Why does the ALSA AAF plugin transmit more than 96,000 AVTP packets per second?

[GitHub issues - opened <github@alsa-project.org>]

alsa-project/alsa-utils issue #334 was opened from liuyq0509-ux:

Question

I'm using an ALSA AAF PCM plugin to transmit audio over AVTP.

My audio configuration is:

Sample rate:      48000 Hz
Format:           S16_BE
Channels:         2
frames_per_pdu:   12

According to my understanding, frames_per_pdu = 12 means each AVTP AAF packet contains 12 audio frames.

Therefore, the expected packet rate should be:

48000 / 12 = 4000 packets/s

However, when I run:

speaker-test -p 25000 \
    -F S16_BE \
    -c 2 \
    -r 48000 \
    -D aaf0

and capture the traffic with tcpdump, I see more than 96,000 AVTP packets per second.

This is much higher than the expected 4,000 packets/s.

For example:

Expected:
48000 frames/s / 12 frames/PDU = 4000 AVTP packets/s

Observed:
> 96000 AVTP packets/s

I would like to understand where this packet transmission rate is coming from.

Questions
Does speaker-test control the actual AVTP packet transmission rate?
Is speaker-test -p 25000 related to the AVTP packet transmission interval, or does it only control the ALSA PCM period size?
With:
48 kHz
2 channels
frames_per_pdu = 12

should the AAF plugin normally transmit approximately:

4000 AVTP packets/s

regardless of the speaker-test period setting?

If I observe more than 96,000 packets/s, could the AAF plugin be sending the same audio data repeatedly, or could there be multiple packets generated for each ALSA frame/period?
What is the recommended way to determine which component controls the actual AVTP transmission timing — speaker-test, ALSA, or the AAF PCM plugin?

I have confirmed that the captured packets are AVTP AAF packets. I would appreciate any explanation of how the ALSA AAF plugin determines the AVTP packet transmission interval.

Issue URL     : https://github.com/alsa-project/alsa-utils/issues/334
Repository URL: https://github.com/alsa-project/alsa-utils