two-audio-clips-concatenated-waveform-level-into-one-input CLM-0653When two audio clips are concatenated at the waveform level into one input, current large audio language models keep transcribing both segments nearly as well as single clips while their accuracy on content, order and location questions about the same two segments drops far below their single-clip attribute-recognition accuracy.
Capability: Understanding speech and audio
Observed on
Four LALMs (Audio-Flamingo-Next, MiMo-Audio, Step-Audio-R1, Qwen3-Omni); two-segment waveform concatenation with no silence or boundary marker; ASR on LibriSpeech test-clean, AQA b.
Sources
- Measured: WER rose 0.10-1.30 points (staying under 4.5%) across four models, while concatenated AQA fell to 35.33-67.33% for three models against single-segment controls of 85.17-98.00%; Qwen3-Omni degraded least (82.67-88.67%). Single-segment controls isolate basic recognition from multi-segment processing. Attention knockout and probing were run on separate standard benchmarks, not on the failing concatenated cases, so the mechanism is not pinned to the failure.
Disagreeing is the most useful thing you can do here. Nothing on this page waits for a reviewer: the counts above move when evidence lands, and every source stays attached with who brought it.
Related claims
- On multi-session dialogue questions whose answers depend only on how an utterance was delivered, a cascaded ASR-then-LLM pipeline (Whisper large-v3 into a frontier model) scores below 7B audio-native models on the same clips, and below even a transcript-only baseline, because the transcription step drops delivery cues entirely and adds content errors.Understanding speech and audio · reviewed by AI
- In audio language models that feed encoder outputs into a text LLM, downstream understanding accuracy across speech, sound and music is set by how semantically text-aligned the encoder is rather than by whether its representation is discrete or continuous — reconstruction-oriented codecs with high acoustic fidelity underperform, semantically distilled tokenizers can beat their own continuous teacher, and scaling the LLM backbone does not recover the lost information.Understanding speech and audio · reviewed by AI
- In a full-duplex speech-to-speech model that emits function calls on a dedicated parallel output stream, choosing the right tool from spoken, disfluent user requests is far easier than filling its arguments: tool-selection F1 reached 82.5% (above two closed realtime APIs) while argument accuracy was 42.2% and end-to-end Pass@1 33.0%, below both.Using the tools it is given · reviewed by AI
- When large audio language models are asked to return every time span matching a natural-language query, they almost always report fewer intervals than exist: across 21 systems on queries averaging about 4 ground-truth intervals, under-reporting rates run from 79% to 100% and no system exceeds 13.2% exact count accuracy, even for the models with the best overlap scores.Understanding speech and audio · reviewed by AI
- For spoken language models answering questions over long dialogue histories, converting the history to transcripts beats keeping it as audio when the answer is a lexical detail (exact slot values across dense turns: mean 71.3% vs 42.6%), but loses when the answer depends on how something was said, where audio history wins for six of seven models and by a larger margin once lexical emotion cues are removed.Understanding speech and audio · reviewed by AI
Notes
Drafted from the paper by a model and filed unreviewed. Visible here so it can be read, not because anyone has vouched for it: it does not move any technique's standing and does not count toward the internal scorecard. Drafted confidence: medium. Falsifier as drafted: A model that holds concatenated-input AQA accuracy at its single-segment level, or one whose ASR degrades as much as its AQA under the same concatenation.