Paper Audio Examples

Table of Contents

1. How To Use This Page
2. Figure Index
3. Paper Sections

Paper Companion

Use figure-oriented listening examples that match the paper narrative

This companion page is organized around paper sections and figures rather than the full challenge taxonomy. Each block includes short narrative framing and example captions so authors and reviewers can jump directly from manuscript claims to curated listening evidence.

Paper sections 4

Annotated examples 8

Playable clips 66

Asset policy Reuse

How To Use This Page

Map Audio To Claims

Each section corresponds to a paper-oriented figure or discussion block so it can be cited directly from the manuscript.

Read The Caption First

Every example includes a short caption that explains what the reader should pay attention to in the comparison.

Reuse Existing Assets

This page currently reuses the challenge audio binaries so paper curation can evolve without duplicating the full audio tree.

• Use headphones when possible.
• The paper page keeps the example ordering fixed to match the intended manuscript sequence.

Jump to top

Figure Index

Figure 1 Transparency Comparisons Across Bitrates

Use clean-speech examples to anchor the paper's discussion of bitrate-dependent transparency. The two examples below keep the comparison pool fixed while moving from a higher to a lower bitrate operating point.

Figure 2 Robustness Under Mild Real-World Distortions

These examples support the paper's discussion of mild noisy and reverberant inputs. The pair contrasts lower- and higher-bitrate operation while preserving the same submission roster.

Figure 3 Simultaneous Speaker Preservation

These examples support the paper's analysis of multi-talker preservation, including the special auxiliary comparison that appears in the 6 kbps condition.

Figure 4 Word Intelligibility Focused Comparisons

This section mirrors the paper's intelligibility discussion and keeps the target and alternative word pair visible so readers can judge which lexical contrast survives each system.

Jump to top

Paper Sections

Figure 1

Transparency Comparisons Across Bitrates

Use clean-speech examples to anchor the paper's discussion of bitrate-dependent transparency. The two examples below keep the comparison pool fixed while moving from a higher to a lower bitrate operating point.

Track 1 Clean Speech, 6 kbps

Higher-bitrate transparent coding example used to highlight differences in residual artifacts and anchor behavior under clean speech conditions.

Track: Track 1 Condition: Clean Speech Bitrate: 6 kbps

Reference

Baseline

aitd-go

Boya Audio

NanoCodec

NJU-AA Lab

Pdura7

TeamWZQAQ

Anchor

Jump to top

Track 1 Clean Speech, 1 kbps

Lower-bitrate clean-speech comparison intended to support discussion of bitrate pressure on speech naturalness and coding transparency.

Track: Track 1 Condition: Clean Speech Bitrate: 1 kbps

Reference

Baseline

aitd-go

Boya Audio

NanoCodec

NJU-AA Lab

Pdura7

TeamWZQAQ

Anchor

Jump to top

Jump to top

Figure 2

Robustness Under Mild Real-World Distortions

These examples support the paper's discussion of mild noisy and reverberant inputs. The pair contrasts lower- and higher-bitrate operation while preserving the same submission roster.

Track 1 Real-World Conditions, 1 kbps

Real-world input example for discussing how mild environmental distortions interact with aggressive compression settings.

Track: Track 1 Condition: Real-World Conditions Bitrate: 1 kbps

Noisy Reference

Baseline

aitd-go

Boya Audio

NanoCodec

NJU-AA Lab

Pdura7

TeamWZQAQ

Jump to top

Track 1 Real-World Conditions, 6 kbps

Higher-bitrate real-world example used to compare robustness trends against the lower-bitrate condition in the paper narrative.

Track: Track 1 Condition: Real-World Conditions Bitrate: 6 kbps

Noisy Reference

Baseline

aitd-go

Boya Audio

NanoCodec

NJU-AA Lab

Pdura7

TeamWZQAQ

Jump to top

Jump to top

Figure 3

Simultaneous Speaker Preservation

These examples support the paper's analysis of multi-talker preservation, including the special auxiliary comparison that appears in the 6 kbps condition.

Track 1 Overlapping Speakers, 6 kbps

Use this example to discuss how systems preserve overlapping content at the higher bitrate and to reference the auxiliary comparison clip.

Track: Track 1 Condition: Overlapping Speakers Bitrate: 6 kbps

Reference

aitd-go

Boya Audio

NanoCodec

NJU-AA Lab

Pdura7

TeamWZQAQ

Additional Comparison

Jump to top

Track 1 Overlapping Speakers, 1 kbps

Lower-bitrate simultaneous-speaker example used to contrast intelligibility preservation against the higher-bitrate condition.

Track: Track 1 Condition: Overlapping Speakers Bitrate: 1 kbps

Reference

Baseline

aitd-go

Boya Audio

NanoCodec

NJU-AA Lab

Pdura7

TeamWZQAQ

Jump to top

Jump to top

Figure 4

Word Intelligibility Focused Comparisons

This section mirrors the paper's intelligibility discussion and keeps the target and alternative word pair visible so readers can judge which lexical contrast survives each system.

Track 2 Word Intelligibility, Example 2

Word-pair example used to discuss lexical confusability in the enhancement-oriented track.

Track: Track 2 Task: Word Intelligibility Target: neck Alternative: deck

Clean Reference

Baseline

aitd-go

Boya Audio

NanoCodec

NJU-AA Lab

Pdura7

TeamWZQAQ

Jump to top

Track 2 Word Intelligibility, Example 4

Second word-pair example for the paper's intelligibility discussion, selected to provide a different lexical contrast from the earlier figure panel.

Track: Track 2 Task: Word Intelligibility Target: shoes Alternative: choose

Clean Reference

Baseline

aitd-go

Boya Audio

NanoCodec

NJU-AA Lab

Pdura7

TeamWZQAQ

Jump to top

Jump to top