Compiler Warnings using Arduino IDE 2.3.2

[Greg-R]

These are not show stoppers, but they are annoying. Compiler output below.
Can any of these be resolved?

Thank you,
Greg

/home/raven/Arduino/libraries/OpenAudio_ArduinoLibrary/output_spdif3_f32.cpp: In member function 'void AudioOutputSPDIF3_F32::begin()':
/home/raven/Arduino/libraries/OpenAudio_ArduinoLibrary/output_spdif3_f32.cpp:77:15: warning: 'void* memset(void*, int, size_t)' clearing an object of type 'class audio_block_f32_t' with no trivial copy-assignment; use value-initialization instead [-Wclass-memaccess]
77 | memset(&block_silent, 0, sizeof(block_silent));
| ~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from /home/raven/Arduino/libraries/OpenAudio_ArduinoLibrary/output_spdif3_f32.h:27,
from /home/raven/Arduino/libraries/OpenAudio_ArduinoLibrary/output_spdif3_f32.cpp:28:
/home/raven/Arduino/libraries/OpenAudio_ArduinoLibrary/AudioStream_F32.h:34:7: note: 'class audio_block_f32_t' declared here
34 | class audio_block_f32_t {
| ^~~~~~~~~~~~~~~~~
/home/raven/Arduino/libraries/OpenAudio_ArduinoLibrary/utility/BTNRH_rfft.cpp: In function 'int BTNRH_FFT::bitrev(int, int)':
/home/raven/Arduino/libraries/OpenAudio_ArduinoLibrary/utility/BTNRH_rfft.cpp:31:30: warning: ISO C++17 does not allow 'register' storage class specifier [-Wregister]
31 | register int jj;
| ^~
Opening Teensy Loader...
Memory Usage on Teensy 4.1:
FLASH: code:211732, data:123428, headers:8896 free for files:7782408
RAM1: variables:146240, code:191368, padding:5240 free for local variables:181440
RAM2: variables:354496 free for malloc/new:169792

[boblark]

I will take these on. The "reg" doesn't apply and the spdif needs to only zero the data, I think. I cannot test the spdif, so after I change the zeroing, maybe somebody could test the functionality. Bob

[Greg-R]

I had to look up SPIDIF at Wikipedia. I checked my media devices and no SPIDIF.

I'm pretty sure I have 2 spare Teensies. So if it's a really important test, it should be theoretically possible to fire audio from one Teensy to another using that interface.

[JayShoe]

SPDIF is also similar to optical/toslink. It is not the same as ADAT because ADAT has 8 channels and SPDIF has 2. I mention this since it sounds like you are just now learning about SPDIF. 

Firing one Teensy into the other is a great way to test it for sure. Get ADAT working and you are a hero.

[Greg-R]

Searching around at Amazon, I found this:

https://www.amazon.com/PROZOR-Digital-Converter-Optical-Toslink/dp/B00KNNSKV0

The project I am working on is the T41 "Software Defined Transceiver". So it seems possible to add the SPDIF to my audio chain, and drive the Amazon device with the SPDIF output, ultimately getting analog audio to the headphones or speaker.
This looks like a great way to test the SPDIF output class, and it gives me an alternative way to get audio out of the radio.

[boblark]

Thanks, guys. I think it is a simple enough fix to be OK without hardware testing (I hate to say that, as it always ...). We just need to zero the data with something like

for(int i=0; i<block_silent.length; i++)
block_silent.data[i] = 0.0f;

and get rid of the memset call.

I will try this, but it will be a couple of days. Bob

[boblark]

I think they are fixed. Let me know of any problems. Bob

[Greg-R]

Compiled clean on T41 code and also the SPDIF output test rig.
Both are operating normally. Looks good here, thank you!

I have an AudioOutputSPDIF3_F32 example sketch if you would like to add it. Needs a little clean-up, and it will be ready to go.

[boblark]

Thanks for the confirmation, Greg. Yesterday I looked at your video on SPDIF test and ordered one of the SPDIF to DAC gadgets. I was planning to take your INO and make it simple as: CODEC in and output to both CODEC and SPDIF. I'll post it as an example, as you suggest. Thanks for all the help. Bob

[Greg-R]

I have a question on SPDIF. In the documentation, it says this:

"Native S/PDIF hardware is used, which is more efficient that use of I2S ports."

What exactly is the efficiency improvement? Lower bandwidth, lower processor cycles, or what?
I'm looking at SPDIF as an alternative to I2S, especially if it saves processor cycles relative to I2S, which
is the current audio output interface. Does anyone know the specifics behind this statement?