Review and Measurements of New Topping D50s DAC

[restorer-john]

It is some kind of spurious tone for which I don't have an explanation as to its source.

Your wife was cleaning her ring in the ultrasonic jewelry cleaner in the next room?

 

[Calexico]

My favorite measurement is the high bandwidth thd+noise vs frequency

 

[Calexico]

@amirm hello

could you please make thd + noise vs frequency in 96khz or 192 until nyquist frequency in your test standards? (higher than 20khz) to see how the dacs behave with high res audio ? I know the reconstruction filter in dacs keep frequencies higher when the samplerate is higher.
I'd like to see if there's ultrasonic stuff
Also i would be interesting to make the same test with dsd format to see if it's interesting to convert to dsd or if it generate more ultrasonics garbage.
It would be perfect to see measurements this way:
- standard cd qualify format measurement made in 16bit 44.1 khz
- high res format measurements made in 24bit 96 or 192 khz with bandwidth of all measurements of at least nyquist frequency
- dsd format measurements
@Blumlein 88 @solderdude @Veri
Wouln't it be a good idea?
It would also test the way the spdif receiver and the usb receiver react with those formats
It would help to choose the dac considering the format mostly used
Maybe some are more optimised for high res or dsd and some for cd

 

[phoenixdogfan]

Pity there's no balanced outputs.

 

[digititus]

Pity there's no balanced outputs.

D70

 

[Veri]

@Blumlein 88 @solderdude @Veri
Wouln't it be a good idea?
It would also test the way the spdif receiver and the usb receiver react with those formats
It would help to choose the dac considering the format mostly used
Maybe some are more optimised for high res or dsd and some for cd

I'm humbled to be included in that shortlist of wonderful persons
I honestly don't think any of these tests you're interested in are worth amir's valuable time, especially in terms of 'ultrasonic stuffs' or something being 'optimised for high res'. I'd be very very surprised if any of this is a high impact matter in audible realm, but rather measuring just to measure.

Amir has an intense backlog and keeps getting new requests, I'd say he best focuses on what matters.

 

[Calexico]

I'm humbled to be included in that shortlist of wonderful persons
I honestly don't think any of these tests you're interested in are worth amir's valuable time, especially in terms of 'ultrasonic stuffs' or something being 'optimised for high res'. I'd be very very surprised if any of this is a high impact matter in audible realm, but rather measuring just to measure.

Amir has an intense backlog and keeps getting new requests, I'd say he best focuses on what matters.

It matters as it can interract with amps.
Why for nos dac ultrasonic matters and for modern dac ultrasonic doesn't matter??
It's not logical.
I thought @amirm was a Retired engeeneer and that it's a hobby and the goal is to show failure in designs so that manufacturers can make it even better.
If not test show really the ultrasonic stuff that's a big misstake.
Filtering is the crucial part of the dac.
That's why there is oversampling also to put the ultrasonic garbage more in the high frequencies so that it can be easily filtered with analog filter.
It's important to test how the digital + analog filter of the dac perform and if it perform the same at different samplerates.
Maybe it's more important than the thd at 1khz measured with bandlimited

 

[Veri]

That's why there is oversampling also to put the ultrasonic garbage more in the high frequencies so that it can be easily filtered with analog filter.
It's important to test how the digital + analog filter of the dac perform and if it perform the same at different samplerates.

Well yes, and basically any DAC chip nowadays oversamples. Unless you yourself select something like super slow roll-off there's no problem.
I disagree that the filtering should be tested at all these different sample rates, for the reason mentioned above.

 

[Calexico]

@Veri
If there are analog and digital low pass filters in dacs that cut of ultrasonics it's not for nothing.

Filtering of the ultrasonics is the most critical part of dacs you cannot say it doesn't matter.

On d50s we saw on amirm thd vs freq measure that when you don't bandlimit the measurements sinad become 100db at 1khz and 80db at 30khz.

So just solder out the lpf of your dac if you think it has no meaning. You will see...
If the bandwidth of measurements is limited to hearable range then you totally mask the evaluation of the low pass filter of the dac.

Different designs of lpf will show no differences in measurements because bandwidth is limited to hearable range.
Then you think there is no hearable differences between low pass filter designs or between op amps.
That's normal because the measurments hide their effect. (Except the thd vs freq)

If you measure with a larger bandwidth you will evaluate the design of the analog and digital filter.

You will see differences in ultrasonics filtering between different designs.
Then you will evaluate in a more complete way that the dac is good engeneered.
As i said good filtering of ultrasonic garbage is critical for dacs.
That will explain why people hear differences between dacs where we see nothing on measurements because of the limited bandwidth of measurements.
The difference is because the ultrasonics garbage can stress other element in the amp or in speakers.

 

[daftcombo]

@Veri
If there are analog and digital low pass filters in dacs that cut of ultrasonics it's not for nothing.

Filtering of the ultrasonics is the most critical part of dacs you cannot say it doesn't matter.

On d50s we saw on amirm thd vs freq measure that when you don't bandlimit the measurements sinad become 100db at 1khz and 80db at 30khz.

So just solder out the lpf of your dac if you think it has no meaning. You will see...
If the bandwidth of measurements is limited to hearable range then you totally mask the evaluation of the low pass filter of the dac.

Different designs of lpf will show no differences in measurements because bandwidth is limited to hearable range.
Then you think there is no hearable differences between low pass filter designs or between op amps.
That's normal because the measurments hide their effect. (Except the thd vs freq)

If you measure with a larger bandwidth you will evaluate the design of the analog and digital filter.

You will see differences in ultrasonics filtering between different designs.
Then you will evaluate in a more complete way that the dac is good engeneered.
As i said good filtering of ultrasonic garbage is critical for dacs.
That will explain why people hear differences between dacs where we see nothing on measurements because of the limited bandwidth of measurements.
The difference is because the ultrasonics garbage can stress other element in the amp or in speakers.

To me, that makes sense. But are you really sure that would be audible? You need other bad devices down the chain, right?

 

[Tatteredmidnight]

@Calexico it seems this is a contentious topic that has sparked quite a conversation. Unfortunately it has dominated this review thread and created a lot of noise (pun) for people interested in specifics of the device mentioned. Would you consider moving this to a dedicated thread to keep this one about D50s issues?

 

[Calexico]

@Calexico it seems this is a contentious topic that has sparked quite a conversation. Unfortunately it has dominated this review thread and created a lot of noise (pun) for people interested in specifics of the device mentioned. Would you consider moving this to a dedicated thread to keep this one about D50s issues?

To me, that makes sense. But are you really sure that would be audible? You need other bad devices down the chain, right?

If it's inaudible then why everybody say nos is not good?

 

[Calexico]

@Calexico it seems this is a contentious topic that has sparked quite a conversation. Unfortunately it has dominated this review thread and created a lot of noise (pun) for people interested in specifics of the device mentioned. Would you consider moving this to a dedicated thread to keep this one about D50s issues?

I don't care. The most important is that amirm has heard my questions and i hope he will consider evaluating the filtering in dacs with different samplerates and high bandwidth for measurements.
Yes the filter graph shows it but it's confusing when you see the sinad so you don't see clearly if hf harmonics or distortion is filtered enough.

 

[gvl]

NOS is not good as the images are very high and close to Nyquist and their IMD products can fall back into audio band at audible levels, not to mention the HF droop. A -80dB 28kHz presence is a non issue being so low.

 

[Jim777]

Well yes, and basically any DAC chip nowadays oversamples. Unless you yourself select something like super slow roll-off there's no problem.
I disagree that the filtering should be tested at all these different sample rates, for the reason mentioned above.

The only reason I can think of for measuring a DAC at its highest sampling rate is that the impact of jitter on SNR is a function of the jitter fraction. If the jitter happens to be constant and you reduce the sampling period (i.e. max sample rate), you maximize the impact of jitter on SNR. (Gersho and Gray, Vector Quantization and Signal Compression, page 66).

 

[vkvedam]

No. I would upgrade for the remote though. It is so nice to change modes using the remote than messing with menus.

Cool, for the duties I have given it I would not need a remote ;-) I shall give it a pass for now. Many thanks!

 

[amirm]

We have a debate thread on measurements: https://www.audiosciencereview.com/...-debate-thread-about-audio-measurements.2580/

It is fine to ask me for some measurement in reviews but arguments about them all should go in the above thread.

And yes, as kindly noted, I am extremely behind in reviews. I usually tear down the gear after a day or two and just the work of reconnecting things and putting aside what is being tested is a lot. What makes our work here valuable is that we test what needs to be tested and not much more. There are other sites with jungle of graphs to confuse readers.

 

[amirm]

Also i would be interesting to make the same test with dsd format to see if it's interesting to convert to dsd or if it generate more ultrasonics garbage.

I like to test DSD. Alas, there are no test signals in DSD and my Audio analyzer can't handle them either. I have tried to convert PCM test signals to DSD but there are losses there and results not very reliable. It is a long overdue feature in Audio Analyzers given the popularity of DSD in DACs.

 

[Jimster480]

Awesome to see an improvement on an already great DAC!
leave it up to topping to just keep moving the bar higher.... a once unknown brand will slowly become a household name as long as they keep this up.

 

[digitalfrost]

No. I would upgrade for the remote though. It is so nice to change modes using the remote than messing with menus.

Anybody know if you can control multiple of these with a single remote?