• 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!

Bluetooth codec comparison

bachatero

Senior Member
Forum Donor
Joined
Dec 25, 2023
Messages
426
Likes
431
I can't believe there is only ONE article on the internet (https://addictedtoaudio.com.au/blog...-the-bluetooth-codecs-which-performs-the-best) comparing all the widely available Bluetooth codec options to see which one performs the best. So, I decided to do my own. Here's the gist of it:

Phone/PC BT -> 1Mii B06HD+ -> Scarlett Solo 3rd gen -> Phone USB-C

The performance of the 1Mii and Scarlett shouldn't be a problem to us because the Scarlett's ADC is surprisingly good (unlike the 4th gen) and the 1Mii uses an ESS DAC. Although I haven't properly tested the 1Mii's performance, it seems good enough as you'll see soon.

I played a 1khz sine through the left channel only (only 1 channel supported in the Solo), and got the following screenshots. There's a rub, though. Android adds some extra junk processing so this doesn't accurately represent how the codecs themselves perform, but rather them plus Android. So, I did some extra testing with my Linux PC (I use Pop!_OS btw) which supports all the codecs Android does plus more.

Here's our baseline, SBC with Android.

Screenshot_20241210-161555-952~2.png

Not great, not terrible. There's a huge spike of SOMETHING to the left, and it's not harmonic distortion (but rather subharmonic). How about AAC?
Screenshot_20241210-161634-874.png

I should tell you now that the huge spike of SOMETHING is due to Android. So, ignore it. AAC doesn't look too terrible either except for all the ugly noise surrounding our test tone. AptX claims to deliver high quality audio. Does it?
Screenshot_20241210-161656-608.png

We've still got noise, but it looks different. AptX HD claims to deliver even higher quality audio. Does it?
Screenshot_20241210-161719-451.png

Our noise is looking a lot better! However, there still remains a little distortion. Let's top it off with LDAC.
Screenshot_20241210-161738-282.png

Now we're talking! Our noise is just about gone. But now our distortion is higher in the higher frequencies. I'm not sure if you can hear it that low, though.

Here's the same screenshots but using the BT from my PC instead. Note that we have two extra codecs here: AptX LL and SBC-XQ. SBC-XQ is just SBC with a larger bandwidth and support that depends on the device. Interestingly, it's also only available on Linux.

Screenshot_20241210-163545-246.png

Screenshot_20241210-163622-856.png

Screenshot_20241210-163647-619.png

Screenshot_20241210-163712-318.png

Screenshot_20241210-163736-540.png

Screenshot_20241210-163759-506.png

Screenshot_20241210-162910-041.png

Conclusion: SBC is by far the worst choice, especially if you're using it on Android. LDAC is by far the best with performance limited largely by DAC/ADC (?). AAC and AptX have less distortion than SBC but add a LOT of extra noise. Can you hear any of this? I've taken the test (https://btcodecs.valdikss.org.ru/sbc-encoder/) and it's honestly hard to hear the junk with real content. But then again, that simulation is in perfect conditions using the codecs you'd find on Linux. Regarding SBC-XQ, support depends on the device and so here it didn't appear to make any tangible difference. AptX-LL didn't either.

Recommendations: Don't use SBC, AAC, or AptX where quality is important, unless you're using an iPhone/iPad which has a better AAC codec. AptX-HD and LDAC are the only options that offer both low noise and low distortion so prioritize those. And as always, remember that no matter what codec you choose, distortion and noise are hard to hear with real content unless you're in an ideal listening environment.

I hope this helps!
 
Last edited:
Your test signals vary by a few dB from 16-19 making comparisons a little more difficult but for the most part better than I expected. Anything in the -80dB is just barely passing on distortion free sound but many of the codexes above dip into 90s aside from a few spikes. Even SBC on Android presents a fairly decent signal but I can see where things might get borderline if you've got some very resolving and sensitive headphones. I don't use bluetooth but based on the above info I wouldn't hesitate to for non-critical listening.
 
What would be a good way?
ABX testing with real program material. ;) These CODECs are dynamic so pure-constant test-tones don't tell you what they are normally dong. And they mostly work by trying to throw-away sounds that are masked by louder "more-important" sounds. Developing a measurement for that would be a lot like they way they develop & tune lossy compression algorithms... With a lot of blind listening tests!

At high bitrates, when an MP3 isn't transparent it's usually a temporal artifact called "pre-echo" that only shows-up (audibly) with transient sounds. I think other CODECs may have similar issues. I'm sure that can be measured with an impulse, but the measurement may not correlate with audibility.
 
Haven't look in on hydrogen audio forums in quite some time. Testing codecs like this was a bread and butter topic over there. ABX testing is the only kind they discuss. So might be worth looking over there to see what has been done.
 
SBC-XQ is also available on several 3rd party Android-based OSs like LineageOS. You've missed Opus which Google added to Android for the Pixel Buds, and PipeWire (linux) has since picked up. PipeWire also has a different Opus profile that nothing else uses.
 
I can't believe there is only ONE article on the internet (https://addictedtoaudio.com.au/blog...-the-bluetooth-codecs-which-performs-the-best) comparing all the widely available Bluetooth codec options to see which one performs the best.
??
From the review:
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.
 
Not seeing much commentary on the actual content but rather a lot of complaining
 
I certainly appreciate the effort. Thank you. Saying that I do believe that blind testing using actual music and multiple subjects/trials would be a great adjunct.
 
I can't believe there is only ONE article on the internet comparing all the widely available Bluetooth codec options to see which one performs the best.
Well, you're right not to believe it ;-)

On top of Amir's BC3 review listed above, there is more here:
- I reviewed the BTA30 Pro.
- There are more measurements in this thread.

I just listed the ones I knew because I was involved.
But you may find quite a few other measurements over here...

Note: The missing protocol in those measurements is AAC, which under performs notably when used from Android or Windows PC.
 
Last edited:
Thanks for making the effort to do these comparisons! I would agree with @DVDdoug that THD and noise are not considered the biggest weak point of these codecs in terms of audible artifacts, but it's good to see the data anyway.
 
AptX has an advantage in common use cases because most users play back and stream lossy rather than lossless sources. This means transcoding is commonly involved. In such situations, AptX, as a subtype of lossy codecs, is less susceptible to transcoding artifacts. I encourage everyone to conduct their own listening tests rather than taking my word for it. Generally, AptX does not significantly increase artifacts with characteristics like warbling, flanging, or smearing. This is because AptX operates on a different principle compared to codecs that produce such artifacts, such as MP3, AAC, and SBC. AptX is more closely related to LossyWAV, ADPCM, and WavPack hybrid mode.
 
Back
Top Bottom