r2r with snapclient

[parabyte]

is there any possibility of a example sketch on how to use a 8bit r2r with snapclient, really appreciate some pointers, with tone generators etc i been playing with the new r2r sound device, but with snap im completely lost,

[pschatzmann]

See the AudioTools Wiki

[parabyte]

i was hoping if anyone had any pointers, i cant seem to get this code to work quiet right,

#include <SPI.h>
#include <Ethernet.h>
#include "AudioTools.h"
#include "SnapClient.h"
#include "AudioLibs/R2ROutput.h"

AudioInfo from;
AudioInfo to(8000, 2, 8); // Resampling to 8000 Hz, 8-bit, 2 channels

EthernetClient eth; // Ethernet client for network communication
WAVDecoder pcm; // PCM decoder
R2ROutput out; // R2R DAC output
FormatConverterStream formatConverter(out); // Initialize the FormatConverterStream

// Synchronization configuration for SnapClient (optional)
SnapTimeSyncDynamic synch(172, 10);

// SnapClient connects the Ethernet client, the decoder, and the resampling stream
SnapClient client(eth, formatConverter, pcm); // Pass formatConverter instead of 'out'

// MAC address for Ethernet shield
byte mac[] = {0xDE, 0xAD, 0xBE, 0xEF, 0xFE, 0xED};

void setup() {
Serial.begin(115200);
AudioLogger::instance().begin(Serial, AudioLogger::Info);

while (!Serial); // Wait for Serial Monitor to open
Serial.println("Setup started");

// Initialize Ethernet
SPI.begin(); // Use default pins
Ethernet.init(A4);
Ethernet.begin(mac);

// Check for Ethernet hardware presence
if (Ethernet.hardwareStatus() == EthernetNoHardware) {
Serial.println("Ethernet shield was not found. Sorry, can't run without hardware.");
while (true);
}

// Check link status
if (Ethernet.linkStatus() == LinkOFF) {
Serial.println("Ethernet cable is not connected.");
}
Serial.print("Local IP: ");
Serial.println(Ethernet.localIP());

// Wait for link
while (Ethernet.linkStatus() != LinkON) {
delay(10);
}

// Initialize the FormatConverterStream
if (!formatConverter.begin(from, to)) { // Setup conversion from 'from' to 'to'
Serial.println("Error initializing FormatConverterStream");
while (true);
}

// Initialize the R2R DAC output with the specified configuration
auto config = out.defaultConfig();
config.copyFrom(to);
config.channel1_pins = {25, 26, 27, 28, 29, 30, 31, 32};
config.channel2_pins = {33, 38, 39, 40, 41, 42, 43, 44};
if (!out.begin(config)) {
Serial.println("Error initializing R2R output");
while (true);
}

// Configure SnapClient to use the specified server IP
client.setServerIP(IPAddress(192, 168, 10, 104));
if (!client.begin(synch)) {
Serial.println("Could not connect to Snap server");
while (true);
}

Serial.println("Setup complete, starting audio streaming...");
}

void loop() {
// Continuously call the SnapClient loop to keep processing
client.doLoop();
}

[pschatzmann]

This looks correct to me. What is the problem exactly ?

[parabyte]

no sound output no matter what i do

[parabyte]

04:10:06.751 -> [E] ResampleStream.h : 191 - write error 8 vs 1020
04:10:06.751 -> [E] ResampleStream.h : 191 - write error 0 vs 1020
04:10:06.751 -> [E] ResampleStream.h : 191 - write error 0 vs 1020

[parabyte]

this is the bit that really confuses me,

04:10:06.751 -> [I] AudioTypes.h : 128 - out: sample_rate: 1000 / channels: 2 / bits_per_sample: 16
04:10:06.751 -> [I] AudioTypes.h : 128 - out: sample_rate: 1000 / channels: 2 / bits_per_sample: 8

[pschatzmann]

Change
AudioInfo to(8000, 2, 8); // Resampling to 8000 Hz, 8-bit, 2 channels
to
AudioInfo to(8000, 2, 16); // Resampling to 8000 Hz, 16-bit, 2 channels

[parabyte]

[parabyte]

same error after suggested change

[parabyte]

04:16:33.053 -> [E] ResampleStream.h : 191 - write error 8 vs 1020
04:16:33.053 -> [E] ResampleStream.h : 191 - write error 0 vs 1020
04:16:33.053 -> [E] ResampleStream.h : 191 - write error 0 vs 1020
04:16:33.053 -> [E] ResampleStream.h : 191 - write error 0 vs 1020

[parabyte]

ill include some of the output from serial on the device

[parabyte]

[I] ResampleStream.h : 129 - -> ResampleStream:
[I] AudioTypes.h : 128 - out: sample_rate: 44100 / channels: 2 / bits_per_sample: 16
[I] AudioTypes.h : 128 - out: sample_rate: 8000 / channels: 2 / bits_per_sample: 16
[I] AudioStreamsConverter.h : 176 - --> ChannelFormatConverterStream
[I] AudioTypes.h : 128 - out: sample_rate: 44100 / channels: 2 / bits_per_sample: 16
[I] AudioStreamsConverter.h : 178 - begin 2 -> 2 channels
[I] AudioStreamsConverter.h : 24 - begin 2 -> 2 channels
[I] AudioStreamsConverter.h : 490 - begin 16 -> 16 bits
[I] AudioStreamsConverter.h : 496 - no bit conversion: 16 -> 16
[I] ResampleStream.h : 64 - begin step_size:
[I] ResampleStream.h : 145 - setStepSize:
[I] ResampleStream.h : 145 - setStepSize:
[I] ResampleStream.h : 129 - -> ResampleStream:
[I] AudioTypes.h : 128 - out: sample_rate: 8000 / channels: 2 / bits_per_sample: 16
[I] R2ROutput.h : 85 - Setup channel1 pin 25
[I] R2ROutput.h : 85 - Setup channel1 pin 26
[I] R2ROutput.h : 85 - Setup channel1 pin 27
[I] R2ROutput.h : 85 - Setup channel1 pin 28
[I] R2ROutput.h : 85 - Setup channel1 pin 29
[I] R2ROutput.h : 85 - Setup channel1 pin 30
[I] R2ROutput.h : 85 - Setup channel1 pin 31
[I] R2ROutput.h : 85 - Setup channel1 pin 32
[I] R2ROutput.h : 89 - Setup channel2 pin 33
[I] R2ROutput.h : 89 - Setup channel2 pin 38
[I] R2ROutput.h : 89 - Setup channel2 pin 39
[I] R2ROutput.h : 89 - Setup channel2 pin 40
[I] R2ROutput.h : 89 - Setup channel2 pin 41
[I] R2ROutput.h : 89 - Setup channel2 pin 42
[I] R2ROutput.h : 89 - Setup channel2 pin 43
[I] R2ROutput.h : 89 - Setup channel2 pin 44
[I] AudioTimerSTM32.h : 50 - virtual bool audio_tools::TimerAlarmRepeatingDriverSTM32::begin(audio_tools::repeating_timer_callback_t, uint32_t, audio_tools::TimeUnit)
[I] AudioTimerSTM32.h : 51 - Using timer TIM2
Setup complete, starting audio streaming...
[I] VolumeStream.h : 199 - setVolume: at 0
[I] VolumeStream.h : 199 - setVolume: at 1
[I] ResampleStream.h : 145 - setStepSize:
[I] ResampleStream.h : 129 - -> ResampleStream:
[I] AudioTypes.h : 128 - out: sample_rate: 8000 / channels: 2 / bits_per_sample: 16
[I] AudioTypes.h : 128 - out: sample_rate: 8000 / channels: 2 / bits_per_sample: 16
[I] AudioTypes.h : 128 - out: sample_rate: 44100 / channels: 2 / bits_per_sample: 16
[I] AudioEncoded.h : 75 - virtual void audio_tools::EncodedAudioOutput::addNotifyAudioChange(audio_tools::AudioInfoSupport&)
[I] ResampleStream.h : 64 - begin step_size:
[I] ResampleStream.h : 145 - setStepSize:
[I] ResampleStream.h : 129 - -> ResampleStream:
[I] AudioTypes.h : 128 - out: sample_rate: 44100 / channels: 2 / bits_per_sample: 16
[I] AudioEncoded.h : 75 - virtual void audio_tools::EncodedAudioOutput::addNotifyAudioChange(audio_tools::AudioInfoSupport&)
[I] CodecWAV.h : 56 - WAVHeader::write: 44 -> 44 -> 44
[I] CodecWAV.h : 63 - WAVHeader::begin: 44
[I] CodecWAV.h : 220 - WAVHeader sound_pos: 44
[I] CodecWAV.h : 221 - WAVHeader channels: 2
[I] CodecWAV.h : 222 - WAVHeader bits_per_sample: 16
[I] CodecWAV.h : 223 - WAVHeader sample_rate: 44100
[I] CodecWAV.h : 224 - WAVHeader format: 1
[I] CodecWAV.h : 420 - WAV sample_rate: 44100
[I] CodecWAV.h : 421 - WAV data_length: 0
[I] CodecWAV.h : 422 - WAV is_streamed: 1
[I] CodecWAV.h : 423 - WAV is_valid: true
[I] ResampleStream.h : 145 - setStepSize:
[I] ResampleStream.h : 129 - -> ResampleStream:
[I] AudioTypes.h : 128 - out: sample_rate: 8000 / channels: 2 / bits_per_sample: 16
[I] AudioTypes.h : 128 - out: sample_rate: 8000 / channels: 2 / bits_per_sample: 16
[I] CodecWAV.h : 443 - WAVDecoder writing first sound data
[I] R2ROutput.h : 70 - is_active = true
[E] ResampleStream.h : 191 - write error 8 vs 1020
[E] ResampleStream.h : 191 - write error 0 vs 1020
[E] ResampleStream.h : 191 - write error 0 vs 1020
[E] ResampleStream.h : 191 - write error 0 vs 1020
[E] ResampleStream.h : 191 - write error 0 vs 1020
[E] ResampleStream.h : 191 - write error 0 vs 1020
[E] ResampleStream.h : 191 - write error 0 vs 1020
[E] ResampleStream.h : 191 - write error 0 vs 1020
[E] ResampleStream.h : 191 - write error 0 vs 1020
[E] ResampleStream.h : 191 - write error 0 vs 1020
[E] ResampleStream.h : 191 - write error 0 vs 1020
[E] ResampleStream.h : 191 - write error 0 vs 1020

[pschatzmann]

Is the sine test sketch working with AudioInfo to(8000, 2, 16);

[parabyte]

yes working fine

[parabyte]

this is my snapserver log concerning this

2024-08-28 04-39-36.954 [Info] (Server) Hello from 00-00-00-00-00, host: arduino-snapclient, v0.0.2, ClientName: libsnapcast, OS: arduino, Arch: xtensa, Protocol version: 2

2024-08-28 04-39-39.085 [Info] (Server) onResync (Notification): 110.749 ms

2024-08-28 04-39-39.085 [Info] (AlsaStream) Too many frames available, fast forwarding from 6474 frames (146.803 ms) to 1323 frames (30 ms)

2024-08-28 04-39-39.757 [Error] (StreamSession) StreamSession write error (msg length: 1700): Connection reset by peer

2024-08-28 04-39-39.757 [Info] (StreamServer) onDisconnect: 00-00-00-00-00

2024-08-28 04-39-39.757 [Error] (StreamSessionTCP) Error reading message header of length 0: End of file

2024-08-28 04-39-39.757 [Error] (StreamSessionTCP) Error in socket shutdown: Transport endpoint is not connected

[pschatzmann]

It seems that I need to make the buffer size configurable for R2ROutput

[parabyte]

do you mean this?

from R2ROutput.h

protected:
TimerAlarmRepeating timer;
NBuffer<uint8_t> buffer{1024, 2};
R2RConfig rcfg;

[parabyte]

ifso larger sizes make no difference

[pschatzmann]

Try to find out the len in the write method(): the buffer size needs to be at least that size

I committed a correction that prints an error message when the buffer size is too small...

ps. I am confused about your pin numbers: for stm32 I would expect something like PA1, PA2, PA3

[parabyte]

Iv always used pin numbers for my stm32 code, stm32duo maps them to the correct places.

now when i start the code on a fresh power cycle of the soc, the dac makes a brief digital noise sound, then nothing.

can you advise on where i should be looking to debug?

if i set AudioInfo to(8000, 2, 8); i get

14:40:37.340 -> [E] ResampleStream.h : 191 - write error 8 vs 1020
14:40:37.340 -> [E] ResampleStream.h : 191 - write error 0 vs 1020

if i set it to AudioInfo to(8000, 1, 8); no error messages, also no sound

But both instances initially have a few ms worth of digital noise,

[parabyte]

i have done a sanity test and tried {PB_12, PB_13, PB_14, PB_15, PB_10, PB_11, PA8, PA9}, im just trying to get 1 channel to work at the moment

[parabyte]

is the microcontroller fast enough to run snapclient and do all of these tasks?

its the CC variant of the blackpill?

[pschatzmann]

It should be fast enough: however I would expect the network to be the bottleneck.
Did you run some performance tests for the network ?

I committed some changes to the R2ROutput to perform blocking writes so that it is waiting if the buffer is full.

ps. don't use PB_12, PB_13... it should be PB12, PB13...

[parabyte]

Still no audio from my code above, but ran the network test with surprising results.

14:33:34.798 -> ==> Bytes per second: 14000
14:33:36.852 -> ==> Bytes per second: 16000
14:33:38.970 -> ==> Bytes per second: 15000
14:33:41.059 -> ==> Bytes per second: 16000
14:33:43.207 -> ==> Bytes per second: 15000
14:33:45.326 -> ==> Bytes per second: 15000
14:33:47.380 -> ==> Bytes per second: 16000

[pschatzmann]

This is way too slow: for 44100 samples per second (16 bit stereo) you would need at least 176400 bytes per second!

Try to increase the SPI frequency...

[parabyte]

oh i missed this before, cause i had it set to info,

E (682) gpio: GPIO can only be used as input mode

[parabyte]

does not specify which one, but i can poke around with my multimeter to work that out

[parabyte]

i believe pin 35 looking at documentation?

[parabyte]

{12,14,27,26,25,33,32, 15}

[parabyte]

[parabyte]

#include "WiFi.h"
#include "AudioTools.h"
#include "SnapClient.h"
#include "AudioLibs/R2ROutput.h"

#define ARDUINO_LOOP_STACK_SIZE (10 * 1024)

AudioInfo from;

AudioInfo to(20000, 1, 8); // Sample rate: 44100 Hz, Mono, 8-bit

WAVDecoder pcm;
R2ROutput out; // R2R DAC output
FormatConverterStream formatConverter(out); // Initialize the FormatConverterStream

WiFiClient wifi;
SnapTimeSyncDynamic synch(172, 10); // optional configuratioin
SnapClient client(wifi, formatConverter, pcm);

const int pins1[] = {12,14,27,26,25,33,32, 15};

void setup() {
Serial.begin(115200);
AudioLogger::instance().begin(Serial, AudioLogger::Info);

// login to wifi -> Define values in SnapConfig.h or replace them here
WiFi.begin("sid", "pwd");
Serial.print("Connecting to WiFi ..");
while (WiFi.status() != WL_CONNECTED) {
Serial.print('.');
delay(1000);
}

// print ip address
Serial.println();
Serial.println(WiFi.localIP());

// Initialize the FormatConverterStream
if (!formatConverter.begin(from, to)) { // Setup conversion from 'from' to 'to'
Serial.println("Error initializing FormatConverterStream");
while (true);
}

// Start R2R DAC
Serial.println("Starting R2R DAC...");
auto config = out.defaultConfig();
config.copyFrom(to); // Copy audio info to config

// Set up pins for the R2R DAC
for (int i = 0; i < sizeof(pins1) / sizeof(pins1[0]); i++) {
int pin = pins1[i]; // Create a non-const temporary variable
config.channel1_pins.push_back(pin);
}

// Initialize R2R output with the config
if (!out.begin(config)) {
Serial.println("Failed to initialize R2R output!");
while (true);
}

// Define CONFIG_SNAPCAST_SERVER_HOST in SnapConfig.h or here
client.setServerIP(IPAddress(192,168,10,104));

// start snap client
client.begin(synch);
}

void loop() {
client.doLoop();
}

17:06:36.721 -> [I] ResampleStream.h : 129 - -> ResampleStream:
17:06:36.754 -> [I] AudioTypes.h : 128 - out: sample_rate: 20000 / channels: 2 / bits_per_sample: 16
17:06:36.754 -> [I] AudioTypes.h : 128 - out: sample_rate: 20000 / channels: 1 / bits_per_sample: 8
17:06:36.754 -> [I] CodecWAV.h : 443 - WAVDecoder writing first sound data
17:06:36.754 ->
17:06:36.754 -> assert failed: size_t audio_tools::ChannelFormatConverterStreamT::write(const uint8_t*, size_t) [with T = short int; size_t = unsigned int; uint8_t = unsigned char] AudioStreamsConverter.h:48 (res

[parabyte]

same rebooting behaviour with AudioInfo to(8000, 1, 8);
just to confirm

[pschatzmann]

I don't understand why you want to convert it to 8 bit.

It is the responsibility of the R2ROutput to provide as many bits as the number of pins that you have defined...

[parabyte]

with correction in place same behaviour

AudioInfo to(8000, 1, 16);

[parabyte]

on plus side, i have just pulled your latest source and built the tone generator,

[pschatzmann]

Did you try to optimize the parameters that are available in the R2RConfig ?
e.g. buffer_count, blocking_retry_delay_ms to improve on the imperfections after every 32/32 waves

[parabyte]

soon as my osciloscope arrives i will have far supior screenshots and measurements, have not forgotten

[pschatzmann]

by the way... did you ever try to increase the sampling rate to 44100 or 48000 ?

[parabyte]

i will try now!

[parabyte]

44100

[parabyte]

48000 not operating at correct frequency

[parabyte]

in the snapclient code if i specify AudioInfo to(44100, 1, 8);

the audio is very choppy the soc occasionally reboots, but there is some sort of life, audio so garbled its uninteligable

tried adding BufferedStream with a 1024 buffer, but that made no difference to audio quality

[parabyte]

any samplerate other then 44100 causes the esp32 to bootloop