DENAFRIPS 'ARES' R2R discrete ladder DAC - close up view

[GoldenSound]

You have presented this opinion in the other thread which I opposed, as the beginning phase of charging capacitor is a straight line. Further tests prove your point, also the other argumentation is correct.

Is Holo Audio response rounded? - Not true. In JA tests it was showing a small ringing that disappeared when testing with 2.5GHz ADC sampling rate of the analyzer. I was initially blaimining digital filtering, but your further test revealed a true reason for such ringing. It is because there is a random ultrasonic scrambling noise added to the music before conversion. Energy of a scrambling noise that is close to the Nyquist frequency cause ringing. This is a truth you have proven yourself, still not admitting...

Digital filtering in Holo Audio in addition to scrambling is still there, JA was pointing to the rolloff figures, giving numbers what is expected from NOS and what is found there.

For the above reason Holo Audio DAC is also (like Denafrips) not NOS, for a different reason.

We've discussed this point before and I've shown practical demonstrations as well.
I'm not going to go round in circles with you for the 100th time.

For those wondering.
The ringing in JA's test was because he used a 192khz ADC sample rate.
Use a higher sample rate to fully capture all content in the impulse response and it is not present.

There is also no 'scrambling noise' and I'm not entirely sure what this would be referring to.

 

[sajunky]

We've discussed this point before and I've shown practical demonstrations as well.
I'm not going to go round in circles with you for the 100th time.

For those wondering.
The ringing in JA's test was because he used a 192khz ADC sample rate.
Use a higher sample rate to fully capture all content in the impulse response and it is not present.

There is also no 'scrambling noise' and I'm not entirely sure what this would be referring to.

Scrambling ultrasonic noise close to the Nyquist frequency of the analyzer's ADC (~100kHz) cause this ringing, it is a fact. A classic case of ringing due to the limited bandwith, but you tried to blame a limited bandwith test pattern. This is not true, as other NOS DACs like from Audio Note do not show ringing when JA tested them using the same test sample and the analyzer.

 

[GoldenSound]

Scrambling ultrasonic noise close to the Nyquist frequency of the analyzer's ADC (~100kHz) cause this ringing, it is a fact. A classic case of ringing due to the limited bandwith, but you tried to blame a limited bandwith test pattern. This is not true, as other NOS DACs like from Audio Note do not show ringing when JA tested them using the same test sample and the analyzer.

It is not a fact.

John Atkinson himself literally said it was because of the ADC sample rate....

As said before I'm not going to go round in circles with you when you're just repeating the same unfounded claims with 0 evidence. So I'll just repost what I posted last time you said this ( https://www.head-fi.org/threads/an-...st-and-holo-may-hqplayer.952934/post-16605353 )

Also if there was added noise/'scrambling' around 100khz then you'd clearly be able to see it when looking at a bandwidth above 100khz. And yet we do not.....

JA test is linked on the top.

And as another user already pointed out, JA in that very article says:

"Ignore the very small amount of symmetrical ringing before and after the single full-scale sample, which is due to the antialiasing filter of Audio Precision's A/D converter operating at a sample rate of 200kHz"

You then claimed that JA was wrong:

JA has skills and years of experience, I can't believe that he made such a wrong assessment.

This is not true at all.
You cannot capture a square wave, true NOS impulse response, or other signal that has content outside the nyquist frequency. It is mathematically impossible.

When we generate as perfect a square wave as we can, it looks like this, that is assuming you have sufficient bandwidth, in the case of this screenshot, the ADC is running at 2.5mhz.

Now if we change nothing other than the ADC sample rate, to the 192khz that JA was running:


Now we can see that the square wave is not perfect. Not because it isn't perfect in reality, but because we are not sampling at a high enough bandwidth to capture it. If we look in further we can see the ringing artifacts more clearly.


If we reduce our ADC sample rate to 44.1khz then because of our further reduced bandwidth limitation, this effect gets stronger:

We can even use a minimum phase filter on the ADC and the type of ringing changes:


So as John Atkinson said in that article. The ringing shown on that IR is due to the ADC sampling rate he was using. There is no such thing as infinite bandwidth, and no amount of production, mastering etc will change the basic principles (nyquist theory) upon which digital audio works.

If you have some actual evidence for your claim that Holo is using DSP (or any of the other claims you've made about various products then refused to back up), please share it.
Otherwise please stop making unsubstantiated claims about brands you don't like, and picking/choosing information to craft the story you want to tell. You can't use JA's graph as 'proof' but then say that his explanation of said graph is completely wrong....

 

[sajunky]

This is true.

John Atkinson himself literally said it was because of the ADC sample rate....

This is also true.

So as John Atkinson said in that article. The ringing shown on that IR is due to the ADC sampling rate he was using.

However I made a point twice in the row (you can't read) that 1kHz test sample do not cause the same limited bandwith ringing when testing Audio Note DACs which we know there are true NOS. Using the same limited bandwith analyzer and a test sample. Find something else, you do not answer to the arguments.

 

[alphaman]

GoldenSound:
Your video on the Ares 2 (and many of the graphs you show, which I assume are your own measurements) are confusing.
What's this (you claim it is an "impulse response"???? wha??????????????? never seen anything like before ... it in textbooks, journal papers, etc):

In any case ...
The dirac impulse response that Stereophile has shown for all NOS dacs -- including all Denafrips -- are concordant with all prev. NOS dacs that Stereophile (and others) have measured. For example:
Ares II/Terminator (NOS)

 

[sajunky]

What's this (you claim it is an "impulse response"???? wha??????????????? never seen anything like before ... it in textbooks, journal papers, etc):

This is right, look at the sampling rate 44.1kHz points, a time scale is on the top. One sample is low, a pulse is only one sample long. A triangle you see looks like a 16 oversampling points (not 8 as claimed), but in principle it is a triangle created by a linear interpolation in software.

 

[alphaman]

I’m literally on da crapper right now hopin’ on squeezin’ out a nice clean ‘un. Savin’ up on da TP… but da bathroom library is well stocked. Uhhhhhhnnnnn…..

 

[GoldenSound]

GoldenSound:
Your video on the Ares 2 (and many of the graphs you show, which I assume are your own measurements) are confusing.
What's this (you claim it is an "impulse response"???? wha??????????????? never seen anything like before ... it in textbooks, journal papers, etc):

An impulse response is an intentionally illegal PCM signal designed to test the filter design of an oversampling filter.
A single PCM sample with silence either side represents all possible frequencies at full scale, and so we can see how the DAC (or external oversampling) filters out content above nyquist frequency. You can derive frequency response from an impulse response and vice versa.

The above image shows the impulse response test file.
The vast majority of dacs will have a sinc filter that aims to attenuate everything past 22.05khz, though there are of course variances in design.

A typical linear phase impulse response test output will look like this (and this will likely be more recognisable):

and a typical minimum phase output will look like this:

I'm not sure if Dirac is doing anything specifically different, but the idea with that is that an impulse response test is played, and the result in your room is recorded.
Based on the result Dirac can then correct for any frequency response or phase discrepancies caused by the speakers/room itself. This isn't specific to Dirac though and is typically called 'convolution' elsewhere.

The dirac impulse response that Stereophile has shown for all NOS dacs -- including all Denafrips -- are concordant with all prev. NOS dacs that Stereophile (and others) have measured. For example:
Ares II/Terminator (NOS)

In this case it's simply that he hasn't zoomed in very much. So you can't really see anything.
The image below was recorded at 768khz on an ADI-2 Pro ADC and is the one I'd posted before.

If we zoom out the timescale to about the same as what john atkinson showed then it looks like this:

JA's test isn't showing anything different, it's just not zoomed close enough to really see the shape of the IR

 

[GoldenSound]

However I made a point twice in the row (you can't read) that 1kHz test sample do not cause the same limited bandwith ringing when testing Audio Note DACs

1khz tests won't cause any ringing on any DAC because it isn't an illegal signal.

Ringing is an artefact caused by a filter trying to represent a signal which it does not have the bandwidth to represent. Such as a square wave.
If you feed only content under the nyquist frequency you will not have any ringing

 

[alphaman]

The image below was recorded at 768khz on an ADI-2 Pro ADC and is the one I'd posted before.


If we zoom out the timescale to about the same as what john atkinson showed then it looks like this:


JA's test isn't showing anything different, it's just not zoomed close enough to really see the shape of the IR

Oh, uh .... okay toots
Another question fer ya toots .... characterize the architecture of a classic Philips DAC.

 

[GoldenSound]

Oh, uh .... okay toots
Another question fer ya toots .... characterize the architecture of a classic Philips DAC.

I'm not sure how this relates to the discussion?

 

[sajunky]

1khz tests won't cause any ringing on any DAC because it isn't an illegal signal.

Ringing is an artefact caused by a filter trying to represent a signal which it does not have the bandwidth to represent. Such as a square wave.
If you feed only content under the nyquist frequency you will not have any ringing

How this relates to the text you are quoting?

Asking because you was just trying to teach us about square 1kHz test sample response, not a sine waveform. A pulse sample as used by JA brings similar results. No ringing on Audio Note DACs, but ringing on the Holo Audio DACs. Answer this question why it happen.

 

[rudy49]

The only problem with the ares2 is that if you start to listen to music you don't stop even though you planned doing something else.

i agree, it's very engaging. not something for easy listening.

 

[alphaman]

How this relates to the text you are quoting?

Asking because you was just trying to teach us about square 1kHz test sample response, not a sine waveform. A pulse sample as used by JA brings similar results. No ringing on Audio Note DACs, but ringing on the Holo Audio DACs. Answer this question why it happen.

SAJUNKY: You make valid points. I suggest you post GoldenSound's video and links to his forum output (Audiophilestyle, THIS thread on head-fi, etc; see Refs below) on diyaudio's forum. Or, I'd encourage GoldenSound himself to post his claims directly. I would do it myself but I have been banned from that place for being overly critical of their moderation.

In any case, DIYAUDIO is the best place to set the record straight on the Denafrips R2R technical issues because:

- they get a lot of daily traffic (from people with actual experience -- both subjective and objective)
- they have a membership from some of Philips' (Holland) orig. design team
- the long Soekris thread and experience from experiments with the Soekris project (which similar to Denafrips design)
- lots of DIYers with non-oversampling knowlwdge
-etc.

You may want to include the following references:

https://audiophilestyle.com/forums/topic/62872-denafrips-dacs-might-not-actually-be-nos/
https://www.audiosciencereview.com/forum/index.php?threads/denafrips-ares-ii-not-nos.25104/

https://goldensound.audio/2021/07/13/denafrips-ares-2-measurements-and-review/

 

[314Rye]

It would be interesting to know how this information that Ares 2 may be a not real NOS dac has effected to the listening experiences related to Ares 2. Has the music started to cause less invigorating listening experiences among Ares listeners after these shocking news?