Fiio BTA30 Pro Bluetooth Transmitter and Receiver + DAC - Review

[Rja4000]

Hello

Here is a review of a special device:

This device is a Bluetooth Audio Transmitter, Receiver and DAC.

This is the Fiio BTA30 Pro.

The BTA30 Pro is about half the dimension of the (1U / half 19") RME ADI-2 Pro fs R in all directions:
Half width, half height (without the antenna), half depth.

There is a switch to select mode:
RX (Bluetooth receiver), TX (BT transmitter) or DAC.
There is also an input selector, for TX or DAC mode, and a BT Codec selector (TX mode). The LED has a different color according to active Codec.

It supports following CODECs
For Transmission:
SBC, aptX, aptX HD, aptX LL, LDAC
For Reception:
SBC, AAC, aptX, aptX HD, LDAC
It may also act as a DAC

Inputs are USB-C, SPDIF Optical or Coaxial; Output are SPIF Optical or Coaxial and a pair of Analog RCA
The USB-C also provides power (NB: The device does not seem to be Power-delivery compatible)

The device costs around 140€ ($120)
It has been kindly lent to me by Audiophonics (@Audiophonics - France).

PART I : Fiio BTA30 Pro as a Bluetooth Transmitter

To test it, I used a Topping BC3 (also lent by Audiophonics) connected to my RME ADI-2 Pro fs R with TOSLink (optical SPDIF).
That's a purely digital connection.
Measurements are performed with Virtins Multi Instrument 3.9.6.2.


Validating measurement chain

So, let's start by validating the Topping BC3 receiver by performing some measurements on it.
It should be easy to validate them, since Amir tested the device (see here)

I used my Samsung Note 9 to send a test signal in LDAC mode to the BC3.
The BC3 is connected to the RME through its optical output (no DAC is involved).
The RME displays the input sampling frequency and bit rate.
For LDAC, it remains at 96kHz 24 bits.

So, first, I tested with a Test signal at 1kHz 96kHz 24 bits at 0dBFS.

Validating Topping BC3, signal sent from Smartphone, SINAD

Obviously, there is some saturation here.

Let's try -1dBFS
Validating Topping BC3, signal sent from Smartphone, SINAD (-1dBFS)

Much better

And close enough to Amir's 112.7dB SINAD

What about 16 bits (24 bits signal) ?

Validating Topping BC3, signal sent from Smartphone, SINAD (16 bits mode)

Again, we are very close to Amir's figure (97.8dB SINAD)

So the Topping BC 3 receiver measures as it should.
We may now proceed to the Fiio BTA30 Pro measurements.


Fiio BTA30 pro as a Bluetooth Transmitter

Measurement chain is as follows:
Virtins Multi Instrument (PC) - (USB) > RME ADI-2 Pro fs R - (SPDIF Optical Out) > Fiio BTA30 Pro - (BT) > Topping BC3 - (SPDIF Optical Out) > RME ADI-2 Pro fs R - (USB) > PC
(For those who wonder, the results with a direct USB connection out of the PC are just identical. I used the RME output since I couldn't get MI to work with Fiio ASIO driver.)

All CODEC except LDAC were tested at 48kHz sampling frequency 24 bits

First, SINAD dashboard

aptX Codec

aptX LL

NB: I have a hard time understanding the difference between aptX and aptX LL. All measurements seem the same, to my eyes.

aptX HD

That's better

SBC

LDAC (@96kHz, -0.2dBFS)
As we know already that the BC3 does saturate at 0dBFS, I played a bit with the level, and maximum I could go is -0.2dBFS.

As seen above, this is a bit shy of the 112dB SINAD that I could get with the Smartphone.
However, this is another use case here, since the signal may directly be sent from the PC.

Very good result, anyway.


Attempt to measure Dynamic range

I tried to measure noise and dynamic range.
As per AES, a -60dBFS signal is to be used, and then you measure the remaining SNR and add the level difference to 0dBFS.

With all Codecs but LDAC, it seems the signal is completely muted below -54dBFS.
There is simply no signal.

EDIT (2): I re-tested with the smartphone sending a test signal at -60dBFS to the Topping BC3 receiver, and it worked.
So the Fiio BTA30 pro seems to be muting its output when the level is below -54dBFS, for all CODEC but LDAC.


So I used -54dBFS as a level to measure the dynamic range for each CODEC (except for LDAC).


aptX (identical to aptX LL)

aptX HD

SBC

Now we see the difference between aptX and SBC: The later seems to favor the noise reduction, while the former has lower distortion.


LDAC (Test signal at -60dBFS)

Now we are speaking !

(Even if we should decrease this by 0.2dB, since we know the signal chain saturates above -0.2dBFS)

Note, first, that this is a purely digital chain. No DAC is involved.
Also, all those CODECs play with the level and frequencies somehow.

So this should not, in my opinion, be compared with a "normal" DAC Dynamic range.

Still, we have an interesting result.


(The BC3 from smartphone gives a similar result)


Attempt to measure Frequency response

NB: Following measurements are using LDAC Codec only.

My "normal" measurement method implies using a signal which is synced between Output and Input.

Unfortunately, the clock in seems to be quite shifting versus the clock out, so the signal is not perfectly in sync, and the measurement is therefore not stable.
In practice this is no problem (you only listen to the output), but for my measurements it makes it more complex.

We see quite a few measurement artifacts.

At 48kHz sampling frequency

At 96kHz

I therefore used a sweep from REW to measure frequency response

Similarly, the response around 20kHz is very weird.

This should not be a problem in practice, though.
We see a -2dB cut at 20kHz.

32 tones

32 tones is also requiring a good sync between Out and In signals.
So I couldn't use the "Rectangle (=No) windowing function for FFT I normaly use
And the low frequencies are therefore mixed.

Nevertheless, here it is

(The test signal was re-generated using Audio Precision's APX500 Flex Demo software for 96kHz.
It should therefore be identical to the signal Amir is using, except for the sampling frequency. It has exactly the same crest factor, frequencies and phases..)

Don't worry about frequencies below 100Hz.

Ok, this is not as clean as a standard DAC.
(Look at vertical scale, reduced to an 100dB range. Usually, we use a much wider range, over 150dB or 200dB)
We also see strong harmonics above 20kHz.

That's not too bad, though.


(Partial) Conclusion - TX mode

As a Bluetooth Transmitter, the Fiio BTA30 pro does a pretty good job.

One may object that it is (slightly) outperformed by a Smartphone, but this is only when you deep dive into the smartphone development parameters that you'll be able to force it to use the best CODEC. Most of the time, this is quite random.

You have much more direct control on the CODEC selection with the Fiio BTA30 pro.
The BTA30 Pro's LED indicator also tells you which CODEC is used.

Also, the Fiio BTA30 Pro works very well from a PC, a TV, or any device with an SPDIF output, which is very convenient.

UPDATE:
A user below complains about the abrupt muting for all CODECs but LDAC below -54dBFS.
This is a low level, but I understand this may be unpleasant in some circumstances.
I mentioned the fact in the dynamic range chapter, but highlighting this may be important, indeed.

If you use LDAC only, this is of no consequence.
But if you plan to use other CODECs, you'd better be aware of this behaviour.

 

[Rja4000]

PART II : Fiio BTA30 Pro as a DAC

To test it, I connected the Fiio BTA30 Pro to my PC using an USB cable. (I used the Topping HS01 USB filter)
One channel of the Fiio BTA30 Pro's analog outputs is connected to my RME ADI-2 Pro fs R analog inputs (with an Y cable, to feed both inputs).
Measurements are performed with Virtins Multi Instrument 3.9.7 (brand new build, released yesterday, including a fix for the Fiio ASIO driver)


SINAD dashboard

Here is the standard Dashboard at 2V (-3dBFS)

The DAC gives 3Vrms at full output.

Best performance is around -6dBFS (1.5Vrms)

The other channel is 2-3dB lower.
It still outperforms the Specs (THD+N <= 0.0008% or -102dB)

SINAD vs Level

If you prefer Amir's graph of SINAD vs output Voltage, here it is

Dynamic range

We see a dynamic range of almost 113dB
A bit short, compared to the 118dB SNR announced.

That's 16dB better than the Topping BC3, though.

IMD vs Level

A hint of ESS hump around -25dBFS, maybe ?


Linearity

This is quite good down to -114dBFS (19 bits)


32 tones (Multitone)
(input SPDIF - 192kHz)

Jitter
NB: the Topping HS01 USB filter is still between the PC and the BTA30 Pro.

Low Pass Filter

Here is the default Low Pass filter

Comparison of the 4 filters

They may be selected from the Android app:

The best option seems the default one.

(Partial) Conclusion - DAC mode
While not as competent as some recent standalone DACs, the Fiio BTA30 Pro DAC section performs better than a throw away - last minute addition.

It's far from state of the art, even at that price, but there is nothing broken here, and performance is coherent with the limitations of Bluetooth audio transmission Codecs, including LDAC.

 

[Rja4000]

PART III : Fiio BTA30 Pro as a Bluetooth Receiver

All tests were performed in LDAC 96kHz 24 bits
Source is my Samsung Note 9.


First, let's check how the Fiio BTA30 pro measures when receiving LDAC Bluetooth and sending the signal through its SPDIF optical output

SINAD Dashboard - Optical output

Woaw. That's disappointing.
This is far from the level of the Topping BC3.

But we see there is some saturation.
Let's lower the level.

Saturation goes away just below -1dBFS.

Still, we are far below the 112dB SINAD we see from the Topping BC3

So what's happening ?

It seems we are limited to 16 bits resolution.

Here is the 24 bits Jitter test signal

This looks like a 16 bits signal truncation to me.

If I try the 16 bits 44.kHz signal, here is what it looks like.

This is also confirmed by the RME ADI-2 Pro fs R display (when I turn "SRC" off on the optical SPDIF input).

So we have something strange here.

It may be that my smartphone negotiates only 16 bits with the BTA30 Pro (although it clearly does it well for 24 bits with the Topping BC3).

Nevertheless, the Fiio BTA30 pro performs as we could expect given this 16 bits limitation :

Dynamic range (16 bits)
Optical output

Frequency response
We still see this -2dB at 20kHz
(Cut in low frequencies is due to the measurement method)

And finally, the

SINAD dashboard (16 bits)
Analog output (2V)

Again, very much the maximum we could expect from a 16 bits resolution.

So, anyway, enough for CD quality reproduction.
And still a few dB better than the Topping BC3 in analog output mode.


UPDATE 2022-11-18
As I explained, I reached to Fiio Support to ask for explanations and guidance in solving this issue.
They answered.

"Please check this option in FiiO Control: Compatible and Up should be selected."

The "Up" option stands for "Up sampling".
Fiio proposes, with this option, to upsample the Bluetooth input to maximum sampling frequency.
As this is modifying the signal for no obvious benefit, I didn't check it.

So here is the result when I do that:

So here we go: We have 24 bits sampling.
OK, we have 192kHz as well.

Then how does it measure ?

SINAD Dashboard - Over sampling mode
(96kHz 24 bits signal, optical output)

This is just a bit better.
We are still VERY FAR from the >112dB of the Topping BC3 in the same conditions

And Dynamic range doesn't get any better.



(Updated) Conclusion - RX mode

Of course, the 16 bits (default) limitation is strange and disappointing.

My view is that the limitation to 16 bits is a result of the combination of my smartphone with the BTA30 pro,
It's quite possible that another smartphone or Bluetooth source would behave differently.
(Read my additional posts below for further analysis.)

I asked Fiio support for feedback but didn't get any answer so far.
After the Fiio support answer, it is likely that this is the normal behaviour for the RX (receiver) mode.

We can say that the Fiio BTA30 Pro measures flawlessly in 16 bits LDAC mode.
Performances are as good as the 16 bits format allows.

And, since it's DAC section is better, it outperforms the Topping BC3, when used as a receiver with the Analog output.

So, although we were hoping better in this mode, the performance is good enough for CD quality use.
Which is the most important.

Oddly enough, the TX and DAC are much more performant.

The up-sampling brings a very slight benefit. I wouldn't bother to switch it on, personnaly.

The "up-sampling" mode seems nothing else than a marketing addition, in my eyes.
It adds very little to the performance.


Overall conclusion

The Fiio BTA30 pro is, overall, a competent device.

It's one of the very few devices that will allow you to Transmit signal over Bluetooth with LDAC CODEC at better than CD quality from a PC or a Toslink/SPDIF source.

Its DAC module is competent as well.

Finally, and although it's less performant than the other modes, the receiver is good enough for CD quallity.
Even if you take the Analog conversion into consideration.

If all you need is a Bluetooth receiver with a digital output, the Topping BC3 is a cheaper and more performant option.
But as soon as you need the analog conversion or if you want a transmitter, the Fiio BTA30 Pro is probably the way to go.

UPDATE:
See also the comment at the end of the Transmitter section:
With all CODEC but LDAC, in Transmitter mode, the BTA30 Pro will mute the signal if the level is below-54dBFS, which may be annoying.
This was mentionned above, in the Dynamic Range section of the first post (Transmitter mode).
If you use LDAC only, this is of no consequence.
But if you plan to use other CODECs, you'd better be aware of this behaviour.

 

[MCH]

Thank you for doing all this work. I believe it is the first time I see aptx hd measured. Good to see that it is somehow better than standard aptx, but also confirms ldac is the way to go. I don't use Bluetooth at home but anyways.
Thanks!

 

[Romario]

Great work! I'm curious how the ES DAC performs in this little Fiio. It's advertised as portable but it never left the main system. CD playing over BT LDAC is quite a treat.

 

[VintageFlanker]

Oh, I missed this...

Fantastic job, @Rja4000 ! Especially regarding BT codecs: I was just planning something similar in the near future (SBC VS AAC VS aptX VS aptX HD VS LDAC, with either E70 or M500 mkIII as a reference).

 

[Rja4000]

Thanks

I'll attempt to test latency as well.
This is what LL (in aptX LL) is about after all:
LL as in "Low Latency".

 

[hodmx]

Thanks a lot, I am trying to decide between this and the SMSL D6. This is less expensive, smaller and have the transmitter but the SMSL seems to have a better DAC.
Will wait for the receiver data before I take my decision.
Right now I have a bta30 connected via optical to a Yamaha wxa 50, and I feel that the DAC of the wxa 50 is better than the one in the Fiio. So, I hope that either, the bta30 pro or the SMSL will be an improvement.

 

[Romario]

Impressive measurements for such a small unit. No need for an upgradet DAC. And it's powered by my Chromebook on battery, so no electrical noise added, I presume. But that's me just thinking out loud because I don't know what those influences are or how clean a chromebook puts out it's power throug USB.

 

[Rja4000]

Firmware upgrade
I've upgraded Fiio BTA30 Pro's Bluetooth firmware to 1.3 (was 1.0)

No cigar. Still stuck in 16 bits in RX mode.

Compared to BC3 as a receiver
With exactly the same settings on smartphone, I switched to the BC3.
Immediately, the RME tells me "24 bits"

And the SINAD dashboard (BT LDAC in optical out) displays values accordingly (even better than last time)

... and to Topping D90LE (analog output)

I tried to feed the Topping D90LE in Bluetooth directly, with same smartphone settings

The Topping doesn't display the bit rate, unfortunately.

But anyway, again, performance is as expected for a 24 bits signal.

Fiio BTA30 pro to Topping D90LE
Switching to the Fiio BTA30 Pro instead of the smartphone as a Bluetooth transmitter, feeding the Topping D90LE, I get this in LDAC

I finally tried it in aptX HD

PC (USB) > Fiio BTA30 Pro (BT) > Topping D90LE (Ana) > RME ADI-2 Pro fs R (USB) > PC

There is obviously an handshaking issue between the Samsung Note 9 I'm using and the Fiio BTA30 Pro.
They systematically end up with 16 bits transmission instead of 24 bits.

I've seen tests elsewhere on the internet, done with another Bluetooth Transmitter, which suggest the Fiio BTA30 pro is capable of 24 bits reception. So this seems a specific issue for my very combination of transmitter (Samsung Note 9) and receiver (Fiio BTA30 Pro).

This is, by the way, what interests me in this device: to be able to use it as a Bluetooth transmitter, directly from the PC (with an ASIO driver), and have those settings much more under control, allowing proper test of various Bluetooth receivers.


I have to say I have no clue what may cause this.
And I have no other LDAC capable Bluetooth transmitter at home to try.

So I give up for now.

If anyone has an idea...

EDIT: I've sent a mail to Fiio support, to check if they have an idea how to fix this.

EDIT 2: They answered.
I updated the Part III above with the results.

 

[Rja4000]

For those of you in debugging mood, here are the setting of my smartphone

 

[Romario]

In the Fiio app there are 4 filter settings for TX/DAC mode. I don't know if such kind of settings could influence the signal.

Edit: When paired in LDAC mode TX you can't connet to the app remotely. ( 1 pair only)

 

[Rja4000]

In the Fiio app there are 4 filter settings for TX/DAC mode. I don't know if such kind of settings could influence the signal.

Edit: When paired in LDAC mode TX you can't connet to the app remotely. ( 1 pair only)

I've added a comparison of the 4 filters in part II above.

 

[Rja4000]

I've re-tested Dynamic range, to update graphs with new format (less confusing with SINAD Dashboard)

At the same time, I ran the test with Smartphone, and corrected the comment: the Topping BC3 doesn't mute the signal at -60DBFS.
The Fiio TX does.

 

[MCH]

Dynamic range of 113 out of the analog for a Bluetooth transceiver seems very decent actually. You measured this plugged, and the dynamic range of the actual Bluetooth transmission on ldac is basically clean. Do you think it can be inferred that the dynamic range of the analog out from a signal received in ldac will have a dynamic range of 113db? That would be quite impressive (to me dynamic range is more important than Sinad hence my interest).
Additionally: I have read that the signal range of the high quality ldac tends to be very limited. What is your experience with the fiio?
Thanks again for all this work!

 

[Rja4000]

Dynamic range of 113 out of the analog for a Bluetooth transceiver seems very decent actually. You measured this plugged, and the dynamic range of the actual Bluetooth transmission on ldac is basically clean. Do you think it can be inferred that the dynamic range of the analog out from a signal received in ldac will have a dynamic range of 113db? That would be quite impressive (to me dynamic range is more important than Sinad hence my interest).

Obviously, yes

Here is the Dynamic Range of the analog output of the Topping D90LE
receiving Bluetooth signal from the Fiio BTA30 Pro (LDAC, of course)

(PC (USB) > Fiio BTA30 Pro (BT LDAC) > Topping D90LE (BT In, Ana out - both channels summed,, 4V range) > RME ADI-2 Pro fs R (Ana In - Both channels summed, +4dBu range, USB out) > PC)

Additionally: I have read that the signal range of the high quality ldac tends to be very limited. What is your experience with the fiio?
Thanks again for all this work!

I didn't test that.

 

[MCH]

Obviously, yes

Here is the Dynamic Range of the analog output of the Topping D90LE
receiving Bluetooth signal from the Fiio BTA30 Pro (LDAC, of course)

(PC (USB) > Fiio BTA30 Pro (BT LDAC) > Topping D90LE (BT In, Ana out - both channels summed,, 4V range) > RME ADI-2 Pro fs R (Ana In - Both channels summed, +4dBu range, USB out) > PC)

View attachment 243156



I didn't test that.

I meant ldac signal > fiio > fiio's analog out

 

[Rja4000]

I've done attempts to measure the Frequency response for various CODEC with REW

Not really successfull, to be fair...

 

[Rja4000]

I meant ldac signal > fiio > fiio's analog out

Well, that's exactly the issue:
As long as the Fiio receiver remains limited to 16 bits, we can't reach more than around 100dB DR.
If this is solved, we'll probably be able to go above that ceiling, as we can see neither the Fiio BTA30 pro DAC stage nor the Bluetooth CODEC is a limiting factor.

 

[Rja4000]

No reply from Fiio support so far
So I updated the conclusions.