Occasional glitches in the audio (very short, momentary dropouts)

With my latest setup and firmware and using very clean, high bandwidth HDMI cables, the number of dropouts has greatly reduced, but not completely. Perhaps once every 5-10 minutes, I hear a little glitch in the sound, a little audio dropout. This is often the symptom of an aggressive jitter-reduction PLL, so I would like to suggest that you offer some slightly more relaxed PLL settings for us to try. I know there may be more jitter possible with a more tolerant PLL, but the dropouts are more annoying than slightly higher jittter. Especially as many good DACs themselves have excellent jitter reduction.



Edited 1 time(s). Last edit at 01/11/2024 01:08PM by bobkatz.

Could the occasional glitches be the FIFO momentarily emptying? You don't explain in the manual how you keep the FIFO maintained?

Hi Bob,
May I ask what your DACs or active speakers are that you feed with the VanityPRO? It is a known phenomenon that ESS based DACs may produce some unwanted pops and dropouts if they have the DPLL set too aggressively. As you say, it is to do with jitter reduction, but on the DAC side. It is virtually impossible for the buffer in the VanityPRO to under or overrun due to the way the control algorithm works. What can happen is that tiny frequency changes which happen at buffer thresholds which is part of the frequency tracking algorithm can upset the resampler in the ESS DAC with too narrow DPLL bandwidth. This is briefly described in the manual in the Troubleshooting section, because most DAC manufacturers have the DPLL setting exposed so the user can change it to more relaxed level if they experience any issues. You can check this yourself if this is the case and if the dropouts are related to oscillator control events in the VanityPRO. Look at the System Status screen (you might need a magnifying glass, sorry) and you can check the Buffer utilization (in %) and as it approaches 20% and 80% thresholds the XO Control voltage changes once or twice by a tiny bit. if your dropouts happen around those moments, then it is likely that the DAC fed by the VanityPRO has very narrow clock recovery bandwidth (whether it is done by means of analogue PLL or digital asynchronous resampling).
Best Regards,
Pavel

Hi, Pavel. I am using a Grace M908 monitor controller. I'm not sure what kind of DAC is used in the Grace, I do think it is an ESS. I've written to Michael Grace to ask about the PLL situation and the nature of their DAC.

To check the status screen, do I have to be viewing the status screen at the exact moment that the dropout occurs?

Hi Bob,
Yes, you would need to be able to relate the dropouts to what's happening on the screen. I had a brief look at your monitor controller and noticed that there is a wordclock IO. If it is not used yet, you could try to connect the VanityPRO word clock output to the GraceM908 word clock input and synchronize both devices. If your dropouts are related to clock recovery in the VanityPRO, syncing the DAC could eliminate them.
Regards,
Pavel

Thanks, Pavel. I thought about the word clock. I'll give it a try!

Hi, Pavel. I tried the word clock. Grace M908 shows an error, but not exactly sure what error. The Grace appears to be locking regardless of the error. I got a clean playback for almost an hour and then 2 seconds of audio dropout. According to Michael Grace, the wordclock uses the same PLL, but skips some of the ESS FPGA so it might eliminate some questions as to the cause. But right now I can't say the exact cause though it is leaning in the direction of a Grace locking issue. Is it possible that when you switch between 20% and 80% buffer that you are changing frequency too quickly or too severely?


I'm going to scope the wordclock output against the AES preamble and see if I can see a clue.



According to the manual:

Flashing
Yellow
Indicated clocksource is
invalid or does not match
the samplerate of the
active input. Active input
clock is valid for digital
inputs or system clock is
valid for analog only
inputs.
m908 plays audio with
active input clock for
digital inputs or system
clocksource for analog
only inputs
For inputs asynchronous to
system clocksource,
configure input clock
override.
Fix invalid clock override
Configure system for a
clocksource that is synchrous
to input

Hi Bob,
A stupid question, you did turn the word clock output on in the menu of the VanityPRO, didn't you?
Can you please paste here or PM me the input specifications of the word clock signal from the Grace manual?
Thanks,
Pavel



Edited 1 time(s). Last edit at 01/14/2024 09:41PM by AP.

Hi, Pavel, yours was not a stupid question! I did not realize that the Vanity wordclock has to be turned on from the menu until I started measuring it with my scope, and discovering there was no output. Thus, the error from the Grace M908. So now I have to re-evaluate the Grace M908 with wordclock. Please stand by for that information!

In the meantime, as long as I was scoping, I connected the Vanity WC out AND ch 1/2 AES output to my Scope. I got these pictures:

[drive.google.com]
[drive.google.com]

As you can see, framing is off a bit in the Vanity. AES Preamble starts about 2 biphase bits before the wordclock transition. If framing is too far off, DACs which lock to wordclock will have trouble locking to the AES preamble, so I suggest you take a look at that, and if your FPGA is capable, retime the wordclock output until it is more in sync with the AES preamble.

In addition, wordclock polarity of the Vanity is the opposite of the Grace standard: "Grace M908: The word clock polarity is high for left channel data and low for right channel data.". I doubt this is an issue, since it is likely that the Grace obeys the AES input for channel identification. I'll watch the movie carefullly and see if the front left and right channels are refersed. Yes, in some rare cases, converters will reverse channels, that is, if they follow the wordclock for channel identification. In those cases, I suggest you put in a wordclock polarity menu item. Who knows what DACs you will encounter in the real world :-).

OK, now that I have the Grace M908 monitor controller hooked up and locked to wordclock from the Vanity, I get the same quantity of occasional audio glitches. I think the next step is to try a different model DAC and listen for an hour or two continuous and see if the problem is rectified.

Best wishes,


Bob

Hi Bob,
Thanks for the update. How is your power distribution consistency? [funny story] My friend used to live in a block of flats and he was getting dropouts due to unlocks of DAC digital input every time someone called elevator in the building. So he only listened to music late in the night when chances of someone needing an elevator was low. [/funny story] Power distribution and cabling can sometimes amplify the effect of mains glitches or instabilities. Anyway, please keep us posted.
Thank you.
Pavel

Power distribution in this room is incredible. I have a dedicated 100 amp (that's 100 amp) 240 volt service coming into Studio A. That goes into 5 Surge-X EMI/RFI filters and 5 20 amp 120 volt circuit breakers. All lines are star run back to this central box through steel conduit. I could go on, but let me finsh by saying that After I finished this renovation, the noise in the power line was reduced by over 20 dB in all modes: Line to neutral, line to ground, and neutral to ground. Even the acoustic noise from power transformers was reduced.

Prior to the installation of this dedicated line, I would play my Shuttle Launch recording that I made and the lights would dim! After the instalation, all power amp headroom has increased over 3 dB! So I can play peaks 3 dB louder in the room --- just by virtue of the new power installation.

So, yeah, my power is taken care of :-)

Dear Pavel: Some progress, but indications are not good.

I played a Blu Ray with a concert video with a 48 kHz 5.1 soundtrack. I took AES output channels 1/2 and fed that to a Prism Callia DAC. As you may know, the Prism DACs are almost unique in the industry, employing a pure, two-stage all-analog Phase Locked Loop. So Prism couldn't be any further apart from the ESS architecture than you can imagine. There were two dropouts in 15 minutes. I then powered the Blu Ray down, Buffer went to 0%. and back up, buffer went to I think 30%. I played for 15 minutes and there were no dropouts and I didn't have any patience, the buffer was pretty stable and hardly changing. I really don't think it was the buffer changing that caused the prism to drop out. I think it was something else. It would be nice if you could further investigate these occasional dropouts.

Are you using 110 ohm output impedance from the AES/EBU?

Best wishes,

Bob

Hi Bob,
Thanks for the additional info. The XLRs are correct output impedance and so are the coax outputs on other modules. If you are getting those dropouts on two different DACs and you think it is not buffer adjustment related, it sounds like there is something else going on. There are a few things we can try, but first would you be happy to record and send us back some diagnostic data from the unit?
Best Regards,
Pavel

Thanks, Pavel. Of course I'm happy to send diagnostic data. Anything you'd like. Well, the dropouts on the ESS DAC do seem to be buffer-related, or at least aggravated by sample frequency micro-adjustments associated with keeping the buffer from over or under flowing. But the other DAC, the Prism, not being ESS and having a pure 2-stage analog PLL, is usually not sensitive to slight errors in source sample rate accuracy. The Prism DAC seems to be very stable and dropout with other sources. And keep in mind that sample rate is 48 kHz for these, which is not a big challenge for any SPDIF or AES/EBU DAC

The hard part, of course is for you to debug at a distance. I have the tools to look at eye pattern, probably, if you show me how to connect them, if that will help you.

Hi Bob,
Thanks for offering your assistance to remotely debug the dropouts you are experiencing. I would suggest to start with logging a diagnostic serial dream from the unit via USB. For this you would need a diagnostic firmware (which has the serial stream enabled) and a serial terminal which can dump the stream into a file.
The firmware can be found here: FW_BUF3_UART.
And brief instructions how to connect to the unit and log the diagnostic stream:
UART Instructions.
The idea is that you are recording the internal diagnostic data while you are playing a film or music and if a dropout occurs you just note down the approximate time it happened since the start of logging. Then we can match that time instant to the diagnostic data around that moment and see if there are any irregularities. You can log a whole listening session, because a couple hours of diagnostic data is just a few hundred kbytes. Any questions, please ask.
Best Regards,
Pavel



Edited 1 time(s). Last edit at 01/27/2024 08:25PM by AP.

Thanks, Pavel. I'm traveling right now so when I get back. Would this diagnostic firmware change any of the desirable functions that are in the current release firmware?

Hi Bob,
No changes to functions and features. Actually there are some minor improvements to the frequency tracking and buffer management which will be later included in a new official fw release. Safe travelling!
Best Regards,
Pavel

Hi Pavel

Just want to add my 2 cents since I'm in a similar situation. I have two chains with a headphone system. When I'm using Mac Pro->VanityPRO->Weiss DAC204, the dropouts barely happened, maybe every 30mins or even longer, it's definitely listenable, but that was only improved by the latest firmware upgrade. My second chain is the recently added Eversolo A6 ME->Vanitypro->Weiss DAC204, it drops about every 5 mins, and did happen when the buffer hits 20%/80% threshold. And I'm using the same Fibbr 8K HDMI cable.

So maybe the dropout has sth. to do with the source as well? I can't find any way to adjust the DAC204 PLL BTW.

Cheers.

I have a further development. Vanity pro --- quite serious dropouts at 44.1 kHz. I tried playing an audio CD from the HDMI output of either of two different Blu Ray players and in both cases, Vanity just can't handle it, lots of dropouts. I conclude that your "PLL" is too "stiff".... intolerant of any incoming sampling frequency issues.

Hi seevic,
We are about to release a new firmware with further improved buffer control and frequency tracking, which should improve the playback even with ESS chips with fairly low DPLL settings. The DAC204 might need a firmware update which I believe Weiss will be happy to do. Please let them know and ask what the best approach would be. They know our products and managed to update their DACs before to work well (at that time for the Vanity203HD). If the dropouts are clearly related to the 20% and 80% buffer points then it is always the downstream DAC which has DPLL set too low.
Regards,
Pavel


seevic Wrote:
-------------------------------------------------------
> Hi Pavel
>
> Just want to add my 2 cents since I'm in a similar
> situation. I have two chains with a headphone
> system. When I'm using Mac Pro->VanityPRO->Weiss
> DAC204, the dropouts barely happened, maybe every
> 30mins or even longer, it's definitely listenable,
> but that was only improved by the latest firmware
> upgrade. My second chain is the recently added
> Eversolo A6 ME->Vanitypro->Weiss DAC204, it drops
> about every 5 mins, and did happen when the buffer
> hits 20%/80% threshold. And I'm using the same
> Fibbr 8K HDMI cable.
>
> So maybe the dropout has sth. to do with the
> source as well? I can't find any way to adjust the
> DAC204 PLL BTW.
>
> Cheers.

Hello Bob,

May I ask you to take a video of the VPRO status screen while you are getting those dropouts? With this data we could see more details and do the first analysis. Thanks.

Regards,
Milan

Hi, Milan. Thanks for your reply and your effort on the new firmware... last time i looked at the status screen, the occasional dropouts did not occur in sync with being near the 20% and 80% buffer points. I would like to try the new firmware and see how it does. I have to move the box upstairs and hook it up to my only remaining windows 10 machine so that's a pain, but if you think the new firmware will help these occasional dropouts I will try.

What do you think about my reactions to 44.1 kHz sample rate? It was totally unuseable, dropouts every 10-15 seconds.


Bob




Milan Wrote:
-------------------------------------------------------
> Hello Bob,
>
> May I ask you to take a video of the VPRO status
> screen while you are getting those dropouts? With
> this data we could see more details and do the
> first analysis. Thanks.
>
> Regards,
> Milan

Hello Bob,

quite werid behavior that only on 44.1kHz. Can you confirm the others FS are ok? Have you had that issue ever before when you bought your VPRO or it just started to happen recently? Anyhow you can make a short video of the status screen while the dropouts are present it would give us some feeling.

Thanks, and regards,
Milan