[MiniMax-H3, Pika H3 Flex] ref2va fails on a reference soundtrack that is not at 32 kHz without torchaudio #15245

Describe the bug

The ref2va setup steps of MiniMax-H3 (MiniMaxH3Ref2VASetupStep) and Pika H3 Flex (PikaH3FlexRef2VASetupStep) resample a reference soundtrack onto the audio VAE's sample rate (32 kHz) with torchaudio.transforms.Resample. torchaudio is not a diffusers dependency, so a reference soundtrack at any other rate, such as the 44.1 or 48 kHz audio track of most videos, fails with an ImportError.

Installing torchaudio is not a fix everywhere: it has no release for recent torch versions (the last on PyPI is 2.11.1).

I intend to fix this in #15199, with a plain-PyTorch port of torchaudio's default band-limited resampler.

Reproduction

In an environment without torchaudio:

import torch
from diffusers.modular_pipelines.minimax_h3.before_encoder import MiniMaxH3Ref2VASetupStep

waveform = torch.randn(1, 48000 * 3)  # 3 s of mono audio at 48 kHz
MiniMaxH3Ref2VASetupStep._normalize_audio_condition(
    waveform, sample_rate=48000, target_sample_rate=32000, max_duration=2.0
)

The same applies to PikaH3FlexRef2VASetupStep._normalize_audio_condition in diffusers.modular_pipelines.pika_h3_flex.before_encoder.

Logs

ImportError: Resampling a MiniMax-H3 reference soundtrack from 48000 Hz to 32000 Hz needs `torchaudio`. Pass a waveform already at the audio VAE's sample rate to do without it.

System Info

• diffusers: main
• torch: a version with no matching torchaudio release (> 2.11)
• torchaudio: not installed

Who can help?

@DN6 @yiyixuxu