• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Topping BC3 Review (Bluetooth Receiver) & BT CODECs

OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,376
Likes
234,529
Location
Seattle Area
So I got my iPhone 6SE fired up and ran Roon there. Here are the results:

Topping BC3 Measurements Bluetooth iPhone AAC.png


Sadly all the issues are there as with Android. I told Roon to output 16 bits instead of 24 and this was that outcome:


Topping BC3 Measurements Bluetooth iPhone 16 bit AAC.png


Things got much worse!

So this test rules out the encoder being faulty. Maybe Qualcomm Bluetooth module's AAC decoder is at fault. It is entirely possible to build a lousy decoder since there are no certifications for AAC, MP3, etc. In this day and age though with so much computing power available, it is hard to imagine this is a problem but what else could it be?
 
Last edited:

Francis Vaughan

Addicted to Fun and Learning
Forum Donor
Joined
Dec 6, 2018
Messages
933
Likes
4,697
Location
Adelaide Australia
That is a pretty interesting result.
What is especially interesting is how the additional energy peak is tucked in close within the critical band of the fundamental. The next step would be to see whether anyone can detect the difference in an ABX test. That would be the real test of the capability of the codec. It may well be that nobody can detect the codec's operation, which would be most interesting.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,376
Likes
234,529
Location
Seattle Area
Went back to basics by just encoding some 1 kHz tones in AAC and examining them on the computer. Tested both with 24-bit test signal that I used here and 16 bits:

AAC 1 kHz tone comparison.png


We clearly see the same signature. There is widening of the main tone ("skirt") and the follow on spikes. 24-bit version is worse due to use of dither to convert it to 16 bits, making it harder on the encoder. I am unclear about the odd harmonics.

I held AAC to higher regard!
 

MayaTlab

Addicted to Fun and Learning
Joined
Aug 15, 2020
Messages
924
Likes
1,512
So this test rules out the encoder being faulty. Maybe Qualcomm Bluetooth module's AAC decoder is at fault. It is entirely possible to build a lousy decoder since there are no certifications for AAC, MP3, etc.

That's something that I'd love to see increasingly investigated. Not just the encoder side, but also the decoder.
I've noticed quite a bit of variation in terms of production of audible artefacts with SBC receiving devices, even with the same emitting device. For example I sent Qudelix a mail because I could reliably produce an issue on my copy with SBC at lower frequencies. The same phenomenon didn't occur with other SBC receivers even with the same emitter.
I'm wondering if AAC suffers from the same problem.
 

Lunafag

Member
Joined
Apr 14, 2021
Messages
89
Likes
127
Interesting AAC results. It could still potentially be an encoder issue, there's like ten of them, just surprising iphone's wasn't one of the "good" ones.
 

Mulder

Addicted to Fun and Learning
Forum Donor
Joined
Sep 2, 2020
Messages
624
Likes
871
Location
Gothenburg, Sweden
I am wondering what conclusions ought to be drawn efter this test. I can't imagine in what user-case anyone would send 24 bits data from ROON to a bluetooth unit. I mean - why? Why not use WiFi? For me bluetooth is a technology för wireless transmission in an environment when you are on the run, for examplewireless headphones, or when you are not at home. In public transportation or at a café. In an home-environment bluetooth is never the optimal Technology from a sonic point of view. To me this raises other kind of questions. Like - are these codecs the same considering battery drain? And can we hear the differenses between different codecs?
 

Francis Vaughan

Addicted to Fun and Learning
Forum Donor
Joined
Dec 6, 2018
Messages
933
Likes
4,697
Location
Adelaide Australia
And can we hear the differenses between different codecs?
And can we hear the difference from lossless? I have a couple of Topping DACs around the place, a DX3 Pro and a D-50s. Both are often fed over bluetooth simply because it is convenient. I might stream from a phone, feeding a pair of powered monitors in the workshop and so on. Even the bedroom system. WiFi means more messing and usually an additional box.
Whilst the traditional audiophile OCD always wants lossless, it remains to be shown that in all bar the most critical of listening environemnts that there is any downside to these lossy codecs. I am very suspicious the answer is a flat "no". Indeed even in critical environments I bet the answer is usually "no".
 

Madlop26

Active Member
Joined
May 2, 2021
Messages
189
Likes
331
First, Thanks to Amir for taking the time to do objective research, as I said you are the man!. Francis, good point there, can we normal folks even tell the difference? well, I decided to do my own AB test (wanted to do a blind test but my wife would not put up to my nonsenses haha) So i used my iPhone (Bluetooth) and MacBook Pro (digital), connected to my S.M.S.L SU-9 DAC, Topping A90 amp and my Dan Clark Audio AEON RT headphones. I used the Music app and later Spotify. i run into some issues , it takes about 5 sec or so to change Bluetooth to digital and visceversa, if the change were instantaneous would be much much better, also the volume of the iPhone slight lower so I had to do continuous corrections and the perception in difference suffered.
Conclusion: Bluetooth sounded inferior.....but just barely, is this a nonblind bias? i don't know; would it stand in a blind test? I don't know; if there is a difference the difference is very small to my ears.
I am pretty sure there are people out there can hear a much more difference than me. after all ACC is a compressed signal. But in my case, even if it turns out I can not tell the difference, just the knowledge of using a signal closer to the natural sound gives me an additional incentive to appreciate the music, yeah I know........ but don't underestimate the things that can put you in a good mood to listen to music.
 

Robbo99999

Master Contributor
Forum Donor
Joined
Jan 23, 2020
Messages
6,876
Likes
6,673
Location
UK
There is an ironic chicken and egg question about measuring the capability of these codecs. All of the lossy compression schemes are perceptual. They are the result of application of enormous amounts of research into the human auditory system. Understanding of the intrinsic limitations due to things like masking, critical bands and so on. Any argument about what matters in actual hearing versus simple measurements would almost certainly find that the largest body of knowledge on the subject is within the design of these codecs. Design is known to have included rigorous listening tests.
Some criticism of the early codecs was that they placed undue emphasis on classical music. (Not that I care.) But modern ones are likely wider in scope.
The reality is that once past a reasonable bit rate they are perceptually transparent. You can tell their presence, not by a reduction in sonic quality, but by listening for specific tells. Like unnatural cut-off of fade outs.

In a curious way one could regard a good perceptual codec, not as a device to try and trip up with weird and contrived waveforms, but as an Oracle. It will pass that part of a signal that is perceptually important, and stuff it doesn’t pass is the part of the signal that isn’t worth worrying about. In this manner perceptual codecs might be a useful tool in trying to bridge the gap in deciding the relative importance of distortion components.
So maybe running artificially generated distortion through codecs (perhaps with a foreground of music) to see which parts get carried through the codec & which parts get cut off.....the parts that get cut off being non-significant from a perceptual level.....thereby you could somehow deduce what "patterns" of distortion are audible based on if the codec removes or includes it. No idea really, just playing with some ideas.
 

Thomas_A

Major Contributor
Forum Donor
Joined
Jun 20, 2019
Messages
3,423
Likes
2,407
Location
Sweden
I have most music in AAC 320 kbit/s. Iif someone can give example of music where AAC 320 kbit/s is audibly different from loss-less I am all ears.
 
Last edited:

Bullwinkle J Moose

Active Member
Joined
Mar 31, 2021
Messages
217
Likes
90
AptX Low Latency is good enough for everything from converting an XLR condenser mic into a wireless camcorder mic with a digital Parametric EQ, perfect lip-sync for computer Games and Movies / 160 foot range / no dropouts even near microwaves or even used in a furnace room with lots of metal objects 2 floors away, and sounds just as good as AptX HD

Anything "better" must be lower latency and longer range, transmit as well as recieve, and have optical input and output as well as analog

Do you have any "better" stereo bluetooth devices like that?

Diddn't think so!

I already have everything I need in a bluetooth device
Anything else is "currently" a waste of time


Maybe next year
 
Last edited:
Joined
Feb 21, 2021
Messages
64
Likes
14
Is this the only LDAC BT Receiver with Line Out? I'm looking for one that I can attach to my AMP permanently.

Also does anyone happen to know the shortest 3.5 to RCA?
 

PeteL

Major Contributor
Joined
Jun 1, 2020
Messages
3,303
Likes
3,838
Is this the only LDAC BT Receiver with Line Out? I'm looking for one that I can attach to my AMP permanently.

Also does anyone happen to know the shortest 3.5 to RCA?
They are plenty, list too long to post here, but this one looks competently designed
 
Joined
May 5, 2021
Messages
49
Likes
23
Location
North Carolina
This is a review and detailed measurements of the Topping BC3 Bluetooth receiver and headphone amplifier. It was sent to me by the online seller, Yaoyaotiger HIFI Audio Store and costs US $69. This is a follow up to our previous tests of Bluetooth receivers and investigating performance of various codecs supported within.

The shiny plastic and slick design sets the BC3 apart for the typical Topping desktop products:

View attachment 132115

The touch screen buttons are pretty but with no tactile feedback, they are kind of inconvenient to use. There is a lock button to deactivate them which is nice:

View attachment 132116

Nice to see both headphone amplifier and digital/analog line outs. There is a convenient Toslink mini to full size adapter in the box which I used for my testing. The unit has battery built-in so you can charge it and use it on the go (car, commute, etc.). I did all of my testing while I powered it through USB-C from my computer monitor's hub (same port I use for all of my DAC testing).

Topping BC3 Digital Measurements
This is the first Bluetooth receiver we are testing that has digital output. This is exciting because we can see the true performance of the various codecs without the performance of the DAC getting in the way. The source for the stream was my Samsung S8+ phone running Roon player with Android developer options enabled so I could select the Bluetooth codec. Prior to this test, I was not successful in getting LDAC to work. I figured out the issue here. Turns out in their infinite wisdom, whoever did this work decided that you must enable LDAC both in the developer options and Bluetooth settings!

Let's start with aptX:

View attachment 132117

We immediate see the signature of aptX with that stepped down response in frequencies above 5 kHz. The higher frequencies are more "random" (have higher entropy) so harder to compress. By lowering their levels or truncating them, there is less for the codec to do.

Switching to AAC shows the poor quality we have seen which I am guessing is due to Android implementation of AAC encoder (meant to test this with an iPhone but forgot -- will do so later):

View attachment 132119

We get a big boost in SINAD but I am not happy about all the artifacts. In my prior testing I found SBC to actually sound better:

View attachment 132120

So our previous conclusions more or less hold with the codecs being the limiting factor not the DACs. However, now we have a new kid on the block, namely the Sony LDAC codec. Is it as good as its hype? Let's start showing its performance with 16 bit setting of the codec (source is still 24 bits):

View attachment 132121

Performance jump quite a bit showing a SINAD which is the limit of 16 bit bit depth. Switching to 24 bit mode is this tester gets a big smile:
View attachment 132132

Wow! What an improvement. Then again, all the other codecs should have done excellently with this simple sine wave that is dead easy to encode. But they didn't so LDAC in 24-bit mode wins big time. Even in 16 bit more it leaves the other codecs behind.

All of these measurements are in digital domain. Let's fee the Toslink to a Topping D30 Pro and measure its analog output using LDAC:
View attachment 132123

Hallelujah! We get state of the art analog reproduction despite using a wireless connection that started life for just voice.

Topping BC3 Analog Output Measurements
The DAC here is not a heroic implementation so we are going to get decent but not state of the art fidelity when using LDAC:

View attachment 132124

Dynamic range is basically representative of 16 bit format:

View attachment 132125

Headphone out is essentially the same as line out:

View attachment 132126

Edit: A couple more measurements:

Here is an asynchronous frequency sweep comparing pairs of codecs:

View attachment 132356

That is one nice and flat response for LDAC. SBC on the other hand is filtering above 14.5 kHz. This is done because high frequencies are much harder to encode due to their high speed variations (in sample value)/noise.

I was surprised that AAC was also filters above 15 kHz:
View attachment 132357

aptX as we saw earlier plays some serious games in treble.

This type of sweep also generates THD+N relative to frequency:
View attachment 132358

We see that LDAC continues to be the cleanest codec by far. Note that this is a dynamic sweep signal so more complex than a single tone that dashboard uses. The other codecs trade places in different parts of the spectrum and are generally a mess. aptX by the fact that it chops off high frequencies does better at very low frequencies.

LDAC Codec Licensing
Searching for LDAC lands you on the English translate of a site that is likely hosted by Sony Japan. In there we see the comforting mention of Topping as proper licensee of the codec:

View attachment 132129

Earlier it mentions that the source code for LDAC is part of the Android Open Source. Don't confuse this with it being fee. You still need to license the underlying patents from Sony or they can come after you. What is involved in said licensing? Sony is unusually transparent about that:

View attachment 132130

Predictably they expect you to spend months negotiating the NDA and actual license. Having been on both sides of such licensing, I can attest to this process being long and drawn out because the agreements likely start very one sided which you then kill yourself across countless conference calls to try to soften. Typical clauses are "covenant not to sue" meaning once you license this technology, then you can't sue Sony or their other licensees over any patents you may own. I don't know if such a provision is in there but I am guessing it would be. On NDA front you likely want to have a residuals clause meaning what you remember can't be unremembered. They will say you have to have a button to erase your memory if NDA is terminated.

And oh, after all the above you have to pay the license fee. I see no provision for RAND licensing (Reasonable and Non-discriminator Licensing), meaning everyone pays the same fee. Without it, a competitor to Sony could be charged much more money than a friend and thereby keeping a competitive advantage.

Did someone complain about MQA being bad? Welcome to the world of audio licensing! :)

LDAC Codec Range
I did most of my testing with the phone being a few feet from the BC3 receiver. I did however test to see if I could get performance to drop but could not even after I moved my phone some 20 to 25 feet (7 to 8 meters) away. I had the codec in adaptive mode by the way so it could have switched encoding rates but I saw no glitches to indicate it did so. Your mileage may vary of course depending on the implementation of Bluetooth transmitter in your device.

Conclusions
The previous Bluetooth receivers we tested suffered both from codec losses and less than optimal implementation of output DAC. Topping has the same limitation as far as codecs but since it supports LDAC, for simple signals at least, it eliminates that bottleneck. Its DAC implementation is what you expect: "good enough for 16 bit content." Fortunately with digital output you can connect it to your favorite DAC with balanced out and by using LDAC, you have a superb input path for Bluetooth.

I have not yet tested enough Bluetooth receivers to have a good calibration for what is good and what is great. Still, I was very pleased with the performance of the Topping BC3 and am putting it on my recommended list.

EDIT: video review just posted:


------------
As always, questions, comments, recommendations, etc. are welcome.

Appreciate any donations using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/


Any chance you could make a video explaining (on certain devices) why the left and the right channels are not the same when it comes to output. One for distortion values and another for volume. Why are they different and is it a manufacturing issue or just poor equipment from the getgo?
 

PeteL

Major Contributor
Joined
Jun 1, 2020
Messages
3,303
Likes
3,838
AptX Low Latency is good enough for everything from converting an XLR condenser mic into a wireless camcorder mic with a digital Parametric EQ, perfect lip-sync for computer Games and Movies / 160 foot range / no dropouts even near microwaves or even used in a furnace room with lots of metal objects 2 floors away, and sounds just as good as AptX HD

Anything "better" must be lower latency and longer range, transmit as well as recieve, and have optical input and output as well as analog

Do you have any "better" stereo bluetooth devices like that?

Diddn't think so!

I already have everything I need in a bluetooth device
Anything else is "currently" a waste of time


Maybe next year
Keep in mind that most OS, Browsers, and media players adapted to the Latency of the various Bluetooth Codecs by now. I don't know of Games, I don't play those, but for Video Playback, syncing image to audio is an issue of the past, I see very limited appeal to AptX LL nowadays.
 

Svet

Member
Forum Donor
Joined
Nov 16, 2018
Messages
17
Likes
4
Location
Canada
There's a fair number of such transmitters that accept optical or analog connections. I'd be dubious of most ADC's on a combo <$100 device *if* highest possible SINAD is your chief concern. Audibility is another story, as always.

@amirm thanks for testing this, I'm always curious about these little adapters because they can offer such easy fixes compared to wires. As an Android user, a few key points:

There's some kind of bug/feature that prevents you from locking LDAC sample rate. It always defaults to Best Effort once the connection is lost, and you have to reset it manually via developer options once connection is re-established. Even as an advanced user, this is tiresome yet I'm not aware of any fix. Maybe rooting? I'm not really in that space because, working from home, I need maximum reliability on my phone.

Many android devices, including my S10 5G, have dropouts/stuttering with the LDAC rate set to 990. It must be a fairly prevalent issue because the Qudelix app offers suggestions on how to work around it during configuration. In my case, I have to completely disable Location services but many people simply have to keep their screen off. Another deal breaker for a typical user.

Finally, latency on 990 LDAC can be as high as 250ms if not higher (device dependent). Obviously a non-issue for simple playback but videos can degrade into kung-fu mouth.
I had a Samsung S8, paired with the Sony XM3. Lot of frustration as the signal was garbage even at 20cm from the phone, in 990. I have now an S21, and the connection is solid all the time, thanks to the better bluetooth radio and dual antennas.
 

dougi

Addicted to Fun and Learning
Joined
May 31, 2020
Messages
845
Likes
763
Location
ACT, Australia
They are plenty, list too long to post here, but this one looks competently designed
I have an iFi Zen Blue and recommend it. It has a good DAC, digital out if required, external antenna, indicates the codec and sample rate used. It also has balanced out via a weird connector if you need it.
 
Top Bottom