We just stumbled upon the new Rock Pi S board and have several questions regarding the audio ports.
(1) Where are the channels 7 & 8 on the on digital output? (they seem to be ignored in the schemantics)
(2) Are the digitial MEMS supported in the ALSA driver stack at all?
(3) Can anyone provide us an output of aplay -L (alsamixed device listening)
Thanks for the info. Regarding the audio, we stumbled upon a few things:
(1) First of all there is NO audio driver for the rock pi S
(2) Regarding the driver: Providing a driver for #### 26-pin Header 2 MICN1-8 will not work,
as the MICS will never have a synchronized clock rate, due to your schemantics. If you
cannot provide synchronized audio channels, any sort of beamforming will NOT work
(3) The only chance with that board design is to use the I2S0_8CH_SDI0 interface. However, channel 7 and 8 are missing. Are you currently working on that specific problem?
We can assist you to provide a driver for this board.
Just keep in mind that until you do not have a valid audio interface, the device is USELESS for any kind of smart-home or ASR. It will not get accepted by the communiy, and it will not be sold accordingly in Millions of pieces such as the RPI3. It will die as quickly as it was launched. Acctually I’m very outraged about your marketing strategy, providing a small embedded system with multichannel MICS, pretending to be usable for Alexa, Siri, …, but NO audio is working at all!
You over exaggerating, honestly. There is enough projects that doesn’t require audio to work which RockPiS can satisfy, so please, restrain from words
Because you != community.
RockPiS is NOT rpi3. The closer is Pi Zero W.
Again, you over exaggerating. You already was told they are working on it. If you wanna help and you do understand how that works - i believe github is open, otherwise, your over exaggerating comments don’t help at all
“An extremely “rocking” mini computer for your projects! Get creative!! :-)”
I don’t see any Siri in there, do you? Or mean “projects”? Well, projects is vague word for sure
We understand your point and your feedback is very important. The issue is what and how the users want to use the mic/audio is not clear for us, ie do the user prefer to use analog mic or digital mic. We also need to consider the maker gpio/spi/uart etc. So the two 26P header is very few and it’s really hard for us to choose which feature to keep and which to abandon. We have realised the i2s issue on the header missing essential clock. Below is what we are going to revise for the 26p header:
Sorry for the late reply. First of all I didn’t want to start a flame war here. Your product is nice, as it is. The only thing that popped up is an unique selling point. Regarding micro arm devices, there are plenty out there, like this one http://linuxgizmos.com/worlds-smallest-quad-core-sbc-starts-at-8-dollars/. However there is no small multi mems board available. The rockchip supports this by design, however you have to implement it accordingly.
Regarding the choice of mics, the best option is to use MEMS instead of analog ones. The use of MEMS in combination with an analog processing is no problem or vice versa, as you can use DAC/ADC.
Please have a look at the following schematic, where 2 ADCS are used to feed the audio input
The use of multiple MEMS does not have any advantage unless you are using it for beamforming, i.e. ASR and smart home usage. If you are planning to support multiple MEMS which might be one real USP of your board, the MEMS have to be clock synchronized.
Regarding the choice of MICS on the pin. The rockchip codec seems to be only processing stereo signals. So after having several discussions with some tech experts, we came to the conclusion that you cannot use the standard interface, otherwise the mics are not clock synchronized. If you are planning to support multiple MEMS, the only way would be to use the I2S interface, i.e. I2S0/I2S1 with 8ch: supports up to 8 channels (8xTX or 8xRX).
At the current state, the board (insert link, order code, etc.) can only connect to 6 digital mems microphones via the i2s interface. This is due to pin Y14 being left unconnected (see page 7 of the schematics pdf). Further, it would be nice to have synchronous sampling of the mems microphones and the line-out. According to the RK3308 datasheet, this happens at different hardware units, (audio-codec, i2s interface). In order to have synchronous sampling, the audio driver must take care of reading/writing to those units at the same time (i.e. with the same DMA cycle).
There is a reference implementation available for one variant of the rockchip. You might find some infos here how the sound module was implemented:
For a reference sound driver design intro, using multiple MICS have a look at
However, the are not using the rockchip
We will come back to you, if we find a valid driver design using the proposed I2S with the rockchip.
I was able to get some sound out of my Rock Pi S. I connected headphones to GND, PIN 25 and PIN 26… BUT: I only got sound on one ear. I switched PIN 25 and 26 and got sound on the other ear so it not broken headphones.
I would be extremly cool, if you add this driver specification to your board, as it seems to fullfill all requirements for multi-channel audio recording and playback. If you are experiencing troubles please drop us a line here.