ProtoDAC TDA1387 X8 project

[eDad2003]

My experience switching from Polarity Inversion to CamillaDSP.

I went to experiment with CamillaDSP today. In Sections 5.2 and 5.3 of the build guide, they both refer to setting Signal Processing via "Menu, Configure, CamillaDSP". What I found is that Signal Processing was not editable in the CamillaDSP config page . Rather, I had to go to "Menu, Configure, Audio, Equalizers" and select "V2-ProtoDAC" there, then "Edit" to follow. Note: when I didn't find "V2-ProtoDAC" listed, I followed Tim's tip from #432.

I then saw this info in Audio Information

Code:

Audio Chain: MPD -> Polarity inversion -> hw -> Device
.
.
CamillaDSP: V2-ProtoDAC
Volume type: CamillaDSP (64-bit)

Polarity inversion is still in the chain, which I didn't think was possible since I thought it was mutually exclusive with CamillaDSP. I suspect I was double-inverting polarity with this setup (but not sure).

The solution was to disable Polarity Inversion before enabling CamillaDSP. Hopefully this helps the next person.

[eDad2003]

The volume increase from turning CDSP off and ALSA polarity inversion on was much greater than I expected. More  tests to come.

@DRONE7, @varunach, @Skip Pack

I notice it too, but really more at lower volumes IMO. Could it be as simple as CDSP uses a different volume curve from ALSA? For example, Spotify Connect offers Logarithmic, Cubic, and Linear as volume curve options.

[hifinet]

Correction: Since there are eight chips in the module, max voltage drop with the 24R is 1.92mV and with 430R it's 34.4mV.  It still doesn't really matter.

Ok, probably not, but I’ll try some other values. With an oscilloscope it could be possible to find the right value, but I don’t got one. So if anyone has one it would be interesting to see the results.

What difference would you expect to see with different values of I2S resistors?

[Skip Pack]

The volume increase from turning CDSP off and ALSA polarity inversion on was much greater than I expected. More  tests to come.

@DRONE7, @varunach, @Skip Pack

I notice it too, but really more at lower volumes IMO. Could it be as simple as CDSP uses a different volume curve from ALSA? For example, Spotify Connect offers Logarithmic, Cubic, and Linear as volume curve options.

I guess my concept is that the dynamic range should not change with volume -- if you adjust the volume (using software or hardware) from one player configuration so it matches the volume from another configuration (CDSP on versus CDSP off) they should sound substantially identical to our ears volumewize. there will be differences from the dithering and other effects. If there seems to be a 'contouring' effect that leads me to think of a loudness profile.

[DRONE7]

CDSP is a little obscured as to how and when and why so I am not surprised that volume and contouring may be involved

[OnkelPH]

Correction: Since there are eight chips in the module, max voltage drop with the 24R is 1.92mV and with 430R it's 34.4mV.  It still doesn't really matter.

Ok, probably not, but I’ll try some other values. With an oscilloscope it could be possible to find the right value, but I don’t got one. So if anyone has one it would be interesting to see the results.

What difference would you expect to see with different values of I2S resistors?

Good question. I don’t know. I’ve only seen some examples where a lower value is preferred. As some of you already mentioned I wouldn’t hear any difference….

[OnkelPH]

Which model WiFi adapter?

@Tim Curtis i bought this simple one since it is based on 5730 chipset that works out of the box. I gotta say its not the greatest since I only get 25 - 30 Mbps on it in my room but gets the job done. The Moode UI loadtime is longer and not snappy which I can live with. I think once the Moode Kernel gets to Linux 6.2 version there is a lot more wifi adapter options available because of compatibility with in-kernel drivers. 

https://www.amazon.com/dp/B0BHW6T96R?psc...ct_details 

I also ordered this dual band adapter from aliexpress because it was on the recommended list for plug n play. If this works better, I'll return the other one above.

https://www.aliexpress.us/item/3256805749323751.html?

Got the dual band from ali yesterday and put together the stack: rpi-Q7ll-ProtoDAC. It worked for a few minutes then the WiFi became unstable, dropping in and out and struggling to connect via Spotify (I see the player but it won’t connect).
Do you have the same experience @varunach

I found a WiFi to Ethernet adapter I had lying and that fixed the problem.
Finally I’m able to enjoy the potential of the ProtoDAC.

[eDad2003]

The volume increase from turning CDSP off and ALSA polarity inversion on was much greater than I expected. More  tests to come.

@DRONE7, @varunach, @Skip Pack

I notice it too, but really more at lower volumes IMO. Could it be as simple as CDSP uses a different volume curve from ALSA? For example, Spotify Connect offers Logarithmic, Cubic, and Linear as volume curve options.

I guess my concept is that the dynamic range should not change with volume -- if you adjust the volume (using software or hardware) from one player configuration so it matches the volume from another configuration (CDSP on versus CDSP off) they should sound substantially identical to our ears volumewize. there will be differences from the dithering and other effects. If there seems to be a 'contouring' effect that leads me to think of a loudness profile.

What I mean is that a volume setting of 32 "volumewise" is totally dependent on the curve that maps the digital volume value to dB volume. I think we can't compare the ALSA volume to the CDSP volume at 32 and say CDSP is reducing the volume. If one is using a linear volume curve, than 32 = 32% of max dB (and 99 = 99% of max), but if the volume curve is logarithmic, then 32 = 10% of max dB (and 99 is still about 99% of max). 

I think I have that right, I am still trying to wrap my head around exactly how the relationships work. I've found no evidence of loudness or contouring in CDSP unless you specifically introduce it in the pipeline.

[Tim Curtis]

Mentioned earlier but the test to determine whether there is some sort of attenuation in the Camila pipeline when there is no -gain anywhere in the pipeline, or whether its just diffs in volume curves is to set moode to output 0dB/100% volume with and w/o camilla on.

There "should" be no difference in SPL at the ear. If there is then maybe someone could post to the camilla thread at diyAudio and ask if there is some sort of built in attenuation.

[varunach]

Which model WiFi adapter?

@Tim Curtis i bought this simple one since it is based on 5730 chipset that works out of the box. I gotta say its not the greatest since I only get 25 - 30 Mbps on it in my room but gets the job done. The Moode UI loadtime is longer and not snappy which I can live with. I think once the Moode Kernel gets to Linux 6.2 version there is a lot more wifi adapter options available because of compatibility with in-kernel drivers. 

https://www.amazon.com/dp/B0BHW6T96R?psc...ct_details 

I also ordered this dual band adapter from aliexpress because it was on the recommended list for plug n play. If this works better, I'll return the other one above.

https://www.aliexpress.us/item/3256805749323751.html?

Got the dual band from ali yesterday and put together the stack: rpi-Q7ll-ProtoDAC. It worked for a few minutes then the WiFi became unstable, dropping in and out and struggling to connect via Spotify (I see the player but it won’t connect).
Do you have the same experience @varunach

I found a WiFi to Ethernet adapter I had lying and that fixed the problem.
Finally I’m able to enjoy the potential of the ProtoDAC.

The adapter I got from AliExpress works great for me