Supplementary listening page

RAV-Net Separated Audio Samples

Round-by-round listening examples corresponding to the comparative presentation in Figure 12. Each row follows the same order: preliminary separation (Round 1), iterative refinement (Rounds 2–4), and the ground-truth speech.

Examples
2 utterances
Format
WAV · mono · 16 kHz
Duration
2.55 s each
Iterations
R = 4

Listening guide

Compare each stage from left to right

Use the players below to hear how the separated speech evolves over successive rounds. Starting another player automatically pauses the one currently playing, which makes direct comparison easier.

Figure 12(a)

Example utterance A

id08696_1SBCWkvRRIc_00021

01

Round 1

Preliminary stage

02

Round 2

Iterative stage

03

Round 3

Iterative stage

04

Round 4

Final iterative stage

GT

Ground truth

Reference speech

View the corresponding spectrogram comparison

PDF preview unavailable. Open the spectrogram comparison.
Open Figure 12(a) in a new tab

Figure 12(b)

Example utterance B

id08696_yZDyZR1JlFo_00494

01

Round 1

Preliminary stage

02

Round 2

Iterative stage

03

Round 3

Iterative stage

04

Round 4

Final iterative stage

GT

Ground truth

Reference speech

View the corresponding spectrogram comparison

PDF preview unavailable. Open the spectrogram comparison.
Open Figure 12(b) in a new tab