(Unofficial) Topping D50 III DAC review and measurements

[discountdaikon]

Hello, just registered after lurking for a long while.
Thank you for the review, as it seems to still be the only one I can find online.
Question about the PEQ via USB.
If using exclusive mode, ASIO or some such in foobar, with PEQ enabled on the DAC, will the DAC automatically change the sample rate?
Or is everything set to 16bit 44.1 when PEQ is enabled on the DAC?

Edit: I ask because when using EAPO/Peace, everything in my set up gets set to whatever the Windows sampling rate is set to.
Cheers and have a great day everyone.

 

[Rja4000]

Offtopic questions: How long will it work after installation? Are there any known limitations of the trial version?

The demo version doesn't allow any measurements. Just to play with the software and look at prerecorded projects and results.

I asked for a fully functional trial, since I'm interested in their subscription model, if that makes enough sense (ie brings enough benefits).
This requires that they send you a key with a temporary license on it.
The key is not a simple USB key: there is some hardware in it (including a clock, so no, you can't just cheat the clock).

Unfortunately, I asked the key for a specific period when I knew I'd be able to spend some time with the software, but they started the period too early (without telling), so when I started using it, I discovered the software would expire soon - April 12th.
So I was only able to evaluate part of it.

Frankly, as far as I could try in this limited time, the benefit for me is short compared to the price they ask (1200€ a year for an amateur...).

There are benefits, for sure, but due to some strong limitations that I wasn't able to circumvent in the time I had, I still have to use another software for several measurements.
(All measurements that require a different DAC and ADC driver or sampling frequency, and also the dynamic range - although I have an idea how to do this last one).

I was hoping that the automation of their "sequence mode" would help justifying the cost, and also bring some measurements I'm currently not able to do, but I didn't have time to check it.

Well, I still have to decide.

 

[johny_2000]

The demo version doesn't allow any measurements. Just to play with the software and look at prerecorded projects and results.

I asked for a fully functional trial, since I'm interested in their subscription model, if that makes enough sense (ie brings enough benefits).
This requires that they send you a key with a temporary license on it.
The key is not a simple USB key: there is some hardware in it (including a clock, so no, you can't just cheat the clock).

Thanks for the details. Yes, for amateurs it's not worth it.
I used REW and RightMark Audio Analyzer for free and was able to get similar results, except for the not-so-familiar appearance of the measurement reports.

 

[ocinn]

Well this is just further proof that the DAC problem is solved. Every few weeks we get a cheaper and cheaper product, which manages essentially state-of-the-art performance

Still waiting for the day that topping/smsl/fosi/etc release a MiniDSP Flex competitor. Would love to see this level of DAC performance alongside modern true two way crossovers and PEQ for room correction, as well as some analog inputs

 

[Sokel]

Well this is just further proof that the DAC problem is solved. Every few weeks we get a cheaper and cheaper product, which manages essentially state-of-the-art performance

Still waiting for the day that topping/smsl/fosi/etc release a MiniDSP Flex competitor. Would love to see this level of DAC performance alongside modern true two way crossovers and PEQ for room correction, as well as some analog inputs

What I would love is that all the above create the same quality filters as RME does for example.
None of them is quite there yet.

 

[Yoku-San]

Well this is just further proof that the DAC problem is solved. Every few weeks we get a cheaper and cheaper product, which manages essentially state-of-the-art performance

Still waiting for the day that topping/smsl/fosi/etc release a MiniDSP Flex competitor. Would love to see this level of DAC performance alongside modern true two way crossovers and PEQ for room correction, as well as some analog inputs

I would like to see an SOA DAC with HDMI Arc input.

 

[yanm]

What I would love is that all the above create the same quality filters as RME does for example.
None of them is quite there yet.

To be honest, this test had made me rethink: I am tempted to buy … a RME ADI-2 DAC Fs instead of my Topping E50. It does everything I need / want (PEQ, loudness, IR control, settable output range…) and it does it extremely well.

The topping D50 III? Amazing device for the asked price but also some limitations (PEQ only on USB, same PEQ parameters for L/R channels, PEQ implementation less optimal than RME, no IR control), which makes it ill-suited for my use case.

 

[Rja4000]

I am tempted to buy … a RME ADI-2 Fs

I suppose you mean the ADI-2 DAC fs ?
You can hardly go wrong with this one, IMO.
I'm listening to some music through the ADI-2 Pro fs R as I write this.

 

[yanm]

I suppose you mean the ADI-2 DAC fs ?
You can hardly go wrong with this one, IMO.
I'm listening to some music through the ADI-2 Pro fs R as I write this.

Yeah, indeed the DAC version (corrected in the above post). The pro is 50% more expensive and I don’t need an analog input. The thing that is kind of holding me back was that there were rumours of an updated design… even though I don’t remember where I read that

 

[JamesYeomans]

I was told that there will be no update to D50III that will allow SPIDF input to have this EQ feature. EQ on USB is pretty useless to me. Why would someone prefer indevice EQ than EQAPO for example, especially when indevice EQ still affect performance a bit.

Fortunately, I was told by Topping that D900 or whatever they would be releasing will have EQ feature on ALL inputs. It's gonna be a stereo device though. They will probably release an 8 channel version if that one sells well.
The lesson I learned from measuring Minidsp Htx is that multichannel device with complicated mappings and EQ is not as simple as I originally thought.

People wanting to use a device like this with iOS/iPadOS devices won’t be able to use something like EQAPO. So in device PEQ useful to them.

 

[Eldus]

Type C port "Power"
You can use any mobile charger.

Personally, I don't like the screen on this DAC. don't like displays of this type; it can fade over time. It would be better if they used a display like the one on the D10. hope there's a function in the settings to turn off it.

I love the display on the E50.

 

[Peib]

thanks for sharing

 

[KaeliKoo]

Thanks for the review! - Question: When connected to BT, will the remote control process (next track / prev track / play / pause) --> Thanks!

From James Steele @ Hi-Fi College --> D50 III cannot control BT pause/play/prev/next functions with remote.

 

[danadam]

I would like to see a random test of this parameter alone on a random sampling of DACs to see how much variation there is between older/newer, cheaper / costlier, this chip / that chip etc

There are some here:

Let's develop an ASR inter-sample test procedure for DACs!

By way of contrast, I'll mention that this problem isn't nearly as bad in video or film. Most of the shows I watch, whether on streaming or disc, have good dynamic range and low overall levels. This applies across a broad spectrum of studios / productions. It appears that in video & film...

www.audiosciencereview.com

 

[Rja4000]

I renamed this thread as "unofficial", since Amir released his own review

Overall, I'm quite happy with the comparison, since results mostly match.

I'm very interested at the differences in some measurements: SINAD, dynamic range and, to some extend, THD+N vs frequency.

I now have my APx Flex trial extended some.

I've been working on SINAD: I re-measured, but still get around 124dB at 4V. Same result with APx Flex and Virtins Multi Instrument now.

Dynamic range, I suspect my CCIR-2k weighting is lacking somehow.
I have an idea of a trick to make it work with APx Flex. I'll give it a try.
I'm also not sure if Amir actually uses CCIR-2k weighting

For THD+N vs frequency, I have lower noise at mid-low frequencies but higher distortion at high frequencies. That's a general trend for this measurement for me.
I need to check why.
May be related to the RME ADC ? Or another difference in the test.
Unfortunately, I can't use APx Flex software for this measurement, since it requires different sampling frequencies for DAC and ADC.

More to come...

 

[srkbear]

Thank you

Yes, of course
An analog 10 bands parameric stereo EQ alone would cost you much more than this little box, I'm afraid. And it has a lot of disadvantages.

One of them being you'd have no way to store and recall presets (although, in the Pro world, the new Sepia platform is supposed to allow it. But that's a completely different topic).

There are very little companies still doing analog EQs nowadays (outside of basic bass/trebble).
All big names of Pro EQs in the past switched to digital a long time ago.

I really don't see why.
Look at the RME Multitone, which shows noise and distortion (or absence of them) for a quite extreme correction.
That's what you get from a properly implemented digital EQ.

I’ve had great success with the analog bass boost on my amp. It has no impact on the overall gain and I’ve experienced no clipping at all. I use Roon’s DSP PEQ when I need to, but the biggest issue I have with it is that no matter what adjustments I make, if I don’t decrease the global gain I get clipping every time. And when I do decrease the global gain there seems to be some degradation of the overall presence of the music—the resulting sound always sounds a bit disembodied or low-energy compared to the native sound without the DSP in the signal chain.

I know that sounds vague, but there seems to be something about the PEQ engine in Roon where even if you make the slightest adjustment in a single band, you have to bring down the global gain quite a bit to stop the clipping light from flashing under the headroom management section of the software. Perhaps the headroom management clipping indicator is too sensitive itself, and may not be accurately reporting actual clipping that may be going on, I don’t know (it’s hard for me to imagine how it knows what’s going on at my amplifier stage). Perhaps others can comment on this or correct my perception/approach?

This is why I’ve been so pleased to have found my Hifiman HE1000 Stealths to use in conjunction with the ASP xBass function on my iFi Pro iCAN Signature amp—the tuning is perfect for my tastes no matter what music I throw at it, and I can eschew messing with PEQ altogether.

I am somewhat surprised that Topping chose to introduce this function in one of their low to mid-tier offerings. I suspect that they may be planning on retrofitting it into some of their earlier products via a firmware update; we’ll have to see.

 

[Rja4000]

Overall, I'm quite happy with the comparison, since results mostly match.

I'm very interested at the differences in some measurements: SINAD, dynamic range and, to some extend, THD+N vs frequency.

I've been working on SINAD: I re-measured, but still get around 124dB at 4V. Same result with APx Flex and Virtins Multi Instrument now.

SINAD values I get seem coherent with the dynamic range and distortion values I'm getting.

i think that the difference just comes from the fact that my setup, with the E1DA APU notch and the RME ADI-2/4, has less noise than the APx555 Amir is using. And the SINAD is mostly limited by noise.
(That may seem shoking, but the E1DA is working only at a fixed frequency, while the AP has a frequency variable notch. A completely different challenge. And the APx555 is 10 years old now.)

is that important ?

It's important that I can explain why I get different figures.

As for audibility, it's Not important at all.

Anything around 120dB SINAD is state of the art.
Anything above 110dB SINAD is excellent, as per Amir's ranking, and should be considered transparent.

Dynamic range, I suspect my CCIR-2k weighting is lacking somehow.

It was.
I found that my figures were using CCIR-1k instead, which is giving 5.6dB lower figures systematically.
So I corrected them.
Now, my figures are slightly better thsn Amir's.
The reason is probably similar to SINAD difference.
And it doesn't matter either.

For THD+N vs frequency, I have lower noise at mid-low frequencies but higher distortion at high frequencies. That's a general trend for this measurement for me.
I need to check why.

That one, I still don't understand.
Still investigating.

 

[nanook]

Why do you think it's technically "impossible"?
I also assume the PEQ is done in the XMOS, which supposedly has significant computing power and which supposedly is basically idle otherwise.
It would however require to push the stream up and down - no idea if the XMOS can support this.

I'd also like to see PEQ for all inputs - my speakers and my room have the same flaws no matter which input of the DAC is used....
For USB I can use APO Eq.
Anyway it's a huge step forward !

As said here, we doubt it's technically possible to enable EQ for all sources for this device.

 

[CedarX]

Why do you think it's technically "impossible"?

I don’t believe we know all the implementation details, how the different inputs are routed internally (through the XMOS 316 or not), and whether the XMOS EQ libraries allow “global” PEQ for all the inputs.

Processing power may not be the issue here, although the PEQ error noted by @Rja4000 (100Hz filter) may be caused by some rounding issue… which could be related to some XMOS library/computing limitations. Again, I don’t think we know that for sure.

See @fabriceo comments about SPDIF input here:
https://audiosciencereview.com/foru...alanced-dac-with-eq-review.53856/post-1976833

In any case, *if* the SPDIF (and Bluetooth) inputs are not routed through the XMOS chip then PEQ on all inputs is indeed impossible.

FiiO has some dongles that also use the XMOS 316 and have BT. They have what appears to be a “global” PEQ, but it is in fact two separate PEQ: one in the XMOS for USB input, and another one in the QCCxxx for BT input. Two separate firmware, one updated through USB, the other updated through BT. They may be dumb… or there are some good technical reasons as why they do that.

 

[nanook]

See @fabriceo comments about SPDIF input here:
https://audiosciencereview.com/foru...alanced-dac-with-eq-review.53856/post-1976833

Thanks, so eventually it might indeed be possible to route the SPDIF through the XMOS, but for good reasons developers hesitate to try to get this going.
In case XMOS would develop an example application that people can copy things may be different. This clearly would be added value for the XMOS.