Denafrips Pontus DAC variable latency issue

[Onurb]

Hello guys,


I am having some latency issues with my Denafrips Pontus DAC.

I measured the latency using the RME Fireface UC USB interface and the REW software (using loop-back as timing reference ).

I made 20 measurements, one after the other with only a short delay (few seconds or less) between the measurements. Here are the results (milliseconds):
1.01, 1.375, 1.557, 2.89, 3.068, 3.203, 0.661, 0.792, 0.927, 1.12, 1.245, 1.359, 1.49, 1.615, 1.734, 1.844, 1.969, 2.099, 2.219, 2.354

The average value is 1.73, the minimum value is 0.661 and the maximum value is 3.20 ms.

For checking my measurement procedure, I measured the latency of a miniSDP SHD. The 11 measurements are exactly the same: 24.646 ms. Thus no variation. These measures confirm that the latency variation I observed is due to the Pontus and not to my measurement procedure.


For most stereo system, having a variable delay is not an issue: a song might start a bit later or sooner, this will not be noticeable for music listening.

For my system, it is a big issue for the following two reasons:

1) I am using the DIRAC Live room calibration system (as implemented in the miniSDP SHD). DIRAC measures first the left speaker and thereafter the right speaker. It is likely that DIRAC send the measurement audio signal for the left and right speaker separately, thus the audio signal will come back to the microphone with a different random delay. Consequently, if the left and right speaker are at exactly the same distance from the microphone, DIRAC will interpret wrongly the delay difference as different distances.

2) I am using a digital crossover (miniSDP nanoDigit 8) for a DIY 4 way system. The Denafrips Pontus DAC is used for the high frequencies (above 1000 Hz). To avoid cancellation at the the crossover frequency, I first adjust the delays in the digital crossover such that the midrange and treble reach the microphone at exactly the same time. This is impossible to achieve with the Pontus. The frequency response of the speaker (midrange and treble playing together) changes for each repeated measurement because each different random delay for the treble induce a different cancellation around the crossover frequency. Furthermore, when performing room correction with DIRAC Live on this system, the left and right speaker have different phase which leads to a lower combined volume, comb filtering and a broken stereo image.


I would like to send the Pontus back for a replacement unit but I am afraid that this is an issue that affects all Denafrips Pontus.

Any suggestions?

Thanks,

Bruno

 

[Vini darko]

Intresting system build. This sounds like clocking inconsistency. Might be a good idea to contact denafrips and ask them if this type of inconsistency is expected or not. I suspect minor temperature changes are causing drift in clock accuracy.

 

[ElNino]

The Pontus uses a FIFO. Unfortunately variable latency like this can be typical with FIFO implementations.

Even without a FIFO, you really want a common clock for a digital crossover system like this to work.

 

[Onurb]

I contacted the DENAFRIPS support and received this clear explanation that confirms ElNino's answer:


"DENAFRIPS DAC uses adaptive FIFO buffer and reclocking to reduce the jitters of the upstream digital sources.
The jitter rejection is excellent but the downside is the variable latency of the FIFO implementation is inevitable."

 

[BWPhil]

Hi Bruno, I am interested in the Pontus DAC, actually the newer II version but this latency would apply to both. I do 2 channel digital room correction using Acourate and convolution in ROON. Can you please confirm if there is any latency difference between the left and right channel of the DAC? I would think that the latency would vary each time a track is played, but it should be the same for both left and right channel at that instance. I sent email to Denafrips with the same question, but he provided the same response he sent you, even though your application is very different than my.

 

[Deleted member 41613]

I've had a Pontus I, and didn't know how quickly I had to get rid of it. Usb was unstable weird kind of noise, starting out of blue during playing. I know 2 other people who have experienced the same with the Pontus II. The sound was also quite disappointing, they say it sounds smooth, in reality it sounds compressed and a bit thin. You can be very impressed with the online sales and marketing.

 

[Ivodam]

I am really surprised to read the statement above. I have had a Pontus II DAC for a year now and it sounds really great with my Simaudio Moon preamp and power amp and Tannoy Glenair 15 speakers. I have not noticed the latency issue either. The problem described seems to me to be some drivers and software players setup issue.

 

[Mnyb]

I am really surprised to read the statement above. I have had a Pontus II DAC for a year now and it sounds really great with my Simaudio Moon preamp and power amp and Tannoy Glenair 15 speakers. I have not noticed the latency issue either. The problem described seems to me to be some drivers and software players setup issue.

Latency is not a problem for normal listening, it’s like if you hesitate before pressing play . Makes no difference.

OP of this tread has different DAC’ s for midrange and treble . So he does have the problem that latency must be consistent aka the same all the time .

And as some pointed out ideally clock synced otherwise they may drift apart over time.

Using the DAC in normal hifi setup is probably uneventful and not an issue .

I don’t know if a DAC with ASRC is more consistent ?

Maybe mini-DSP shd or similar with their inbuilt DAC’s would behave better and consistent. A non working xover is much bigger problem than what’s DAC is used for the treble and mid ? Mini DSP SHD or similar would probably use the same internal clock ?

Note . I know that SHD migth not have enough channels, maybe som other product from them ?

 

[Decebel]

How well does the pontus 2 12th work with an optical input? Am I going to use it with deq2496?