[igt-dev] [PATCH i-g-t v6 3/7] tests/kms_chamelium: test we receive a signal from both audio channels
Martin Peres
martin.peres at linux.intel.com
Tue Apr 23 07:41:21 UTC 2019
On 17/04/2019 15:43, Simon Ser wrote:
> This commit updates the audio test to make sure we receive a signal from both
> audio channels. However this commit doesn't check that left and right channels
> are not swapped. Such a check requires some more work (because the Chamelium
> device does swap left and right channels) and will be implemented in a future
> commit.
>
> This commit adds a new channel argument to audio_signal_add_frequency, to add
> a frequency to a single channel only.
>
> Some light refactoring has been performed (a proper audio_signal_deinit
> function has been introduced) and logging has been improved.
>
> Signed-off-by: Simon Ser <simon.ser at intel.com>
> ---
> lib/igt_alsa.c | 26 +++++++--
> lib/igt_audio.c | 119 +++++++++++++++++++++++++-----------------
> lib/igt_audio.h | 11 ++--
> tests/kms_chamelium.c | 56 ++++++++++++--------
> 4 files changed, 137 insertions(+), 75 deletions(-)
>
> diff --git a/lib/igt_alsa.c b/lib/igt_alsa.c
> index fc6d336b..a478686a 100644
> --- a/lib/igt_alsa.c
> +++ b/lib/igt_alsa.c
> @@ -182,6 +182,8 @@ static char *alsa_resolve_indentifier(const char *device_name, int skip)
> continue;
> }
>
> + igt_debug("Matched device \"%s\"\n", pcm_name);
> +
Should have been a separate commit.
> snprintf(name, sizeof(name), "hw:%d,%d", card,
> dev);
>
> @@ -329,6 +331,9 @@ static bool alsa_test_configuration(snd_pcm_t *handle, int channels,
> {
> snd_pcm_hw_params_t *params;
> int ret;
> + unsigned int min_channels, max_channels;
> + unsigned int min_rate, max_rate;
> + int min_rate_dir, max_rate_dir;
>
> snd_pcm_hw_params_alloca(¶ms);
>
> @@ -337,12 +342,24 @@ static bool alsa_test_configuration(snd_pcm_t *handle, int channels,
> return false;
>
> ret = snd_pcm_hw_params_test_rate(handle, params, sampling_rate, 0);
> - if (ret < 0)
> + if (ret < 0) {
> + snd_pcm_hw_params_get_rate_min(params, &min_rate, &min_rate_dir);
> + snd_pcm_hw_params_get_rate_max(params, &max_rate, &max_rate_dir);
> + igt_debug("Output device supports rates between %u and %u, "
> + "requested %d\n",
> + min_rate, max_rate, sampling_rate);
> return false;
> + }
This likely should be in patch 5/7.
>
> ret = snd_pcm_hw_params_test_channels(handle, params, channels);
> - if (ret < 0)
> + if (ret < 0) {
> + snd_pcm_hw_params_get_channels_min(params, &min_channels);
> + snd_pcm_hw_params_get_channels_max(params, &max_channels);
> + igt_debug("Output device supports between %u and "
> + "%u channels, requested %d\n",
> + min_channels, max_channels, channels);
> return false;
> + }
>
> return true;
> }
> @@ -409,13 +426,16 @@ void alsa_configure_output(struct alsa *alsa, int channels,
> snd_pcm_t *handle;
> int ret;
> int i;
> + int soft_resample = 0; /* Don't allow ALSA to resample */
> + unsigned int latency = 0;
>
> for (i = 0; i < alsa->output_handles_count; i++) {
> handle = alsa->output_handles[i];
>
> ret = snd_pcm_set_params(handle, SND_PCM_FORMAT_S16_LE,
> SND_PCM_ACCESS_RW_INTERLEAVED,
> - channels, sampling_rate, 0, 0);
> + channels, sampling_rate,
> + soft_resample, latency);
Probably should have been a separate commit again ;)
> igt_assert(ret >= 0);
> }
>
> diff --git a/lib/igt_audio.c b/lib/igt_audio.c
> index 7624f565..a2a5c594 100644
> --- a/lib/igt_audio.c
> +++ b/lib/igt_audio.c
> @@ -35,7 +35,7 @@
> #include "igt_audio.h"
> #include "igt_core.h"
>
> -#define FREQS_MAX 8
> +#define FREQS_MAX 64
>
> /**
> * SECTION:igt_audio
> @@ -49,9 +49,10 @@
>
> struct audio_signal_freq {
> int freq;
> + int channel;
>
> - short *period;
> - int frames;
> + int16_t *period;
> + size_t period_len;
> int offset;
> };
>
> @@ -60,7 +61,7 @@ struct audio_signal {
> int sampling_rate;
>
> struct audio_signal_freq freqs[FREQS_MAX];
> - int freqs_count;
> + size_t freqs_count;
> };
>
> /**
> @@ -89,21 +90,28 @@ struct audio_signal *audio_signal_init(int channels, int sampling_rate)
> * audio_signal_add_frequency:
> * @signal: The target signal structure
> * @frequency: The frequency to add to the signal
> + * @channel: The channel to add this frequency to, or -1 to add it to all
> + * channels
> *
> * Add a frequency to the signal.
> *
> * Returns: An integer equal to zero for success and negative for failure
> */
> -int audio_signal_add_frequency(struct audio_signal *signal, int frequency)
> +int audio_signal_add_frequency(struct audio_signal *signal, int frequency,
> + int channel)
Is the alignment correct here? Doesn't look like it!
> {
> - int index = signal->freqs_count;
> + size_t index = signal->freqs_count;
> + struct audio_signal_freq *freq;
>
> - if (index == FREQS_MAX)
> - return -1;
> + igt_assert(index < FREQS_MAX);
> + igt_assert(channel < signal->channels);
>
> /* Stay within the Nyquist–Shannon sampling theorem. */
> - if (frequency > signal->sampling_rate / 2)
> + if (frequency > signal->sampling_rate / 2) {
> + igt_debug("Skipping frequency %d: too high for a %d Hz "
> + "sampling rate\n", frequency, signal->sampling_rate);
> return -1;
> + }
Should have been a separate commit.
>
> /* Clip the frequency to an integer multiple of the sampling rate.
> * This to be able to store a full period of it and use that for
> @@ -111,11 +119,14 @@ int audio_signal_add_frequency(struct audio_signal *signal, int frequency)
> */
> frequency = signal->sampling_rate / (signal->sampling_rate / frequency);
>
> - igt_debug("Adding test frequency %d\n", frequency);
> + igt_debug("Adding test frequency %d to channel %d\n",
> + frequency, channel);
> +
> + freq = &signal->freqs[index];
> + memset(freq, 0, sizeof(*freq));
> + freq->freq = frequency;
> + freq->channel = channel;
>
> - signal->freqs[index].freq = frequency;
> - signal->freqs[index].frames = 0;
> - signal->freqs[index].offset = 0;
> signal->freqs_count++;
>
> return 0;
> @@ -133,20 +144,17 @@ void audio_signal_synthesize(struct audio_signal *signal)
> {
> int16_t *period;
> double value;
> - int frames;
> + size_t period_len;
> int freq;
> int i, j;
>
> - if (signal->freqs_count == 0)
> - return;
> -
> for (i = 0; i < signal->freqs_count; i++) {
> freq = signal->freqs[i].freq;
> - frames = signal->sampling_rate / freq;
> + period_len = signal->sampling_rate / freq;
>
> - period = calloc(1, frames * sizeof(short));
> + period = calloc(1, period_len * sizeof(int16_t));
>
> - for (j = 0; j < frames; j++) {
> + for (j = 0; j < period_len; j++) {
> value = 2.0 * M_PI * freq / signal->sampling_rate * j;
> value = sin(value) * INT16_MAX / signal->freqs_count;
>
> @@ -154,26 +162,34 @@ void audio_signal_synthesize(struct audio_signal *signal)
> }
>
> signal->freqs[i].period = period;
> - signal->freqs[i].frames = frames;
> + signal->freqs[i].period_len = period_len;
> }
> }
>
> /**
> - * audio_signal_synthesize:
> + * audio_signal_deinit:
> + *
> + * Release the signal.
> + */
> +void audio_signal_deinit(struct audio_signal *signal)
Is the coding style to use deinit or fini?
> +{
> + audio_signal_reset(signal);
> + free(signal);
> +}
> +
> +/**
> + * audio_signal_reset:
> * @signal: The target signal structure
> *
> * Free the resources allocated by audio_signal_synthesize and remove
> * the previously-added frequencies.
> */
> -void audio_signal_clean(struct audio_signal *signal)
> +void audio_signal_reset(struct audio_signal *signal)
> {
> - int i;
> + size_t i;
>
> for (i = 0; i < signal->freqs_count; i++) {
> - if (signal->freqs[i].period)
> - free(signal->freqs[i].period);
> -
> - memset(&signal->freqs[i], 0, sizeof(struct audio_signal_freq));
> + free(signal->freqs[i].period);
> }
>
> signal->freqs_count = 0;
> @@ -183,44 +199,45 @@ void audio_signal_clean(struct audio_signal *signal)
> * audio_signal_fill:
> * @signal: The target signal structure
> * @buffer: The target buffer to fill
> - * @frames: The number of frames to fill
> + * @samples: The number of samples to fill
> *
> - * Fill the requested number of frames to the target buffer with the audio
> + * Fill the requested number of samples to the target buffer with the audio
> * signal data (in interleaved S16_LE format), at the requested sampling rate
> * and number of channels.
> */
> -void audio_signal_fill(struct audio_signal *signal, int16_t *buffer, int frames)
> +void audio_signal_fill(struct audio_signal *signal, int16_t *buffer,
> + size_t buffer_len)
> {
> int16_t *destination, *source;
> + struct audio_signal_freq *freq;
> int total;
> - int freq_frames;
> - int freq_offset;
> int count;
> int i, j, k;
>
> - memset(buffer, 0, sizeof(int16_t) * signal->channels * frames);
> + memset(buffer, 0, sizeof(int16_t) * signal->channels * buffer_len);
>
> for (i = 0; i < signal->freqs_count; i++) {
> + freq = &signal->freqs[i];
> total = 0;
>
> - while (total < frames) {
> - freq_frames = signal->freqs[i].frames;
> - freq_offset = signal->freqs[i].offset;
> + igt_assert(freq->period);
>
> - source = signal->freqs[i].period + freq_offset;
> + while (total < buffer_len) {
> + source = freq->period + freq->offset;
> destination = buffer + total * signal->channels;
>
> - count = freq_frames - freq_offset;
> - if (count > (frames - total))
> - count = frames - total;
> + count = freq->period_len - freq->offset;
> + if (count > buffer_len - total)
> + count = buffer_len - total;
>
> - freq_offset += count;
> - freq_offset %= freq_frames;
> -
> - signal->freqs[i].offset = freq_offset;
> + freq->offset += count;
> + freq->offset %= freq->period_len;
>
> for (j = 0; j < count; j++) {
> for (k = 0; k < signal->channels; k++) {
> + if (freq->channel >= 0 &&
> + freq->channel != k)
> + continue;
> destination[j * signal->channels + k] += source[j];
> }
> }
> @@ -237,11 +254,11 @@ void audio_signal_fill(struct audio_signal *signal, int16_t *buffer, int frames)
> * sampling_rate is given in Hz. data_len is the number of elements in data.
> */
> bool audio_signal_detect(struct audio_signal *signal, int sampling_rate,
> - double *data, size_t data_len)
> + int channel, double *data, size_t data_len)
> {
> size_t bin_power_len = data_len / 2 + 1;
> double bin_power[bin_power_len];
> - bool detected[signal->freqs_count];
> + bool detected[FREQS_MAX];
> int ret, freq_accuracy, freq, local_max_freq;
> double max, local_max, threshold;
> size_t i, j;
> @@ -308,6 +325,10 @@ bool audio_signal_detect(struct audio_signal *signal, int sampling_rate,
> * invalid. */
> if (bin_power[i] < threshold) {
> for (j = 0; j < signal->freqs_count; j++) {
> + if (signal->freqs[j].channel >= 0 &&
> + signal->freqs[j].channel != channel)
> + continue;
> +
> if (signal->freqs[j].freq >
> local_max_freq - freq_accuracy &&
> signal->freqs[j].freq <
> @@ -340,6 +361,10 @@ bool audio_signal_detect(struct audio_signal *signal, int sampling_rate,
>
> /* Check that all frequencies we generated have been detected. */
> for (i = 0; i < signal->freqs_count; i++) {
> + if (signal->freqs[i].channel >= 0 &&
> + signal->freqs[i].channel != channel)
> + continue;
> +
> if (!detected[i]) {
> igt_debug("Missing frequency: %d\n",
> signal->freqs[i].freq);
> diff --git a/lib/igt_audio.h b/lib/igt_audio.h
> index 4aa43e69..fe26bb57 100644
> --- a/lib/igt_audio.h
> +++ b/lib/igt_audio.h
> @@ -35,12 +35,15 @@
> struct audio_signal;
>
> struct audio_signal *audio_signal_init(int channels, int sampling_rate);
> -int audio_signal_add_frequency(struct audio_signal *signal, int frequency);
> +void audio_signal_deinit(struct audio_signal *signal);
> +int audio_signal_add_frequency(struct audio_signal *signal, int frequency,
> + int channel);
> void audio_signal_synthesize(struct audio_signal *signal);
> -void audio_signal_clean(struct audio_signal *signal);
> -void audio_signal_fill(struct audio_signal *signal, int16_t *buffer, int frames);
> +void audio_signal_reset(struct audio_signal *signal);
> +void audio_signal_fill(struct audio_signal *signal, int16_t *buffer,
> + size_t buffer_len);
> bool audio_signal_detect(struct audio_signal *signal, int sampling_rate,
> - double *data, size_t data_len);
> + int channel, double *data, size_t data_len);
> size_t audio_extract_channel_s32_le(double *dst, size_t dst_cap,
> int32_t *src, size_t src_len,
> int n_channels, int channel);
> diff --git a/tests/kms_chamelium.c b/tests/kms_chamelium.c
> index 014a22b3..d336612f 100644
> --- a/tests/kms_chamelium.c
> +++ b/tests/kms_chamelium.c
> @@ -777,16 +777,16 @@ do_test_display_audio(data_t *data, struct chamelium_port *port,
> struct alsa *alsa, int playback_channels,
> int playback_rate)
> {
> - int ret, capture_rate, capture_channels, msec;
> + int ret, capture_rate, capture_channels, msec, freq;
> struct chamelium_audio_file *audio_file;
> struct chamelium_stream *stream;
> enum chamelium_stream_realtime_mode stream_mode;
> struct audio_signal *signal;
> int32_t *recv, *buf;
> double *channel;
> - size_t i, streak, page_count;
> + size_t i, j, streak, page_count;
> size_t recv_len, buf_len, buf_cap, buf_size, channel_len;
> - bool ok;
> + bool ok, success;
> char dump_suffix[64];
> char *dump_path = NULL;
> int dump_fd = -1;
> @@ -794,10 +794,15 @@ do_test_display_audio(data_t *data, struct chamelium_port *port,
> struct audio_state state = {};
>
> if (!alsa_test_output_configuration(alsa, playback_channels,
> - playback_rate))
> + playback_rate)) {
> + igt_debug("Skipping test with sample rate %d and %d channels "
> + "because selected output devices don't support this "
> + "configuration\n", playback_rate, playback_channels);
> return false;
> + }
>
> - igt_debug("Testing with playback sampling rate %d\n", playback_rate);
> + igt_debug("Testing with playback sampling rate %d and %d channels\n",
> + playback_rate, playback_channels);
> alsa_configure_output(alsa, playback_channels, playback_rate);
>
> chamelium_start_capturing_audio(data->chamelium, port, false);
> @@ -825,8 +830,12 @@ do_test_display_audio(data_t *data, struct chamelium_port *port,
> signal = audio_signal_init(playback_channels, playback_rate);
> igt_assert(signal);
>
> - for (i = 0; i < test_frequencies_count; i++)
> - audio_signal_add_frequency(signal, test_frequencies[i]);
> + for (i = 0; i < test_frequencies_count; i++) {
> + for (j = 0; j < playback_channels; j++) {
> + freq = test_frequencies[i];
> + audio_signal_add_frequency(signal, freq, j);
> + }
> + }
> audio_signal_synthesize(signal);
>
> state.signal = signal;
> @@ -851,10 +860,11 @@ do_test_display_audio(data_t *data, struct chamelium_port *port,
> recv = NULL;
> recv_len = 0;
>
> + success = false;
> streak = 0;
> msec = 0;
> i = 0;
> - while (streak < MIN_STREAK && msec < AUDIO_TIMEOUT) {
> + while (!success && msec < AUDIO_TIMEOUT) {
> ok = chamelium_stream_receive_realtime_audio(stream,
> &page_count,
> &recv, &recv_len);
> @@ -872,21 +882,27 @@ do_test_display_audio(data_t *data, struct chamelium_port *port,
> igt_assert(write(dump_fd, buf, buf_size) == buf_size);
> }
>
> - /* TODO: check other channels too, not just the first one */
> - audio_extract_channel_s32_le(channel, channel_len, buf, buf_len,
> - capture_channels, 0);
> -
> msec = i * channel_len / (double) capture_rate * 1000;
> igt_debug("Detecting audio signal, t=%d msec\n", msec);
>
> - if (audio_signal_detect(signal, capture_rate, channel,
> - channel_len))
> - streak++;
> - else
> - streak = 0;
> + for (j = 0; j < playback_channels; j++) {
> + igt_debug("Processing channel %zu\n", j);
> +
> + audio_extract_channel_s32_le(channel, channel_len,
> + buf, buf_len,
> + capture_channels, j);
> +
> + if (audio_signal_detect(signal, capture_rate, j,
> + channel, channel_len))
> + streak++;
> + else
> + streak = 0;
> + }
>
> buf_len = 0;
> i++;
> +
> + success = streak == MIN_STREAK * playback_channels;
> }
>
> igt_debug("Stopping audio playback\n");
> @@ -921,12 +937,10 @@ do_test_display_audio(data_t *data, struct chamelium_port *port,
> chamelium_destroy_audio_file(audio_file);
> }
>
> - audio_signal_clean(signal);
> - free(signal);
> -
> + audio_signal_deinit(signal);
> chamelium_stream_deinit(stream);
>
> - igt_assert(streak == MIN_STREAK);
> + igt_assert(success);
> return true;
> }
>
With the coding style issues fixed:
Reviewed-by: Martin Peres <martin.peres at linux.intel.com>
More information about the igt-dev
mailing list