This is common effect with many DACs when using DoP. USB interface should delay data stream enough that unmute happens about 100 ms after DSD data has been detected, and mute 100 ms before DSD data ends.
Delaying unmute at start is easier, but really the only way to do proper muting at stop is to have enough buffer/delay in the USB interface so that it can see "ahead of the time".
This is just one of the many down sides of DoP.
Any suggestions for Mac users?
I am not aware of any way to use other USB drivers to avoid DoP.
I am not aware of any way to use other USB drivers to avoid DoP.
Is there a way to do a gapless playback from software side to make sure the data stream wont be cut off?
I noticed on the Amanero site that it is possible to reflash their card to add a 500 ms delay to the the start of the i2S stream. Would this do the trick?
Yes, it is gapless, but when the player enters stop mode then the stream will cut...
(unless you have configuration where output sampling rate would need to change, rate cannot be changed on the fly)
Not really, there should be delay before Amanero unmutes the output after the data stream has started. And vice versa it should mute the output before data stream stops. I understood that the stop case is particularly problematic?
Yes, the stop click is louder, but only marginally. It is a shame because the combination of HQ Player Desktop and DSC1 is a real winner. The best sound I have experienced in my systems, both Stax headphone based and the large room system.
What I am now curious about is why the Resonessence Labs Herus and Concero which I also have, do not make noises doing DSD under DOP in the same systems. Neither does a friend's M2Tech Young DSD.
Is it an Amanero shortcoming with any hope of overcoming for Mac users?
N.
Hi Nic.M,
In my system - which is PC based - the Amanero also when shifting tracks or on stop & start of playback. I somehow (although not entirely satisfactorily) solved it by setting up the clock for asynchronous playback - this reduced the clicks substantially but also caused playback to not really work at 384 kHz. And then I haven't had the time to amend this ...
BTW - you are aware of the Amanero "support" thread:
http://www.diyaudio.com/forums/vendors-bazaar/216474-usb-i2s-384khz-dsd-converter.html
Although very long there's a good bit of information here about various challenges with the Amanero board ...
Cheers,
Jesper
In my system - which is PC based - the Amanero also when shifting tracks or on stop & start of playback. I somehow (although not entirely satisfactorily) solved it by setting up the clock for asynchronous playback - this reduced the clicks substantially but also caused playback to not really work at 384 kHz. And then I haven't had the time to amend this ...
BTW - you are aware of the Amanero "support" thread:
http://www.diyaudio.com/forums/vendors-bazaar/216474-usb-i2s-384khz-dsd-converter.html
Although very long there's a good bit of information here about various challenges with the Amanero board ...
Cheers,
Jesper
It just depends on the USB implementation controlling output mute... For example iFi iDSD Micro also has some clicks and pops and Mytek also. While Fostex, TEAC and exaSound are quiet.
For DoP it requires a look-ahead buffer in the USB receiver, so it can control mute before changes in the output data happen.
Sabre is especially nasty chip in this respect because it cannot be programmatically switched between PCM and DSD modes but instead uses autodetect based on the input signals (so clever mute is mandatory).
You could ask Amanero about the mute control and what possibilities there would be for having longer unmute/mute time delta in respect to the output data.
Thank you Jasper for sharing your experience and for the asynchronous clock tip. How do you set up asynchronous clocking with the DSC1 and HQ Player?
Excuse my ignorance, but I do not know how to do this. I have gone into the settings and see no dialog for doing this in HQ Player.
Yes, I have been slowly wading through the long Amanero thread, tedious reading(I am on page 93 of 150 or so.) There was only one post so far of
clicking noises but nobody seems to have paid much attention to it. I will struggle on with it in the hope of finding something helpful.
N.
I will do that and hope for a response.
In the meantime, I could always install Windows on my music dedicated Mac mini. (SSD, outboard power supply, custom fan control, separate USB voltage, etc., etc.)
I like the sound of your discrete DAC and the resampling to DSD enough to take a shot at it, if I can get those nasty noises under control.
Nic
@Nic.M:
Yes, I also find it to be a quite long thread and I haven't read it myself in its entirety ... A suggestion could be that you post a question about your issues in the thread ...
I've read through my notes on the amanero thread and I haven't noted anything about solving the click issue although I remember reading something about it somewhere in the thread.
About the asynchronous setup it's done in hardware and to my knowledge not in the software. Acko has made a reclocker for the Amanero (link here: http://www.diyaudio.com/forums/group-buys/227502-amanero-isolator-reclocker-gb.html) and via this reclocker it's possible to lead the clock signal back to the amanero board and use it here instead of the onboard mounted oscillators (it can be setup in the amanero software).
This - which TMK is synchronous playback - may reduce jitter but in my experience introduces clicks and playback stops. Yet to my ears it's somewhat better sounding than asynchronous playback ...
And the sound with the reclocker (again to my ears) I would say is quite beyond the sound without the reclocker.
Asynchronous playback is where the amanero uses its onboard oscillators and outputs the data & clock signals to the Acko board timed by these onboard oscillators, however, the acko board's flip-flops are clocked using the clock coming from the acko board's oscillators.
In my understanding unless both the Amanero & the Acko board's oscillators have entirely identical timings (which over time is unlikely) the timing distances between the two clocks may vary - potentially also introducing slips in the clock/data stream. I don't notice this, though - except possibly at 384 kHz.
I hope this may be an accessible explanation - & good luck in finding a solution 😉
Cheers,
Jesper
Yes, I also find it to be a quite long thread and I haven't read it myself in its entirety ... A suggestion could be that you post a question about your issues in the thread ...
I've read through my notes on the amanero thread and I haven't noted anything about solving the click issue although I remember reading something about it somewhere in the thread.
About the asynchronous setup it's done in hardware and to my knowledge not in the software. Acko has made a reclocker for the Amanero (link here: http://www.diyaudio.com/forums/group-buys/227502-amanero-isolator-reclocker-gb.html) and via this reclocker it's possible to lead the clock signal back to the amanero board and use it here instead of the onboard mounted oscillators (it can be setup in the amanero software).
This - which TMK is synchronous playback - may reduce jitter but in my experience introduces clicks and playback stops. Yet to my ears it's somewhat better sounding than asynchronous playback ...
And the sound with the reclocker (again to my ears) I would say is quite beyond the sound without the reclocker.
Asynchronous playback is where the amanero uses its onboard oscillators and outputs the data & clock signals to the Acko board timed by these onboard oscillators, however, the acko board's flip-flops are clocked using the clock coming from the acko board's oscillators.
In my understanding unless both the Amanero & the Acko board's oscillators have entirely identical timings (which over time is unlikely) the timing distances between the two clocks may vary - potentially also introducing slips in the clock/data stream. I don't notice this, though - except possibly at 384 kHz.
I hope this may be an accessible explanation - & good luck in finding a solution 😉
Cheers,
Jesper
I am happy to report that I finally got noise free playback with the DSC1.
This was using my Mac Book Pro which runs Windows 7 with Boot Camp.
With the invaluable help of my computer whiz friend Oscar, and after loading the Amanero drivers, Asio and various tools, we set up JRiver to play back in non-DoP mode and converted everything to 2 times DSD.
This works with very few glitches and minimal switching and stop-start noise.
I have not tried using HQ Player, because I have not installed it on the laptop, yet. I miss its flexibility on how the conversion/upsampling is done.
Will set it up soon.
Thanks for the help and advice, Gentlevoice and Signalyst especially to Signalyst for designing such a great piece of gear and making it open to DIYers.
Nic.
This was using my Mac Book Pro which runs Windows 7 with Boot Camp.
With the invaluable help of my computer whiz friend Oscar, and after loading the Amanero drivers, Asio and various tools, we set up JRiver to play back in non-DoP mode and converted everything to 2 times DSD.
This works with very few glitches and minimal switching and stop-start noise.
I have not tried using HQ Player, because I have not installed it on the laptop, yet. I miss its flexibility on how the conversion/upsampling is done.
Will set it up soon.
Thanks for the help and advice, Gentlevoice and Signalyst especially to Signalyst for designing such a great piece of gear and making it open to DIYers.
Nic.
Here's some recent changes I've made to the DAC.
First, resistor network boards using 0.05% SMD resistors:
Second, use diyinhk isolator board and powered by local 3.3v.
I also got a QA400 that I've been playing around with. Here are some result:
Unmodified:
Stock DSC1 + TPS 5v power supply:
Stock DSC1 + TPS 5v power supply + 0.05% resistors:
Stock DSC1 + TPS 5v power supply + 0.05% resistors + isolator:
First, resistor network boards using 0.05% SMD resistors:
Second, use diyinhk isolator board and powered by local 3.3v.
I also got a QA400 that I've been playing around with. Here are some result:
Unmodified:
Stock DSC1 + TPS 5v power supply:
Stock DSC1 + TPS 5v power supply + 0.05% resistors:
Stock DSC1 + TPS 5v power supply + 0.05% resistors + isolator:
Moreover, I'm playing with a i/v stage inspired by this thread: http://www.diyaudio.com/forums/digital-source/227677-using-ad844-i-v.html
Plus a pull down resistor, so the output from the resistor networks wont be 0-5v with 2.5v DC.
I'd like to see by doing that, if I could remove the capacitor after i/v.
Plus a pull down resistor, so the output from the resistor networks wont be 0-5v with 2.5v DC.
I'd like to see by doing that, if I could remove the capacitor after i/v.
Hi mcluxun,
Interesting changes you've made ... your QA400 measurements makes me wonder, though, if you might have some kind of "systematic difference" somewhere since Jussi's own measurements of the DSC-1 are somewhat better (1 kHz sine wave fig 5 of 6; DSD128):
Preliminary measurement results for the DSC1 DAC - Blogs - Computer Audiophile
Just FYI.
Cheers,
Jesper
Interesting changes you've made ... your QA400 measurements makes me wonder, though, if you might have some kind of "systematic difference" somewhere since Jussi's own measurements of the DSC-1 are somewhat better (1 kHz sine wave fig 5 of 6; DSD128):
Preliminary measurement results for the DSC1 DAC - Blogs - Computer Audiophile
Just FYI.
Cheers,
Jesper
... just came to think of a question possibly related to this sentence of yours:
Have you tried to measure what the DC level is just after the 15k resistors (i.e. before the first AD844) when not playing music? And if so can you say if it is stable DC-wise (maybe at ½ * VCC)?
Cheers,
Jesper
Have you tried to measure what the DC level is just after the 15k resistors (i.e. before the first AD844) when not playing music? And if so can you say if it is stable DC-wise (maybe at ½ * VCC)?
Cheers,
Jesper
For one QA400 measures 20-20k. Jussi's measurements go all the way up to 70k.
Also if you look closer to the predominant harmonic: 2kHz, our measures are pretty close around -80db.
Digital filters also affect the result.
QAAnalyzer provides lots of ways to sample and present the data, so it's better to compare apples to apples with things all measured by QA400.
Hi,
Thanks once more for your feedback ...
I recognize your comments so just briefly to clarify: My reference was only to your fig. 1 and Jussi's fig. 5 where the difference may be from 2H at 73 dB (your fig. 1) and Jussi's ~78 dBs. I also noticed there were more noise around the 1 kHz & 2 kHz 1H & 2H in your figure as well as an overall higher noise level.
Anyway, just an FYI comment of mine.
Have a good day 🙂
Jesper
Thanks once more for your feedback ...
I recognize your comments so just briefly to clarify: My reference was only to your fig. 1 and Jussi's fig. 5 where the difference may be from 2H at 73 dB (your fig. 1) and Jussi's ~78 dBs. I also noticed there were more noise around the 1 kHz & 2 kHz 1H & 2H in your figure as well as an overall higher noise level.
Anyway, just an FYI comment of mine.
Have a good day 🙂
Jesper