• 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!

Linear Phase filter settings. Topping: How we (ASR) can help?

Guermantes

Senior Member
Joined
Feb 19, 2018
Messages
484
Likes
561
Location
Brisbane, Australia
The Topping D30 uses a CS4398 chip for DAC purposes. Its datasheet can be found here and shows two selectable filter settings: fast and slow rolloff. The appendix at the end shows frequency response measurements for each filter type. Page 11 indicates the fast filter has a stopband attenuation of 102 dB in single speed mode (48 kHz sample rate), 80 dB at 96 kHz and 90 dB at 192 kHz. Page 12 lists the equivalent specs for the slow rolloff filter as 64 dB, 70 dB and 75 dB respectively. Cirrus do not indicate whether either filter is linear or minimum phase, though I suspect the fast rolloff filter is linear phase. Perhaps someone else might like to shed light on this. The DSD module has it's own filter fixed at 50 kHz.

I have a device (Fiio X3 II) with this DAC chip that allows selection of the filter type. It seems the Topping does not have that option. The Fiio doesn't produce the aural artefacts you are describing with either filter setting.

I can't see anything in the datasheet that might allow Topping to spoof the measurements as alleged. Amir has done a teardown of the D30 with no sign of extra DSP: https://www.audiosciencereview.com/...ds/hardware-teardown-of-topping-d30-dac.2230/

I've done my own ad hoc tests on my D30 and compared it to a Topping D10, Apogee Rosetta 200, HifiBerry DAC+ Pro, HifimeDIY Sabre USB DAC, RME Fireface UFX and whatever DAC is inside the Genelec 8260A monitors (listening only). I used track 15 from this classical piano album as one of my 24 bit 96 kHz test tracks: https://www.deutschegrammophon.com/au/cat/4798308.

All of the DACs had their pros and cons. The D30 had more electrical hum than the pro gear but slightly less than the D10 and HifiBerry. None of the DACs exhibited anything like the artefacts on the piano recording you are describing, even when switching between linear and minimum phase filters on the HifiBerry (in fact, the HifiBerry measured relatively worse than the others for aliasing artefacts because of poor stopband attenuation). I felt the D30 was good enough that I pressed it into service as my main DAC in my music streaming rig at home. I listen to a lot of complex orchestral, choral and solo instrumental music (like this) and don't have any issues.

Perhaps your Topping D30 is performing below spec and producing more harmonic distortion than it should?
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
I have a device (Fiio X3 II) with this DAC chip that allows selection of the filter type. It seems the Topping does not have that option. The Fiio doesn't produce the aural artefacts you are describing with either filter setting.

It certainly doesn't. I have to admit mine is always on fast but did play with the setting.
 
Last edited:

Veri

Master Contributor
Joined
Feb 6, 2018
Messages
9,596
Likes
12,036
//ignored//
You're not seriously ignoring solderdude probably the most helpful and insightful dude we have on the forum right?
 
Last edited:

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
He ignored Krunok and answerred my questions/remarks. ;)
 

Eirikur

Senior Member
Joined
Jul 9, 2019
Messages
318
Likes
509
So, may I propose a relatively simple protocol?
  1. First capture some null baseline (no input)
  2. Play and capture a pure sine wave (create it with Audacity) for the piano note you will be playing
  3. Play and capture the piano note
Then start looking at the data, using DeltaWave, Audacity, spek, others?
Do you have another DAC to do the same tests and compare?

Possible piano samples:
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
Natural sound of piano is very complex, so I gather it would be difficult to analyze.

Actually you have a good filter, you just don't like it for some reason.

Natural sound of piano is very complex, so I gather it would be difficult to analyze. A simplified test pattern is probably better way to do that.

A knowledgeable person would not have much problem analyzing/comparing it.

Just submit the original file you used (see tips @Eirikur ) and a recording of the analog out of the D30 from that used file (not clipped)
Other owners could do this but won't be bothered I reckon, aside from that one has to make sure there isn't some weird issue with your D30 or gear.
 
OP
S

sajunky

Active Member
Joined
Jul 14, 2019
Messages
186
Likes
68
Location
South Africa
Ohhhhh. Couldn't see the post he quoted so made the wrong assumption, my bad.
Competition? You've got your recognition in the OP, so please don't be ashamed. Maybe you want to withdraw your comment as quoted in OP?
 

Veri

Master Contributor
Joined
Feb 6, 2018
Messages
9,596
Likes
12,036
Competition? You've got your recognition in the OP, so please don't be ashamed. Maybe you want to withdraw your comment as quoted in OP?
My point was a simple one. When a digital filter sounds broken, it is. Look for grace m900 measurents by Amir for example, bad filter messed with the measurents in a big way, of course it's then possible one might start to hear differences...
 
OP
S

sajunky

Active Member
Joined
Jul 14, 2019
Messages
186
Likes
68
Location
South Africa
So, may I propose a relatively simple protocol?
  1. First capture some null baseline (no input)
  2. Play and capture a pure sine wave (create it with Audacity) for the piano note you will be playing
  3. Play and capture the piano note
Then start looking at the data, using DeltaWave, Audacity, spek, others?
Do you have another DAC to do the same tests and compare?

Possible piano samples:
Not sure what to do after step 2&3. Slow sweep FFT? How to compare bits? If we forget for a moment about piano phenomenon (which is exclusively my observation), a single simple tone sounds very clean, but orchestral intensive parts or other harmonic complex sound like choir come out distorted. More complex harmonic in a sound, more distorsions. The other person (quoted in OP) describe it as a harshness and only with default Linear Phase filter, while @Veri says, with properly designed filter it shouldn't be a problem.

I am going now to make proposal for a lab test sample, will post soon. As for another DAC, I should have one before end of this month, I ordered another one in an act of desperacy. A half price of Topping D30 that comes with CapXon capaitors, BTW.
 

Veri

Master Contributor
Joined
Feb 6, 2018
Messages
9,596
Likes
12,036
I like that you at the very least, are really quite thorough in looking into this ;)
 

BDWoody

Chief Cat Herder
Moderator
Forum Donor
Joined
Jan 9, 2019
Messages
6,948
Likes
22,625
Location
Mid-Atlantic, USA. (Maryland)
I like that you at the very least, are really quite thorough in looking into this ;)

Indeed.

While he and i have had our moments in several threads, at least he seems to want to find a scientific explanation, rather than waving it off as immeasurable.

I can appreciate the zeal to possibly show something previously unnoticed, and he seems to appreciate finally that he is not dealing with a bunch of lightweights that he can hand wave away with casual banality...but I'd still like to see an actual blind test.
 

SIY

Grand Contributor
Technical Expert
Joined
Apr 6, 2018
Messages
10,383
Likes
24,749
Location
Alfred, NY
Indeed.

While he and i have had our moments in several threads, at least he seems to want to find a scientific explanation, rather than waving it off as immeasurable.

I can appreciate the zeal to possibly show something previously unnoticed, and he seems to appreciate finally that he is not dealing with a bunch of lightweights that he can hand wave away with casual banality...but I'd still like to see an actual blind test.

Yes, the most difficult problems to solve are the ones that don't exist. :cool:
 
OP
S

sajunky

Active Member
Joined
Jul 14, 2019
Messages
186
Likes
68
Location
South Africa
My point was a simple one. When a digital filter sounds broken, it is. Look for grace m900 measurents by Amir for example, bad filter messed with the measurents in a big way, of course it's then possible one might start to hear differences...
You responded in the same context, so I am afraid you can only confirm or withdraw. :)
This case is different to the grace. The filter is tuned to achieve the best measurements in tests with complete disregard to how it sounds.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
The filter is tuned to achieve the best measurements in tests with complete disregard to how it sounds.

You should add ... IMO... if you want to be right.

The filter is what is standard in the used DAC chip it is not designed by Topping nor selected solely for best measurements with disregard for sound. That is merely your opinion and not a fact.
 
OP
S

sajunky

Active Member
Joined
Jul 14, 2019
Messages
186
Likes
68
Location
South Africa
You should add ... IMO... if you want to be right.

The filter is what is standard in the used DAC chip it is not designed by Topping nor selected solely for best measurements with disregard for sound.
All you say that Topping cannot change filter parameters, that's interesting...
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
They DECIDED not to make the filters user select-able.
They could have if they wanted to.
Topping just 'hardwared' the chip so it uses the standard filter in the DAC chip. It is not their 'own' filter, tailored to their personal wishes or 'beliefs'.
Obviously they went for the best setting in the DAC chip (which arguably is the linear phase filter)
The fact that some prefer something else only says something about their preference.
 
OP
S

sajunky

Active Member
Joined
Jul 14, 2019
Messages
186
Likes
68
Location
South Africa
Topping just 'hardwared' the chip so it uses the standard filter in the DAC chip. It is not their 'own' filter, tailored to their personal wishes or 'beliefs'.
All you are saying ... again... that Topping cannot change filter parameters, that's interesting...
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
No, that's not what I said... so I will try to explain again

Topping just 'hardwared' the DAC chip so it uses the linear-phase filter present in the DAC chip.
It is not their 'own' filter, tailored to their personal wishes or 'beliefs' nor is the other filter that is available in the DAC chip user-selectable.

Below is the description of the used DAC chip. (some words are bolded by me)

The CS4398 is a complete stereo 24 bit/192 kHz digital-to-analog system. This D/A system includes digital de-emphasis, half dB step size volume control, ATAPIchannel mixing, selectable fast and slow digital interpolation filters followed by an oversampled multi-bit delta-sigma modulator that includes mismatch shaping technology that eliminates distortion due to capacitor mismatch. Following this stage is a multi-element switched capacitor stage and low pass filter with differential analog outputs.
The CS4398 also has a proprietary DSD processor that allows for volume control and 50 kHz on-chip filtering without an intermediate decimation stage. It also offers an optional path for direct DSD conversion by directly using the multi-element switched capacitor array. The CS4398 accepts PCM data at sample rates from32 kHz to 216 kHz, DSD audio data, has selectable digital filters, consumes little power, and delivers excellent sound quality.

That's the only thing a manufacturer that uses the chip can easily alter. For the D30 Topping decided to NOT make the filter option the DAC chip offers user selectable. They chose the best filter (linear phase).
 
Top Bottom