this bug. And because Samsung androids only does 44.1 on aptx and aac. (no apt adaptive).

Edit: 2a: Also there seems to be some form of cross-device volume stickiness accumulated over time. e.g. I usually listen @ volume 16, but have noticed over time I only require volume 10 to get same volume, but lots of clipping when music becomes loud. Seems to be +gain distortion that i'm experiencing. Factory reset solves the symptom, but not root cause. Could be related to point 1 - DVC. Am unsure, but problem does exist.

Edit: 2b: Either disable the aggressive device pairing and stick to traditional single-point connectivity to host (not much use anyway for me), or in Fiio Control allow user to sort pairing order in priority + ability to forget previous devices paired. Redundant feature + inconvenience -> should remove the feature. Or fix.
---------
Improvement points other than bugs:

3. The 'pseudo loseless' codecs that are actually useful, like aptx-hd, aptx ll, l3, aptx-loseless, they dont include. But they throw in a LHDC thingy that nothing can support except in 0.001% of devices. LDAC would be really nice - but i get it - $$$ for codec licensing on a 1 year old product at mid-life cycle. ok fine. maybe not cost effective on your side. But the others though! Highly implementable! - therefore:

Would be great if FiiO could include codecs mentioned above - the chipset does support it - i dont think it's a licensing issue.
Or if im wrong, at least give the user control on the SBC and AAC codec bitpool, switching between dual channel, stereo, joint stereo, - etc. , since that's open source. This alone would already de-limit a high potential market offering, without much need for more capital outlay and development.

3a. FiiO - might be worth to have your engineers look into changing the SBC maximum bitpool from 33 to maybe 50. And give user ability to force dual channel mode The hardware is capable of it, and it will negotiate downward if its not. And if it doesnt, you can write a simple automated script, factoring in RSSI strength and reducing/increasing bitpool on the fly - this could perhaps be implemented in the Fiio Control App since serial port is always connected to the host anyway This might help automatically increase negotiated bitrates.

3b. IMPORTANT + EASY TO IMPLEMENT: Would be good to put - EQ presets for the big 3 codecs - to compensate somewhat for the high frequency roll-off that the popular codecs suffer from near the 12-18khz band. I think this would be much more helpful than putting some "Rock" or "Pop" EQ preset. Or maybe thats what you guys implemented the high-pass filter thingy for - which is s good segue to next point below -

3c. Please make the high pass filter graph decipherable. Thats some snake oil crap u got going on there. Just doesnt make sense. Some graphs, the axes are differently labled. idkwtf is going on.

4. Some recognition is in order - This product, upon inspection, seems to have some rudimentary form of Bluetooth Low Energy implementation - Edit - removed. Qcc51xx has LE, but seems unable to use channel for streaming bandwidth increase. damn.
--------

If FiiO could release the firmware source, I believe modders will make the product good again in a way that FiiO can benefit by - a) selling more UTWS5s, and b) building their engineering team's software competencies to rival giants like Shure and Sony.

win-win scenario - and life cycle of a USD129.00 is increased as well - at least until the hardware breaks due to wear-tear.

Really great product. But really poor implementation. Send help :frowning2:

Finally - to the FiiO executive who reads this - its not hard to get a good Qualcomm SoC, case it up, value add on the software, brand it as yours - the sell it to wiling customers used to paying good money for good sound. But first and foremost, you are a systems integrator, which means this is heavily software dependent, and your true value proposition differentiating you from Shanling, iBasso, and others. Meaning - the software dev is the last thing u wanna be lazy on, cos thats the heart and soul of the device, and also the key touch-point the consumer either loves, or hates. But half the battle is already won - u have integrated good hardware.
If you fix up the software and the UI/UX stuff like how Xiaomi managed to do with their market offerings, you'll be able to compete at mass market level with the big established Japanese and American/Euro brands. Else, you'll be at the CHi-Fi plateau that you're at currently - good hardware, but tons of pain points and poor user experience - resulting in suboptimal revenue streams and untapped potential of your product line and business unit. Perhaps then i'll send u my resume.

tq and regards.
 
Last edited:
Feb 13, 2023 at 7:40 PM Post #1,908 of 3,144
@fmaorui , God bless you! Hope they will listen to you!! I hope someone will read you, not just a rep who can only say: friend, I will convey your message to the engineer (lmao!)
 
Feb 13, 2023 at 7:58 PM Post #1,909 of 3,144
Appendix A -

Seems like you FiiO guys kinda made something work with the FiiO Music app - the HWA toggle (though not working on my phone. but then again its Samsung, so not your fault)
If you could, in the Fiio Control app, implement a toggle to hijack audio streams and output through the Fiio Control app in Hi-Fi, this could be useful for the eventual A2DP stack you may be thinking of making (or not.)

But hey - Poweramp did it, in a very beautiful UI.

Also highly useful in troubleshooting What went wrong with your sound as a audiophile end-user's first foray into wireless tech. (which btw is shady af cos bluetooth is fundamentally flawed vs low powered wifi-direct @ 5mbps bandwidth - dont even need to compress into codec lol. but again. sorry for digressing -

Look at that. Beautiful yes? One glance - trace the whole chain from source to DSP to resampler to whatever to device. Then u realize where the bottleneck is, and how to to compensate.

840059146_PowerAmpAudioDetails1of2.thumb.png.51eb10d9aecc1b2bfb6dfe9b1fa0965b.png
2037038992_PowerAmpAudioDetails2of2.png.b006586c6888d22f53e4e36f6cb3933e.png

Or for modders that doesnt mind bricking their FiiO to high hell

https://lineageos.org/engineering/Bluetooth-SBC-XQ/

https://appuals.com/how-to-modify-b...for-greatly-enhanced-bluetooth-audio-quality/

https://sudonull.com/post/6801-Bluetooth-based-wireless-audio-whats-better

https://habr.com/en/post/456182/

https://4pda.to/forum/index.php?showtopic=914135


Or you know what? screw it. maybe you just make your BTWS15 able to receive chromecast and airplay audio streams. at 2mbps minimum @ 100ms latency, confirm will beat any competitor currently. not sure battery life though lol, but hey i'm no engineer, just a disgruntled user trying to help solve this stupid thing.
 
Last edited:
Feb 13, 2023 at 8:00 PM Post #1,910 of 3,144
Well, I broke down and ordered a set of MMCX-3.5mm adapters -- gonna try these driving my HEKse. They're relatively efficient, and the most efficient full-size cans I own, but do need a fair amount of current, so I'm not sure the UTWS5s can drive them adequately. Oh well, a few bucks to test won't kill me.
Arrived — and they work w/ my HEKse, although nearly all the way up to get to ‘loud enough’.

Annoying rattle when you move your head and the adapter bangs against the headphone grill.
 
Feb 13, 2023 at 8:06 PM Post #1,911 of 3,144
Hi everyone.
Finally cant take it anymore. bought this thing on stellar reviews, and its quite a compelling product with great potential - great SoC + great amp + flashable - in a nice TWS-ish form factor, in a well refined case.

FiiO wasn't wrong, technically. They put a piece of solid hardware indeed capable of doing 24/96, with decent battery life, in a small form factor. Hardware wise - this thing is the bees knees. Good stuff.

However, FiiO has managed to build something worth less than the sum of its parts. What was promised to be 24/96 - was quickly discovered to be limited to lossy codecs with annoying clipoffs at 15khz frequency range, no matter the codec used.

And the promised LHDC was only just released, and but quite buggy and lack of host support, tbh. Might as well just give us a good open source codec - more utility.

Since FiiO has managed to deliver on strong hardware, yet lacking firmware - i think its time to make our own.

Able to serial into the device and issue commands to change codec, tweak bitpool settings, negotiate higher bitrate on an old codec which delivers aptx HD-like quality. (go google SBC XQ). My personal best was 640kbps SBC - dual channel, Loudness, 12 blocks, depth 8 @ bitpool of 55. Have given up on aptx and LHDC. aptx adaptive havent tried because my samsung exynos phone didnt provide it. But thats another story.

Have also bricked my left earpiece in this trial and error process, but thank goodness for dfu-mode.

Anyway. FiiO - if you are able to post publicly on a git - or send me a decompiled version of the latest *.bin file, that would be great.

--------------
Also, i noticed that majority of user pain points are centric around 2 things (besides giving crappy codecs and calling it a Hi-Fi unit) -

Priority fix for these bugs please -
1. Sticky DVC disable mode - since the DVC settings i set on android (in developer settings) are sticking to the next device i connect to, which results in wavering volume upon hitting some gain ceiling and dithering. Factory reset makes the symptom go away. but doesn't solve the problem cos it'll come back.

2. Not sure why, but this thing somehow auto downsamples/upsamples the stuff it plays sometimes, contributing to a peculiar muddy, slightly modulated metallic tone in the music. This is what people have experienced after upgrading to 1.5, and went back to 1.42, but actually not required. Just flash twice and factory reset. After that make sure it stays on one frequency - 44.1 or 48. - and stick to it all the time. (This is quite impractical for me personally, but i've made everything 44.1 because of this bug. And because Samsung androids only does 44.1 on aptx and aac. (no apt adaptive).
---------
Improvement points other than bugs:

3. The loseless codecs that are actually useful, like aptx-hd, l3, aptx-loseless, they dont include. But they throw in a LHDC thingy that nothing can support except in 0.001% of devices.
Would be great if FiiO could include these codecs - the chipset does support it - i dont think it's a licensing issue.
Or if im wrong, at least give the user control on the SBC and AAC codec bitpool, switching between dual channel, stereo, joint stereo, - etc. , since that's open source. This alone would already de-limit a high potential market offering, without much need for more capital outlay and development.

3a. FiiO - might be worth to have your engineers look into changing the SBC maximum bitpool from 33 to maybe 50. The hardware is capable of it, and it will negotiate downward if its not. And if it doesnt, you can write a simple automated script, factoring in RSSI strength and reducing/increasing bitpool on the fly - this could perhaps be implemented in the Fiio Control App since serial port is always connected to the host anyway This might help automatically increase negotiated bitrates.

3b. IMPORTANT + EASY TO IMPLEMENT: Would be good to put - EQ presets for the big 3 codecs - to compensate somewhat for the high frequency roll-off that the popular codecs suffer from near the 12-18khz band. I think this would be much more helpful than putting some "Rock" or "Pop" EQ preset. Or maybe thats what you guys implemented the high-pass filter thingy for - which is s good segue to next point below -

3c. Please make the high pass filter graph decipherable. Thats some snake oil crap u got going on there. Just doesnt make sense. Some graphs, the axes are differently labled. idkwtf is going on.

4. Some recognition is in order - This product, upon inspection, seems to have some rudimentary form of Bluetooth Low Energy implementation (possibly for L3 codec stuff) (infrastructure not coded perhaps). I was able to install BT LE isochronous streams, hoping it would increase bandwidth as I tried to break the 700kbps barrier. but was lacking BT LE Audio stack, so unable to progress. Seems good that the infrastructure is there though. Theoratically this thing, with EDR, is capable of sustaining a theoratical 1-2Mbps link. Sad to see it capped at 600-700kbps, like putting a rpm limit on a Ferrari. But hardware-wise capable - means software unlockable, which is awesome.
--------

If FiiO could release the firmware source, I believe modders will make the product good again in a way that FiiO can benefit by - a) selling more UTWS5s, and b) building their engineering team's software competencies to rival giants like Shure and Sony.

win-win scenario - and life cycle of a USD129.00 is increased as well - at least until the hardware breaks due to wear-tear.

Really great product. But really poor implementation. Send help :frowning2:
@FiiO please read all of the above text.
 
Feb 13, 2023 at 8:16 PM Post #1,912 of 3,144
Hi everyone.
Finally cant take it anymore. bought this thing on stellar reviews, and its quite a compelling product with great potential - great SoC + great amp + flashable - in a nice TWS-ish form factor, in a well refined case.

FiiO wasn't wrong, technically. They put a piece of solid hardware indeed capable of doing 24/96, with decent battery life, in a small form factor. Hardware wise - this thing is the bees knees. Good stuff.

However, FiiO has managed to build something worth less than the sum of its parts. What was promised to be 24/96 - was quickly discovered to be limited to lossy codecs with annoying clipoffs at 15khz frequency range, no matter the codec used.

And the promised LHDC was only just released, and but quite buggy and lack of host support, tbh. Might as well just give us a good open source codec - more utility.

Since FiiO has managed to deliver on strong hardware, yet lacking firmware - i think its time to make our own.

Able to serial into the device and issue commands to change codec, tweak bitpool settings, negotiate higher bitrate on an old codec which delivers aptx HD-like quality. (go google SBC XQ). My personal best was 640kbps SBC - dual channel, Loudness, 12 blocks, depth 8 @ bitpool of 55. Have given up on aptx and LHDC. aptx adaptive havent tried because my samsung exynos phone didnt provide it. But thats another story.

Have also bricked my left earpiece in this trial and error process, but thank goodness for dfu-mode.

Anyway. FiiO - if you are able to post publicly on a git - or send me a decompiled version of the latest *.bin file, that would be great.

--------------
Also, i noticed that majority of user pain points are centric around 2 things (besides giving crappy codecs and calling it a Hi-Fi unit) -

Priority fix for these bugs please -
1. Sticky DVC disable mode - since the DVC settings i set on android (in developer settings) are sticking to the next device i connect to, which results in wavering volume upon hitting some gain ceiling and dithering. Factory reset makes the symptom go away. but doesn't solve the problem cos it'll come back.

2. Not sure why, but this thing somehow auto downsamples/upsamples the stuff it plays sometimes, contributing to a peculiar muddy, slightly modulated metallic tone in the music. This is what people have experienced after upgrading to 1.5, and went back to 1.42, but actually not required. Just flash twice and factory reset. After that make sure it stays on one frequency - 44.1 or 48. - and stick to it all the time. (This is quite impractical for me personally, but i've made everything 44.1 because of this bug. And because Samsung androids only does 44.1 on aptx and aac. (no apt adaptive).
---------
Improvement points other than bugs:

3. The loseless codecs that are actually useful, like aptx-hd, l3, aptx-loseless, they dont include. But they throw in a LHDC thingy that nothing can support except in 0.001% of devices.
Would be great if FiiO could include these codecs - the chipset does support it - i dont think it's a licensing issue.
Or if im wrong, at least give the user control on the SBC and AAC codec bitpool, switching between dual channel, stereo, joint stereo, - etc. , since that's open source. This alone would already de-limit a high potential market offering, without much need for more capital outlay and development.

3a. FiiO - might be worth to have your engineers look into changing the SBC maximum bitpool from 33 to maybe 50. The hardware is capable of it, and it will negotiate downward if its not. And if it doesnt, you can write a simple automated script, factoring in RSSI strength and reducing/increasing bitpool on the fly - this could perhaps be implemented in the Fiio Control App since serial port is always connected to the host anyway This might help automatically increase negotiated bitrates.

3b. IMPORTANT + EASY TO IMPLEMENT: Would be good to put - EQ presets for the big 3 codecs - to compensate somewhat for the high frequency roll-off that the popular codecs suffer from near the 12-18khz band. I think this would be much more helpful than putting some "Rock" or "Pop" EQ preset. Or maybe thats what you guys implemented the high-pass filter thingy for - which is s good segue to next point below -

3c. Please make the high pass filter graph decipherable. Thats some snake oil crap u got going on there. Just doesnt make sense. Some graphs, the axes are differently labled. idkwtf is going on.

4. Some recognition is in order - This product, upon inspection, seems to have some rudimentary form of Bluetooth Low Energy implementation (possibly for L3 codec stuff) (infrastructure not coded perhaps). I was able to install BT LE isochronous streams, hoping it would increase bandwidth as I tried to break the 700kbps barrier. but was lacking BT LE Audio stack, so unable to progress. Seems good that the infrastructure is there though. Theoratically this thing, with EDR, is capable of sustaining a theoratical 1-2Mbps link. Sad to see it capped at 600-700kbps, like putting a rpm limit on a Ferrari. But hardware-wise capable - means software unlockable, which is awesome.
--------

If FiiO could release the firmware source, I believe modders will make the product good again in a way that FiiO can benefit by - a) selling more UTWS5s, and b) building their engineering team's software competencies to rival giants like Shure and Sony.

win-win scenario - and life cycle of a USD129.00 is increased as well - at least until the hardware breaks due to wear-tear.

Really great product. But really poor implementation. Send help :frowning2:
@FiiO ,@FiiO Willson , please read the above message carefully and reply!
 
Feb 13, 2023 at 9:17 PM Post #1,913 of 3,144
@fmaorui , God bless you! Hope they will listen to you!! I hope someone will read you, not just a rep who can only say: friend, I will convey your message to the engineer (lmao!)
tq for the kind words.
ultimately the underlying theme in pushing for more end-user control of their hardware - is a theme that reverberates through history no matter the subject at hand:

power to the people.
 
Feb 14, 2023 at 4:57 AM Post #1,915 of 3,144
this thing somehow auto downsamples/upsamples the stuff it plays sometimes, contributing to a peculiar muddy, slightly modulated metallic tone in the music. This is what people have experienced after upgrading to 1.5, and went back to 1.42,
Very unlikely this has anything to do with resampling. I described the tonal shifts produced by trying to push LHDC too hard back in November, and it has nothing to do with the new 1.5 firmware. Trying to squeeze too high a bitrate into an antenna sitting next to a big bag of radio-absorbent jelly is always going to cause problems - high bitrate BT is not for wearable tech.

As for the rest ... well it's not hard to see why Fiio wouldn't want to deal with the support nightmare of users bricking their devices after following some garbled instructions from an amateur hacker. And I can't see how these mods would offer any actual utility. More != Better
 
Feb 14, 2023 at 5:36 AM Post #1,916 of 3,144
Poco F2 Pro, supports LHDC V3/V4, among a ton of other codecs. In the Developer settings, I set Bluetooth audio sample rate to 44.1kHz and the codec to 500/560 kbps, the only one supported.

Two pairs of the UTWS5, both running 1.5 firmware. In one of them I disabled all other codecs besides LHDC.

I live in a 3 floor house. I leave my phone on the first floor, on my desk. I go upstairs and downstairs. Almost no loss of connection at all. Only in the upstairs bathroom, near the roof window, it loses connection but it regains it if I will move towards the middle of the bathroom.

When I go outside of my property, I am always wearing one of my 2 pairs of the UTWS5. I put my phone in the inside pocket of my jacket or in one of the outside pockets. I have never-ever faced any issues with disconnections, sound distortion or anything else. Never! Even the transition from Wi-Fi to Mobile data on my phone is completely smooth and I do not hear anything.

As for the rest.. FiiO should be much more responsive and responsible. They appear very rarely in this thread, they are giving some generic-robotic replies, they are not taking in consideration users' requests and it takes them centuries to update something. They could consider this request and give some control to the users, with a full disclaimer that they will not be responsible for any issues that might occur.
 
Last edited:
Feb 14, 2023 at 6:11 AM Post #1,917 of 3,144
Very unlikely this has anything to do with resampling. I described the tonal shifts produced by trying to push LHDC too hard back in November, and it has nothing to do with the new 1.5 firmware. Trying to squeeze too high a bitrate into an antenna sitting next to a big bag of radio-absorbent jelly is always going to cause problems - high bitrate BT is not for wearable tech.

As for the rest ... well it's not hard to see why Fiio wouldn't want to deal with the support nightmare of users bricking their devices after following some garbled instructions from an amateur hacker. And I can't see how these mods would offer any actual utility. More != Better
Hi -
1. would suggest you read the Qcc5141 chipset technical sheet - should answer your issue on high bitrate BT. Also seems to be in line with regard to chip size vs transistor density (and corresponding bitrate increase). Would postulate that high bitrate whatever tech, wearable or not - is basically a product of Moore's Law. Which is how technology progresses in general.

To put this in perspective, people were getting a 32kbps rate using BT2.0 about 15 years ago using pure BT Sco output.(think old school mono Jabra earpices) With LDAC high quality output now, thats about a 2500% increase in bitrate output, while size of product shrank by about 50% or more.

https://www.qualcomm.com/content/da...qcc5100-series-product-brief_87-cf482-1-i.pdf

You will see that Qcc5141 - the chip used in UTWS5 - does not support LE Audio. It does, however, support LE *bandwidth* implementations, allowing for 1 - 2mbps stable rate (albeit at expense of battery) - which is why, when you pair it with your Windows PC, and have a look at device manager, you will see 4 unrecognizable bluetooth peripherals without drivers.

Thats your unused BLE stack channels sitting idle. Which is why its only plateauing @ 600-700kbps pure throughput. Its actually capable of much more, but the LE stack has to first be adopted, consolidated, and pushed to mass market. I believe Bose has already done that, because they have a Qcc5141 product which works well. Of course, their app is much more refined, including programmable DSPs, excellent tweakable ANC modes, etc. packaged in a nice shiny easy-to-use UI.

Also the reason why it has multipoint connectivity is because theres tons of idle bandwidth not utilised, as its capable of 32 simultaneous theoretical channels.

Would be good to note at this point, that, as wireless audiophiles, we are essentially early adopters of bleeding edge tech, which is why bugs occur. But thats good because law of diminishing marginal returns hasn't set in yet - any small effort in optimisation will yield high marginal benefit for now.

2. I haven't put this device to test with LHDC yet either - so not sure about your issue.

Am however experiencing the tonal shifts here independent of firmware version, bitrate output, nor LHDC, as i believe others might as well.

Have found upon trial and error that tonal shifts - at least for my side only occur during sample hz mismatch between track and output.
- Could be possible that the tonal shifts you're experiencing is a different issue to mine? Or different manifestations of the same root cause.

3. Finally - am proposing some features be removed, not added, which is why i divided my long post into sub-parts in order of priority.


All being said, definitely an amateur tinkerer here though.

Serial port commands through the GAIA API were obtained through this reference - which is basically a deep dive into developing and flashing a factory default Qcc51xx chipset from scratch. Definitely not for the end-user. And definitely not what i'm proposing. You will brick your device.

Which is why am hoping FiiO steps in.

https://blog.csdn.net/maomao199306/article/details/128735417?spm=1001.2101.3001.6650.5&utm_medium=distribute.pc_relevant.none-task-blog-2~default~BlogCommendFromBaidu~Rate-5-128735417-blog-126021600.pc_relevant_recovery_v2&depth_1-utm_source=distribute.pc_relevant.none-task-blog-2~default~BlogCommendFromBaidu~Rate-5-128735417-blog-126021600.pc_relevant_recovery_v2&utm_relevant_index=10
 
Last edited:
Feb 14, 2023 at 7:46 AM Post #1,918 of 3,144
Hi -
1. would suggest you read the Qcc5141 chipset technical sheet - should answer your issue on high bitrate BT. Also seems to be in line with regard to chip size vs transistor density (and corresponding bitrate increase). Would postulate that high bitrate whatever tech, wearable or not - is basically a product of Moore's Law. Which is how technology progresses in general.

To put this in perspective, people were getting a 32kbps rate using BT2.0 about 15 years ago using pure BT Sco output.(think old school mono Jabra earpices) With LDAC high quality output now, thats about a 2500% increase in bitrate output, while size of product shrank by about 50% or more.

https://www.qualcomm.com/content/da...qcc5100-series-product-brief_87-cf482-1-i.pdf

You will see that Qcc5141 - the chip used in UTWS5 - does not support LE Audio. It does, however, support LE *bandwidth* implementations, allowing for 1 - 2mbps stable rate (albeit at expense of battery) - which is why, when you pair it with your Windows PC, and have a look at device manager, you will see 4 unrecognizable bluetooth peripherals without drivers.

Thats your unused BLE stack channels sitting idle. Which is why its only plateauing @ 600-700kbps pure throughput. Its actually capable of much more, but the LE stack has to first be adopted, consolidated, and pushed to mass market. I believe Bose has already done that, because they have a Qcc5141 product which works well. Of course, their app is much more refined, including programmable DSPs, excellent tweakable ANC modes, etc. packaged in a nice shiny easy-to-use UI.

Also the reason why it has multipoint connectivity is because theres tons of idle bandwidth not utilised, as its capable of 32 simultaneous theoretical channels.

Would be good to note at this point, that, as wireless audiophiles, we are essentially early adopters of bleeding edge tech, which is why bugs occur. But thats good because law of diminishing marginal returns hasn't set in yet - any small effort in optimisation will yield high marginal benefit for now.

2. I haven't put this device to test with LHDC yet either - so not sure about your issue.

Am however experiencing the tonal shifts here independent of firmware version, bitrate output, nor LHDC, as i believe others might as well.

Have found upon trial and error that tonal shifts - at least for my side only occur during sample hz mismatch between track and output.
- Could be possible that the tonal shifts you're experiencing is a different issue to mine? Or different manifestations of the same root cause.

3. Finally - am proposing some features be removed, not added, which is why i divided my long post into sub-parts in order of priority.


All being said, definitely an amateur tinkerer here though.

Serial port commands through the GAIA API were obtained through this reference - which is basically a deep dive into developing and flashing a factory default Qcc51xx chipset from scratch. Definitely not for the end-user. And definitely not what i'm proposing. You will brick your device.

Which is why am hoping FiiO steps in.

https://blog.csdn.net/maomao199306/article/details/128735417?spm=1001.2101.3001.6650.5&utm_medium=distribute.pc_relevant.none-task-blog-2~default~BlogCommendFromBaidu~Rate-5-128735417-blog-126021600.pc_relevant_recovery_v2&depth_1-utm_source=distribute.pc_relevant.none-task-blog-2~default~BlogCommendFromBaidu~Rate-5-128735417-blog-126021600.pc_relevant_recovery_v2&utm_relevant_index=10
It's not surprising that Qualcomm's marketing materials advertise extremely high data rates. I'm sure they can squeeze 3Mbps through the pipe under lab conditions. Real-world radio is a different matter, though, especially with wearable tech where both sending and receiving antennae are right next to a highly absorbent structure, and the actual rate you get will vary from moment to moment in a way that's very hard to predict.

The key factor is graceful degradation rather than trying to squeeze in as much as you can. In my experience neither LHDC nor LDAC degrade in a particularly graceful manner. LHDC in particular displays several different mannerisms when failing: sometimes it will drop the audio or crackle, sometimes it will only drop a few sub-bands (and thus produce a tonal shift as it loses part of the payload's frequency spectrum), sometimes it will drop bitdepth and sound like it's clipping; sometimes this will persist and sometimes it will snap back and work fine for a while. I tried LHDC extensively on 1.42 and there were times when it would work great for extended periods, but most of the time it just wasn't worth it as it would suddenly fail and screw up the audio in odd ways.

All these high-rate codecs were designed for static installations with good connectivity. AptX-Adaptive was designed to fail gracefully and is just a much better solution for wearables with highly variable radio conditions, though I strongly suspect that most of the time it's not actually providing much more bandwidth than ordinary AptX. If you try cramming a quart into a pint pot you'll end up unhappy.
 
Feb 14, 2023 at 8:00 AM Post #1,919 of 3,144
It's not surprising that Qualcomm's marketing materials advertise extremely high data rates. I'm sure they can squeeze 3Mbps through the pipe under lab conditions. Real-world radio is a different matter, though, especially with wearable tech where both sending and receiving antennae are right next to a highly absorbent structure, and the actual rate you get will vary from moment to moment in a way that's very hard to predict.

The key factor is graceful degradation rather than trying to squeeze in as much as you can. In my experience neither LHDC nor LDAC degrade in a particularly graceful manner. LHDC in particular displays several different mannerisms when failing: sometimes it will drop the audio or crackle, sometimes it will only drop a few sub-bands (and thus produce a tonal shift as it loses part of the payload's frequency spectrum), sometimes it will drop bitdepth and sound like it's clipping; sometimes this will persist and sometimes it will snap back and work fine for a while. I tried LHDC extensively on 1.42 and there were times when it would work great for extended periods, but most of the time it just wasn't worth it as it would suddenly fail and screw up the audio in odd ways.

All these high-rate codecs were designed for static installations with good connectivity. AptX-Adaptive was designed to fail gracefully and is just a much better solution for wearables with highly variable radio conditions, though I strongly suspect that most of the time it's not actually providing much more bandwidth than ordinary AptX. If you try cramming a quart into a pint pot you'll end up unhappy.
yo - u tried LHDC vs aptx adaptive. i tried neither.
ask u -
in your opinion - how is perceived sound quality of aptx adaptive quality at peak bitrate vs LHDC?
Asking cos am thinking of getting a usbc aptx adaptive transmitter
Does it still have the high freq cutoff?
These codecs. sigh. not very accessible to me at the moment. because **** samsung.
Hopefully next gen L3 codec lives up to the hype. At least L3 compression algorithm was developed more recently - 2005, vs aptx -1988

Edit: AptX adaptive is highly overrated.
 
Last edited:
Feb 14, 2023 at 8:59 AM Post #1,920 of 3,144
Edit:

Request on LE test - withdrawn.

Dude that replied after this post is factually accurate. Only 5171 has LE Audio for now.
However despite Audio not txed over BLE, LE infrastructure still exists, since both based on same architecture.

My UTWS5 LE stack dump -

BTHLEDevice\{00001801-0000-1000-8000-00805f9b34fb}_40ed981abee8\9&24e2aad7&7&0001 Generic Attribute
BTHLEDevice\{00001800-0000-1000-8000-00805f9b34fb}_40ed981abee8\9&24e2aad7&7&0007 Generic Accesss
BTHLEDevice\{0000180a-0000-1000-8000-00805f9b34fb}_40ed981abee8\9&24e2aad7&7&000c Device Information
BTHLEDevice\{00001100-d102-11e1-9b23-00025b00a5a5}_40ed981abee8\9&24e2aad7&7&001f CSR GAIA - Command/Response Channel - SDK driver req'd
BTHLEDevice\{0000eb10-d102-11e1-9b23-00025b00a5a5}_40ed981abee8\9&24e2aad7&7&0028 GAIA stuff? unknown/Vendor specific - SDK driver req'd
BTHLEDevice\{0000fd92-0000-1000-8000-00805f9b34fb}_40ed981abee8\9&24e2aad7&7&0032 Qualcomm stuff - unknown/Vendor specific --???
BTHLEDevice\{0000180f-0000-1000-8000-00805f9b34fb}_40ed981abee8\9&24e2aad7&7&003b Battery telemetry L
BTHLEDevice\{0000180f-0000-1000-8000-00805f9b34fb}_40ed981abee8\9&24e2aad7&7&0040 Battery telemetry R

could have more, but PHY connection layers wasnt called up yet, so couldn't detect.
If I were to guess i'd say maybe this exists partly for firmware update + interfacing with GAIA Api - i.e. FiiO Control. Initially thought was through BT stack serial port.

From qcc5141 papers, seems like audio tx/rx occurs on BT stack, though apparently PHY links occurs on both BT and BLE level for more bandwidth channels - if phone is paired to compatible host. Therfore technically not LE Audio, unlike power efficient elder brother Qcc5171.

Or BLE link may just be dedicated to firmware/app interface related stuff, though seems like a waste to not utilize as situation requires. 🤷‍♂️

Can get quite technical, this BT stuff. Will experiment again when free.

Meantime, should allow thread to go back to audio discussion.
 
Last edited:

Users who are viewing this thread

Back
Top