Sadness under rescue rule S3, k=2. the literal 'not sad -> clearly sad' pair. 90.0 % of clips score at or below zero on this emotion and the largest gap on its normalised axis is 0.449 (WIDER than the 0.25 step cap). This rule found 7,826 chains over 40,000 tracks; the strict rule found 0 at k=3.
S3 changes: Absent to present. The chain must start at or below 0.05 (the emotion is absent) and end at or above 1.0 (it is clearly present). The most literal reading of 'from not sad to sad'. What it costs: Says nothing about the SHAPE of the path -- only that it begins absent and ends present. Any intermediate clip is unconstrained. Full explanation →Interest has a median of 2.08 and is never zero, while Infatuation is zero on 87.7 % of clips — and the caption named an emotion whenever its raw score cleared an absolute 1.0. Interest therefore appeared in 94.8 % of captions and Sadness in almost none: the clause was reporting the scale of the head, not the emotion of the clip. An emotion is now named only when it lands in the top 10 % for that emotion, against a pooled tie-aware mid-rank ECDF over 132,833,726 utterances spanning every dataset and language. Interest now appears in 6.2 %, all 40 emotions occur, and a clip that is ordinary on all 40 says “no dominant emotion” rather than being forced to pick one (21.8 % of clips). This is the same scale the trajectory miner selects on, so the caption and the mining now refer to the same quantity: the mined target emotion is named in the final clip's caption on 73 % of chains, up from 46 %.
This is not a strict-rule trajectory. It comes from rescue rule S3, which exists because the strict rule returns nothing at all for Sadness.
The raw scorer output across the chain is 0.00, 1.09. In the first clip the scorer found no Sadness whatsoever (0.00); by the last it is at 1.09.
On the corpus-wide percentile scale those become 0.43, 0.98 — a total move of +0.55.
That is why the strict rule cannot build this chain. Roughly 90 % of the corpus scores exactly zero on Sadness, and tied values all collapse onto one point (about 0.45). The clips that genuinely carry Sadness sit above 0.85. There is simply nothing in between to step onto, so the only move available is one jump far wider than the 0.25 per-step cap.
What this rule changes: Absent to present. The chain must start at or below 0.05 (the emotion is absent) and end at or above 1.0 (it is clearly present). The most literal reading of 'from not sad to sad'.
What it costs: Says nothing about the SHAPE of the path -- only that it begins absent and ends present. Any intermediate clip is unconstrained.
Same speaker? No similarity score is available here — the emolia clips in this chain are not covered by either speaker-embedding store. The chain therefore rests on the corpus's own speaker/track labelling, which is not the same as a measured check.
Voice consistency: these clips are separate recordings joined together, and no voice-similarity check could be run for this sample, so there is no measurement of how closely the voices match. You may hear the voice shift between segments. Voice conversion has not been applied yet in this build. A planned pass will re-render every segment onto the first segment's voice, which removes this effect entirely.
What was done to this chain. All 2 segments were re-synthesised with ChatterboxVC onto segment 1's voice — including segment 1 itself, converted with itself as the target — and then restored with SIDON. Converting the anchor too is what keeps the room and the reverb the same across the whole chain; leaving it raw put a change of acoustic at the first join. The words, timing and delivery still come from each original clip. The joins are 150–150 ms equal-power crossfades. The chain is normalised as one signal, so the loudness differences between segments are the ones the conversion produced, not a per-clip reset.
Did it unify the voice? On the 250-dimensional Speaker-wavLM-id verification embedding, the worst similarity between any segment and segment 1 was 0.716 before conversion and 0.698 after — it fell by 0.017. Neighbour-to-neighbour the worst pair went 0.716 → 0.698. (The earlier render, with segment 1 left raw, scores 0.585 here.) This chain started 0.70-0.80 — close, but under the identity threshold, a band where the conversion is close to a wash on this measure.
Did the emotion survive? Re-scored end to end through the same emotion stack and the same corpus-percentile scale the chain was mined on, Sadness moved +0.551 in the original and +0.520 after conversion — 94 % of the delta retained, which is essentially all of it.
Quality. Mean predicted overall quality across the segments went 2.83 → 2.98 (+0.15) on the Empathic-Insight head. The corrected render is 48 kHz because SIDON outputs 48 kHz; the original is the 24 kHz source. Some of what you hear as “cleaner” is that bandwidth, not the conversion.
This is not a strict-rule trajectory. It comes from rescue rule S3, which exists because the strict rule returns nothing at all for Sadness.
The raw scorer output across the chain is 0.00, 1.47. In the first clip the scorer found no Sadness whatsoever (0.00); by the last it is at 1.47.
On the corpus-wide percentile scale those become 0.44, 0.99 — a total move of +0.55.
That is why the strict rule cannot build this chain. Roughly 90 % of the corpus scores exactly zero on Sadness, and tied values all collapse onto one point (about 0.45). The clips that genuinely carry Sadness sit above 0.85. There is simply nothing in between to step onto, so the only move available is one jump far wider than the 0.25 per-step cap.
What this rule changes: Absent to present. The chain must start at or below 0.05 (the emotion is absent) and end at or above 1.0 (it is clearly present). The most literal reading of 'from not sad to sad'.
What it costs: Says nothing about the SHAPE of the path -- only that it begins absent and ends present. Any intermediate clip is unconstrained.
Same speaker? No similarity score is available here — the emolia clips in this chain are not covered by either speaker-embedding store. The chain therefore rests on the corpus's own speaker/track labelling, which is not the same as a measured check.
Voice consistency: these clips are separate recordings joined together, and no voice-similarity check could be run for this sample, so there is no measurement of how closely the voices match. You may hear the voice shift between segments. Voice conversion has not been applied yet in this build. A planned pass will re-render every segment onto the first segment's voice, which removes this effect entirely.
What was done to this chain. All 2 segments were re-synthesised with ChatterboxVC onto segment 1's voice — including segment 1 itself, converted with itself as the target — and then restored with SIDON. Converting the anchor too is what keeps the room and the reverb the same across the whole chain; leaving it raw put a change of acoustic at the first join. The words, timing and delivery still come from each original clip. The joins are 100–100 ms equal-power crossfades. The chain is normalised as one signal, so the loudness differences between segments are the ones the conversion produced, not a per-clip reset.
Did it unify the voice? On the 250-dimensional Speaker-wavLM-id verification embedding, the worst similarity between any segment and segment 1 was 0.304 before conversion and 0.459 after — it rose by 0.155. Neighbour-to-neighbour the worst pair went 0.304 → 0.459. (The earlier render, with segment 1 left raw, scores 0.369 here.) This chain started below 0.50 — the segments really were different people, the band the conversion helps most: chains starting below 0.50 improve on this measure almost without exception.
Did the emotion survive? Re-scored end to end through the same emotion stack and the same corpus-percentile scale the chain was mined on, Sadness moved +0.556 in the original and +0.529 after conversion — 95 % of the delta retained, which is essentially all of it.
Quality. Mean predicted overall quality across the segments went 2.68 → 3.14 (+0.46) on the Empathic-Insight head. The corrected render is 48 kHz because SIDON outputs 48 kHz; the original is the 24 kHz source. Some of what you hear as “cleaner” is that bandwidth, not the conversion.
This is not a strict-rule trajectory. It comes from rescue rule S3, which exists because the strict rule returns nothing at all for Sadness.
The raw scorer output across the chain is 0.00, 1.24. In the first clip the scorer found no Sadness whatsoever (0.00); by the last it is at 1.24.
On the corpus-wide percentile scale those become 0.44, 0.99 — a total move of +0.55.
That is why the strict rule cannot build this chain. Roughly 90 % of the corpus scores exactly zero on Sadness, and tied values all collapse onto one point (about 0.45). The clips that genuinely carry Sadness sit above 0.85. There is simply nothing in between to step onto, so the only move available is one jump far wider than the 0.25 per-step cap.
What this rule changes: Absent to present. The chain must start at or below 0.05 (the emotion is absent) and end at or above 1.0 (it is clearly present). The most literal reading of 'from not sad to sad'.
What it costs: Says nothing about the SHAPE of the path -- only that it begins absent and ends present. Any intermediate clip is unconstrained.
Same speaker? No similarity score is available here — the emolia clips in this chain are not covered by either speaker-embedding store. The chain therefore rests on the corpus's own speaker/track labelling, which is not the same as a measured check.
Voice consistency: these clips are separate recordings joined together, and no voice-similarity check could be run for this sample, so there is no measurement of how closely the voices match. You may hear the voice shift between segments. Voice conversion has not been applied yet in this build. A planned pass will re-render every segment onto the first segment's voice, which removes this effect entirely.
What was done to this chain. All 2 segments were re-synthesised with ChatterboxVC onto segment 1's voice — including segment 1 itself, converted with itself as the target — and then restored with SIDON. Converting the anchor too is what keeps the room and the reverb the same across the whole chain; leaving it raw put a change of acoustic at the first join. The words, timing and delivery still come from each original clip. The joins are 150–150 ms equal-power crossfades. The chain is normalised as one signal, so the loudness differences between segments are the ones the conversion produced, not a per-clip reset.
Did it unify the voice? On the 250-dimensional Speaker-wavLM-id verification embedding, the worst similarity between any segment and segment 1 was 0.741 before conversion and 0.775 after — it rose by 0.034. Neighbour-to-neighbour the worst pair went 0.741 → 0.775. (The earlier render, with segment 1 left raw, scores 0.534 here.) This chain started 0.70-0.80 — close, but under the identity threshold, a band where the conversion is close to a wash on this measure.
The emotional move did not survive. Re-scored end to end, Sadness moved +0.560 in the original and -0.012 after conversion — it changed direction. On this chain the corrected audio is not an improvement.
Quality. Mean predicted overall quality across the segments went 2.45 → 2.87 (+0.42) on the Empathic-Insight head. The corrected render is 48 kHz because SIDON outputs 48 kHz; the original is the 24 kHz source. Some of what you hear as “cleaner” is that bandwidth, not the conversion.
This is not a strict-rule trajectory. It comes from rescue rule S3, which exists because the strict rule returns nothing at all for Sadness.
The raw scorer output across the chain is -0.00, 1.02. In the first clip the scorer found no Sadness whatsoever (-0.00); by the last it is at 1.02.
On the corpus-wide percentile scale those become 0.43, 0.98 — a total move of +0.55.
That is why the strict rule cannot build this chain. Roughly 90 % of the corpus scores exactly zero on Sadness, and tied values all collapse onto one point (about 0.45). The clips that genuinely carry Sadness sit above 0.85. There is simply nothing in between to step onto, so the only move available is one jump far wider than the 0.25 per-step cap.
What this rule changes: Absent to present. The chain must start at or below 0.05 (the emotion is absent) and end at or above 1.0 (it is clearly present). The most literal reading of 'from not sad to sad'.
What it costs: Says nothing about the SHAPE of the path -- only that it begins absent and ends present. Any intermediate clip is unconstrained.
Same speaker? No similarity score is available here — the emolia clips in this chain are not covered by either speaker-embedding store. The chain therefore rests on the corpus's own speaker/track labelling, which is not the same as a measured check.
Voice consistency: these clips are separate recordings joined together, and no voice-similarity check could be run for this sample, so there is no measurement of how closely the voices match. You may hear the voice shift between segments. Voice conversion has not been applied yet in this build. A planned pass will re-render every segment onto the first segment's voice, which removes this effect entirely.
What was done to this chain. All 2 segments were re-synthesised with ChatterboxVC onto segment 1's voice — including segment 1 itself, converted with itself as the target — and then restored with SIDON. Converting the anchor too is what keeps the room and the reverb the same across the whole chain; leaving it raw put a change of acoustic at the first join. The words, timing and delivery still come from each original clip. The joins are 150–150 ms equal-power crossfades. The chain is normalised as one signal, so the loudness differences between segments are the ones the conversion produced, not a per-clip reset.
Did it unify the voice? On the 250-dimensional Speaker-wavLM-id verification embedding, the worst similarity between any segment and segment 1 was 0.695 before conversion and 0.685 after — it fell by 0.010. Neighbour-to-neighbour the worst pair went 0.695 → 0.685. (The earlier render, with segment 1 left raw, scores 0.562 here.) This chain started 0.50-0.70 — audibly different, a band where the conversion is close to a wash on this measure.
Did the emotion survive? Re-scored end to end through the same emotion stack and the same corpus-percentile scale the chain was mined on, Sadness moved +0.548 in the original and +0.526 after conversion — 96 % of the delta retained, which is essentially all of it.
Quality. Mean predicted overall quality across the segments went 2.98 → 3.03 (+0.05) on the Empathic-Insight head. The corrected render is 48 kHz because SIDON outputs 48 kHz; the original is the 24 kHz source. Some of what you hear as “cleaner” is that bandwidth, not the conversion.
This is not a strict-rule trajectory. It comes from rescue rule S3, which exists because the strict rule returns nothing at all for Sadness.
The raw scorer output across the chain is -0.00, 1.09. In the first clip the scorer found no Sadness whatsoever (-0.00); by the last it is at 1.09.
On the corpus-wide percentile scale those become 0.43, 0.98 — a total move of +0.55.
That is why the strict rule cannot build this chain. Roughly 90 % of the corpus scores exactly zero on Sadness, and tied values all collapse onto one point (about 0.45). The clips that genuinely carry Sadness sit above 0.85. There is simply nothing in between to step onto, so the only move available is one jump far wider than the 0.25 per-step cap.
What this rule changes: Absent to present. The chain must start at or below 0.05 (the emotion is absent) and end at or above 1.0 (it is clearly present). The most literal reading of 'from not sad to sad'.
What it costs: Says nothing about the SHAPE of the path -- only that it begins absent and ends present. Any intermediate clip is unconstrained.
Same speaker? No similarity score is available here — the emolia clips in this chain are not covered by either speaker-embedding store. The chain therefore rests on the corpus's own speaker/track labelling, which is not the same as a measured check.
Voice consistency: these clips are separate recordings joined together, and no voice-similarity check could be run for this sample, so there is no measurement of how closely the voices match. You may hear the voice shift between segments. Voice conversion has not been applied yet in this build. A planned pass will re-render every segment onto the first segment's voice, which removes this effect entirely.
What was done to this chain. All 2 segments were re-synthesised with ChatterboxVC onto segment 1's voice — including segment 1 itself, converted with itself as the target — and then restored with SIDON. Converting the anchor too is what keeps the room and the reverb the same across the whole chain; leaving it raw put a change of acoustic at the first join. The words, timing and delivery still come from each original clip. The joins are 100–100 ms equal-power crossfades. The chain is normalised as one signal, so the loudness differences between segments are the ones the conversion produced, not a per-clip reset.
Did it unify the voice? On the 250-dimensional Speaker-wavLM-id verification embedding, the worst similarity between any segment and segment 1 was 0.725 before conversion and 0.835 after — it rose by 0.109. Neighbour-to-neighbour the worst pair went 0.725 → 0.835. (The earlier render, with segment 1 left raw, scores 0.760 here.) This chain started 0.70-0.80 — close, but under the identity threshold, a band where the conversion is close to a wash on this measure.
Did the emotion survive? Re-scored end to end through the same emotion stack and the same corpus-percentile scale the chain was mined on, Sadness moved +0.552 in the original and +0.457 after conversion — 83 % of the delta retained, which is most of it.
Quality. Mean predicted overall quality across the segments went 3.07 → 3.24 (+0.16) on the Empathic-Insight head. The corrected render is 48 kHz because SIDON outputs 48 kHz; the original is the 24 kHz source. Some of what you hear as “cleaner” is that bandwidth, not the conversion.
This is not a strict-rule trajectory. It comes from rescue rule S3, which exists because the strict rule returns nothing at all for Sadness.
The raw scorer output across the chain is 0.04, 1.00. In the first clip the scorer found no Sadness whatsoever (0.04); by the last it is at 1.00.
On the corpus-wide percentile scale those become 0.87, 0.98 — a total move of +0.10.
That is why the strict rule cannot build this chain. Roughly 90 % of the corpus scores exactly zero on Sadness, and tied values all collapse onto one point (about 0.45). The clips that genuinely carry Sadness sit above 0.85. There is simply nothing in between to step onto, so the only move available is one jump far wider than the 0.25 per-step cap.
What this rule changes: Absent to present. The chain must start at or below 0.05 (the emotion is absent) and end at or above 1.0 (it is clearly present). The most literal reading of 'from not sad to sad'.
What it costs: Says nothing about the SHAPE of the path -- only that it begins absent and ends present. Any intermediate clip is unconstrained.
Same speaker? No similarity score is available here — the emolia clips in this chain are not covered by either speaker-embedding store. The chain therefore rests on the corpus's own speaker/track labelling, which is not the same as a measured check.
Voice consistency: these clips are separate recordings joined together, and no voice-similarity check could be run for this sample, so there is no measurement of how closely the voices match. You may hear the voice shift between segments. Voice conversion has not been applied yet in this build. A planned pass will re-render every segment onto the first segment's voice, which removes this effect entirely.
What was done to this chain. All 2 segments were re-synthesised with ChatterboxVC onto segment 1's voice — including segment 1 itself, converted with itself as the target — and then restored with SIDON. Converting the anchor too is what keeps the room and the reverb the same across the whole chain; leaving it raw put a change of acoustic at the first join. The words, timing and delivery still come from each original clip. The joins are 100–100 ms equal-power crossfades. The chain is normalised as one signal, so the loudness differences between segments are the ones the conversion produced, not a per-clip reset.
Did it unify the voice? On the 250-dimensional Speaker-wavLM-id verification embedding, the worst similarity between any segment and segment 1 was 0.880 before conversion and 0.909 after — it rose by 0.029. Neighbour-to-neighbour the worst pair went 0.880 → 0.909. (The earlier render, with segment 1 left raw, scores 0.728 here.) This chain started at or above 0.80 — already effectively one voice. Across the whole build that is the band where conversion tends to cost identity agreement rather than add it, and a hard identity cut would keep this chain without converting it at all. Judge it by ear against the original above.
Did the emotion survive? Re-scored end to end through the same emotion stack and the same corpus-percentile scale the chain was mined on, Sadness moved +0.104 in the original and +0.053 after conversion — 51 % of the delta retained.
Quality. Mean predicted overall quality across the segments went 2.84 → 3.25 (+0.42) on the Empathic-Insight head. The corrected render is 48 kHz because SIDON outputs 48 kHz; the original is the 24 kHz source. Some of what you hear as “cleaner” is that bandwidth, not the conversion.
This is not a strict-rule trajectory. It comes from rescue rule S3, which exists because the strict rule returns nothing at all for Sadness.
The raw scorer output across the chain is 0.00, 1.12. In the first clip the scorer found no Sadness whatsoever (0.00); by the last it is at 1.12.
On the corpus-wide percentile scale those become 0.44, 0.98 — a total move of +0.54.
That is why the strict rule cannot build this chain. Roughly 90 % of the corpus scores exactly zero on Sadness, and tied values all collapse onto one point (about 0.45). The clips that genuinely carry Sadness sit above 0.85. There is simply nothing in between to step onto, so the only move available is one jump far wider than the 0.25 per-step cap.
What this rule changes: Absent to present. The chain must start at or below 0.05 (the emotion is absent) and end at or above 1.0 (it is clearly present). The most literal reading of 'from not sad to sad'.
What it costs: Says nothing about the SHAPE of the path -- only that it begins absent and ends present. Any intermediate clip is unconstrained.
Same speaker? No similarity score is available here — the emolia clips in this chain are not covered by either speaker-embedding store. The chain therefore rests on the corpus's own speaker/track labelling, which is not the same as a measured check.
Voice consistency: these clips are separate recordings joined together, and no voice-similarity check could be run for this sample, so there is no measurement of how closely the voices match. You may hear the voice shift between segments. Voice conversion has not been applied yet in this build. A planned pass will re-render every segment onto the first segment's voice, which removes this effect entirely.
What was done to this chain. All 2 segments were re-synthesised with ChatterboxVC onto segment 1's voice — including segment 1 itself, converted with itself as the target — and then restored with SIDON. Converting the anchor too is what keeps the room and the reverb the same across the whole chain; leaving it raw put a change of acoustic at the first join. The words, timing and delivery still come from each original clip. The joins are 100–100 ms equal-power crossfades. The chain is normalised as one signal, so the loudness differences between segments are the ones the conversion produced, not a per-clip reset.
Did it unify the voice? On the 250-dimensional Speaker-wavLM-id verification embedding, the worst similarity between any segment and segment 1 was 0.802 before conversion and 0.729 after — it fell by 0.073. Neighbour-to-neighbour the worst pair went 0.802 → 0.729. (The earlier render, with segment 1 left raw, scores 0.493 here.) This chain started at or above 0.80 — already effectively one voice. Across the whole build that is the band where conversion tends to cost identity agreement rather than add it, and a hard identity cut would keep this chain without converting it at all. Judge it by ear against the original above.
Did the emotion survive? Re-scored end to end through the same emotion stack and the same corpus-percentile scale the chain was mined on, Sadness moved +0.556 in the original and +0.088 after conversion — 16 % of the delta retained, so a meaningful part of the trajectory was flattened.
Quality. Mean predicted overall quality across the segments went 2.46 → 2.81 (+0.35) on the Empathic-Insight head. The corrected render is 48 kHz because SIDON outputs 48 kHz; the original is the 24 kHz source. Some of what you hear as “cleaner” is that bandwidth, not the conversion.
This is not a strict-rule trajectory. It comes from rescue rule S3, which exists because the strict rule returns nothing at all for Sadness.
The raw scorer output across the chain is 0.00, 1.01. In the first clip the scorer found no Sadness whatsoever (0.00); by the last it is at 1.01.
On the corpus-wide percentile scale those become 0.43, 0.98 — a total move of +0.55.
That is why the strict rule cannot build this chain. Roughly 90 % of the corpus scores exactly zero on Sadness, and tied values all collapse onto one point (about 0.45). The clips that genuinely carry Sadness sit above 0.85. There is simply nothing in between to step onto, so the only move available is one jump far wider than the 0.25 per-step cap.
What this rule changes: Absent to present. The chain must start at or below 0.05 (the emotion is absent) and end at or above 1.0 (it is clearly present). The most literal reading of 'from not sad to sad'.
What it costs: Says nothing about the SHAPE of the path -- only that it begins absent and ends present. Any intermediate clip is unconstrained.
Same speaker? No similarity score is available here — the emolia clips in this chain are not covered by either speaker-embedding store. The chain therefore rests on the corpus's own speaker/track labelling, which is not the same as a measured check.
Voice consistency: these clips are separate recordings joined together, and no voice-similarity check could be run for this sample, so there is no measurement of how closely the voices match. You may hear the voice shift between segments. Voice conversion has not been applied yet in this build. A planned pass will re-render every segment onto the first segment's voice, which removes this effect entirely.
What was done to this chain. All 2 segments were re-synthesised with ChatterboxVC onto segment 1's voice — including segment 1 itself, converted with itself as the target — and then restored with SIDON. Converting the anchor too is what keeps the room and the reverb the same across the whole chain; leaving it raw put a change of acoustic at the first join. The words, timing and delivery still come from each original clip. The joins are 150–150 ms equal-power crossfades. The chain is normalised as one signal, so the loudness differences between segments are the ones the conversion produced, not a per-clip reset.
Did it unify the voice? On the 250-dimensional Speaker-wavLM-id verification embedding, the worst similarity between any segment and segment 1 was 0.913 before conversion and 0.915 after — it rose by 0.002. Neighbour-to-neighbour the worst pair went 0.913 → 0.915. (The earlier render, with segment 1 left raw, scores 0.888 here.) This chain started at or above 0.80 — already effectively one voice. Across the whole build that is the band where conversion tends to cost identity agreement rather than add it, and a hard identity cut would keep this chain without converting it at all. Judge it by ear against the original above.
Did the emotion survive? Re-scored end to end through the same emotion stack and the same corpus-percentile scale the chain was mined on, Sadness moved +0.547 in the original and +0.549 after conversion — 100 % of the delta retained, which is essentially all of it.
Quality. Mean predicted overall quality across the segments went 2.61 → 2.96 (+0.34) on the Empathic-Insight head. The corrected render is 48 kHz because SIDON outputs 48 kHz; the original is the 24 kHz source. Some of what you hear as “cleaner” is that bandwidth, not the conversion.
This is not a strict-rule trajectory. It comes from rescue rule S3, which exists because the strict rule returns nothing at all for Sadness.
The raw scorer output across the chain is -0.00, 1.30. In the first clip the scorer found no Sadness whatsoever (-0.00); by the last it is at 1.30.
On the corpus-wide percentile scale those become 0.43, 0.99 — a total move of +0.56.
That is why the strict rule cannot build this chain. Roughly 90 % of the corpus scores exactly zero on Sadness, and tied values all collapse onto one point (about 0.45). The clips that genuinely carry Sadness sit above 0.85. There is simply nothing in between to step onto, so the only move available is one jump far wider than the 0.25 per-step cap.
What this rule changes: Absent to present. The chain must start at or below 0.05 (the emotion is absent) and end at or above 1.0 (it is clearly present). The most literal reading of 'from not sad to sad'.
What it costs: Says nothing about the SHAPE of the path -- only that it begins absent and ends present. Any intermediate clip is unconstrained.
Same speaker? No similarity score is available here — the emolia clips in this chain are not covered by either speaker-embedding store. The chain therefore rests on the corpus's own speaker/track labelling, which is not the same as a measured check.
Voice consistency: these clips are separate recordings joined together, and no voice-similarity check could be run for this sample, so there is no measurement of how closely the voices match. You may hear the voice shift between segments. Voice conversion has not been applied yet in this build. A planned pass will re-render every segment onto the first segment's voice, which removes this effect entirely.
What was done to this chain. All 2 segments were re-synthesised with ChatterboxVC onto segment 1's voice — including segment 1 itself, converted with itself as the target — and then restored with SIDON. Converting the anchor too is what keeps the room and the reverb the same across the whole chain; leaving it raw put a change of acoustic at the first join. The words, timing and delivery still come from each original clip. The joins are 150–150 ms equal-power crossfades. The chain is normalised as one signal, so the loudness differences between segments are the ones the conversion produced, not a per-clip reset.
Did it unify the voice? On the 250-dimensional Speaker-wavLM-id verification embedding, the worst similarity between any segment and segment 1 was 0.805 before conversion and 0.801 after — it fell by 0.004. Neighbour-to-neighbour the worst pair went 0.805 → 0.801. (The earlier render, with segment 1 left raw, scores 0.701 here.) This chain started at or above 0.80 — already effectively one voice. Across the whole build that is the band where conversion tends to cost identity agreement rather than add it, and a hard identity cut would keep this chain without converting it at all. Judge it by ear against the original above.
The emotional move did not survive. Re-scored end to end, Sadness moved +0.560 in the original and -0.044 after conversion — it changed direction. On this chain the corrected audio is not an improvement.
Quality. Mean predicted overall quality across the segments went 2.69 → 3.07 (+0.38) on the Empathic-Insight head. The corrected render is 48 kHz because SIDON outputs 48 kHz; the original is the 24 kHz source. Some of what you hear as “cleaner” is that bandwidth, not the conversion.
This is not a strict-rule trajectory. It comes from rescue rule S3, which exists because the strict rule returns nothing at all for Sadness.
The raw scorer output across the chain is 0.00, 1.12. In the first clip the scorer found no Sadness whatsoever (0.00); by the last it is at 1.12.
On the corpus-wide percentile scale those become 0.43, 0.98 — a total move of +0.55.
That is why the strict rule cannot build this chain. Roughly 90 % of the corpus scores exactly zero on Sadness, and tied values all collapse onto one point (about 0.45). The clips that genuinely carry Sadness sit above 0.85. There is simply nothing in between to step onto, so the only move available is one jump far wider than the 0.25 per-step cap.
What this rule changes: Absent to present. The chain must start at or below 0.05 (the emotion is absent) and end at or above 1.0 (it is clearly present). The most literal reading of 'from not sad to sad'.
What it costs: Says nothing about the SHAPE of the path -- only that it begins absent and ends present. Any intermediate clip is unconstrained.
Same speaker? No similarity score is available here — the emolia clips in this chain are not covered by either speaker-embedding store. The chain therefore rests on the corpus's own speaker/track labelling, which is not the same as a measured check.
Voice consistency: these clips are separate recordings joined together, and no voice-similarity check could be run for this sample, so there is no measurement of how closely the voices match. You may hear the voice shift between segments. Voice conversion has not been applied yet in this build. A planned pass will re-render every segment onto the first segment's voice, which removes this effect entirely.
What was done to this chain. All 2 segments were re-synthesised with ChatterboxVC onto segment 1's voice — including segment 1 itself, converted with itself as the target — and then restored with SIDON. Converting the anchor too is what keeps the room and the reverb the same across the whole chain; leaving it raw put a change of acoustic at the first join. The words, timing and delivery still come from each original clip. The joins are 150–150 ms equal-power crossfades. The chain is normalised as one signal, so the loudness differences between segments are the ones the conversion produced, not a per-clip reset.
Did it unify the voice? On the 250-dimensional Speaker-wavLM-id verification embedding, the worst similarity between any segment and segment 1 was 0.780 before conversion and 0.773 after — it fell by 0.007. Neighbour-to-neighbour the worst pair went 0.780 → 0.773. (The earlier render, with segment 1 left raw, scores 0.596 here.) This chain started 0.70-0.80 — close, but under the identity threshold, a band where the conversion is close to a wash on this measure.
Did the emotion survive? Re-scored end to end through the same emotion stack and the same corpus-percentile scale the chain was mined on, Sadness moved +0.553 in the original and +0.000 after conversion — 0 % of the delta retained, so a meaningful part of the trajectory was flattened.
Quality. Mean predicted overall quality across the segments went 2.56 → 2.82 (+0.26) on the Empathic-Insight head. The corrected render is 48 kHz because SIDON outputs 48 kHz; the original is the 24 kHz source. Some of what you hear as “cleaner” is that bandwidth, not the conversion.
This is not a strict-rule trajectory. It comes from rescue rule S3, which exists because the strict rule returns nothing at all for Sadness.
The raw scorer output across the chain is -0.00, 1.67. In the first clip the scorer found no Sadness whatsoever (-0.00); by the last it is at 1.67.
On the corpus-wide percentile scale those become 0.43, 1.00 — a total move of +0.57.
That is why the strict rule cannot build this chain. Roughly 90 % of the corpus scores exactly zero on Sadness, and tied values all collapse onto one point (about 0.45). The clips that genuinely carry Sadness sit above 0.85. There is simply nothing in between to step onto, so the only move available is one jump far wider than the 0.25 per-step cap.
What this rule changes: Absent to present. The chain must start at or below 0.05 (the emotion is absent) and end at or above 1.0 (it is clearly present). The most literal reading of 'from not sad to sad'.
What it costs: Says nothing about the SHAPE of the path -- only that it begins absent and ends present. Any intermediate clip is unconstrained.
Same speaker? No similarity score is available here — the emolia clips in this chain are not covered by either speaker-embedding store. The chain therefore rests on the corpus's own speaker/track labelling, which is not the same as a measured check.
Voice consistency: these clips are separate recordings joined together, and no voice-similarity check could be run for this sample, so there is no measurement of how closely the voices match. You may hear the voice shift between segments. Voice conversion has not been applied yet in this build. A planned pass will re-render every segment onto the first segment's voice, which removes this effect entirely.
What was done to this chain. All 2 segments were re-synthesised with ChatterboxVC onto segment 1's voice — including segment 1 itself, converted with itself as the target — and then restored with SIDON. Converting the anchor too is what keeps the room and the reverb the same across the whole chain; leaving it raw put a change of acoustic at the first join. The words, timing and delivery still come from each original clip. The joins are 100–100 ms equal-power crossfades. The chain is normalised as one signal, so the loudness differences between segments are the ones the conversion produced, not a per-clip reset.
Did it unify the voice? On the 250-dimensional Speaker-wavLM-id verification embedding, the worst similarity between any segment and segment 1 was 0.884 before conversion and 0.849 after — it fell by 0.035. Neighbour-to-neighbour the worst pair went 0.884 → 0.849. (The earlier render, with segment 1 left raw, scores 0.792 here.) This chain started at or above 0.80 — already effectively one voice. Across the whole build that is the band where conversion tends to cost identity agreement rather than add it, and a hard identity cut would keep this chain without converting it at all. Judge it by ear against the original above.
Did the emotion survive? Re-scored end to end through the same emotion stack and the same corpus-percentile scale the chain was mined on, Sadness moved +0.567 in the original and +0.565 after conversion — 100 % of the delta retained, which is essentially all of it.
Quality. Mean predicted overall quality across the segments went 2.79 → 2.98 (+0.19) on the Empathic-Insight head. The corrected render is 48 kHz because SIDON outputs 48 kHz; the original is the 24 kHz source. Some of what you hear as “cleaner” is that bandwidth, not the conversion.
This is not a strict-rule trajectory. It comes from rescue rule S3, which exists because the strict rule returns nothing at all for Sadness.
The raw scorer output across the chain is 0.00, 1.02. In the first clip the scorer found no Sadness whatsoever (0.00); by the last it is at 1.02.
On the corpus-wide percentile scale those become 0.44, 0.98 — a total move of +0.54.
That is why the strict rule cannot build this chain. Roughly 90 % of the corpus scores exactly zero on Sadness, and tied values all collapse onto one point (about 0.45). The clips that genuinely carry Sadness sit above 0.85. There is simply nothing in between to step onto, so the only move available is one jump far wider than the 0.25 per-step cap.
What this rule changes: Absent to present. The chain must start at or below 0.05 (the emotion is absent) and end at or above 1.0 (it is clearly present). The most literal reading of 'from not sad to sad'.
What it costs: Says nothing about the SHAPE of the path -- only that it begins absent and ends present. Any intermediate clip is unconstrained.
Same speaker? No similarity score is available here — the emolia clips in this chain are not covered by either speaker-embedding store. The chain therefore rests on the corpus's own speaker/track labelling, which is not the same as a measured check.
Voice consistency: these clips are separate recordings joined together, and no voice-similarity check could be run for this sample, so there is no measurement of how closely the voices match. You may hear the voice shift between segments. Voice conversion has not been applied yet in this build. A planned pass will re-render every segment onto the first segment's voice, which removes this effect entirely.
What was done to this chain. All 2 segments were re-synthesised with ChatterboxVC onto segment 1's voice — including segment 1 itself, converted with itself as the target — and then restored with SIDON. Converting the anchor too is what keeps the room and the reverb the same across the whole chain; leaving it raw put a change of acoustic at the first join. The words, timing and delivery still come from each original clip. The joins are 150–150 ms equal-power crossfades. The chain is normalised as one signal, so the loudness differences between segments are the ones the conversion produced, not a per-clip reset.
Did it unify the voice? On the 250-dimensional Speaker-wavLM-id verification embedding, the worst similarity between any segment and segment 1 was 0.907 before conversion and 0.860 after — it fell by 0.048. Neighbour-to-neighbour the worst pair went 0.907 → 0.860. (The earlier render, with segment 1 left raw, scores 0.840 here.) This chain started at or above 0.80 — already effectively one voice. Across the whole build that is the band where conversion tends to cost identity agreement rather than add it, and a hard identity cut would keep this chain without converting it at all. Judge it by ear against the original above.
Did the emotion survive? Re-scored end to end through the same emotion stack and the same corpus-percentile scale the chain was mined on, Sadness moved +0.549 in the original and +0.464 after conversion — 84 % of the delta retained, which is most of it.
Quality. Mean predicted overall quality across the segments went 3.03 → 3.18 (+0.15) on the Empathic-Insight head. The corrected render is 48 kHz because SIDON outputs 48 kHz; the original is the 24 kHz source. Some of what you hear as “cleaner” is that bandwidth, not the conversion.
This is not a strict-rule trajectory. It comes from rescue rule S3, which exists because the strict rule returns nothing at all for Sadness.
The raw scorer output across the chain is 0.01, 1.21. In the first clip the scorer found no Sadness whatsoever (0.01); by the last it is at 1.21.
On the corpus-wide percentile scale those become 0.87, 0.99 — a total move of +0.12.
That is why the strict rule cannot build this chain. Roughly 90 % of the corpus scores exactly zero on Sadness, and tied values all collapse onto one point (about 0.45). The clips that genuinely carry Sadness sit above 0.85. There is simply nothing in between to step onto, so the only move available is one jump far wider than the 0.25 per-step cap.
What this rule changes: Absent to present. The chain must start at or below 0.05 (the emotion is absent) and end at or above 1.0 (it is clearly present). The most literal reading of 'from not sad to sad'.
What it costs: Says nothing about the SHAPE of the path -- only that it begins absent and ends present. Any intermediate clip is unconstrained.
Same speaker? No similarity score is available here — the emolia clips in this chain are not covered by either speaker-embedding store. The chain therefore rests on the corpus's own speaker/track labelling, which is not the same as a measured check.
Voice consistency: these clips are separate recordings joined together, and no voice-similarity check could be run for this sample, so there is no measurement of how closely the voices match. You may hear the voice shift between segments. Voice conversion has not been applied yet in this build. A planned pass will re-render every segment onto the first segment's voice, which removes this effect entirely.
What was done to this chain. All 2 segments were re-synthesised with ChatterboxVC onto segment 1's voice — including segment 1 itself, converted with itself as the target — and then restored with SIDON. Converting the anchor too is what keeps the room and the reverb the same across the whole chain; leaving it raw put a change of acoustic at the first join. The words, timing and delivery still come from each original clip. The joins are 100–100 ms equal-power crossfades. The chain is normalised as one signal, so the loudness differences between segments are the ones the conversion produced, not a per-clip reset.
Did it unify the voice? On the 250-dimensional Speaker-wavLM-id verification embedding, the worst similarity between any segment and segment 1 was 0.479 before conversion and 0.456 after — it fell by 0.023. Neighbour-to-neighbour the worst pair went 0.479 → 0.456. (The earlier render, with segment 1 left raw, scores 0.517 here.) This chain started below 0.50 — the segments really were different people, the band the conversion helps most: chains starting below 0.50 improve on this measure almost without exception.
The emotional move did not survive. Re-scored end to end, Sadness moved +0.122 in the original and -0.531 after conversion — it changed direction. On this chain the corrected audio is not an improvement.
Quality. Mean predicted overall quality across the segments went 2.63 → 2.86 (+0.22) on the Empathic-Insight head. The corrected render is 48 kHz because SIDON outputs 48 kHz; the original is the 24 kHz source. Some of what you hear as “cleaner” is that bandwidth, not the conversion.
This is not a strict-rule trajectory. It comes from rescue rule S3, which exists because the strict rule returns nothing at all for Sadness.
The raw scorer output across the chain is -0.00, 1.38. In the first clip the scorer found no Sadness whatsoever (-0.00); by the last it is at 1.38.
On the corpus-wide percentile scale those become 0.43, 0.99 — a total move of +0.56.
That is why the strict rule cannot build this chain. Roughly 90 % of the corpus scores exactly zero on Sadness, and tied values all collapse onto one point (about 0.45). The clips that genuinely carry Sadness sit above 0.85. There is simply nothing in between to step onto, so the only move available is one jump far wider than the 0.25 per-step cap.
What this rule changes: Absent to present. The chain must start at or below 0.05 (the emotion is absent) and end at or above 1.0 (it is clearly present). The most literal reading of 'from not sad to sad'.
What it costs: Says nothing about the SHAPE of the path -- only that it begins absent and ends present. Any intermediate clip is unconstrained.
Same speaker? No similarity score is available here — the emolia clips in this chain are not covered by either speaker-embedding store. The chain therefore rests on the corpus's own speaker/track labelling, which is not the same as a measured check.
Voice consistency: these clips are separate recordings joined together, and no voice-similarity check could be run for this sample, so there is no measurement of how closely the voices match. You may hear the voice shift between segments. Voice conversion has not been applied yet in this build. A planned pass will re-render every segment onto the first segment's voice, which removes this effect entirely.
What was done to this chain. All 2 segments were re-synthesised with ChatterboxVC onto segment 1's voice — including segment 1 itself, converted with itself as the target — and then restored with SIDON. Converting the anchor too is what keeps the room and the reverb the same across the whole chain; leaving it raw put a change of acoustic at the first join. The words, timing and delivery still come from each original clip. The joins are 100–100 ms equal-power crossfades. The chain is normalised as one signal, so the loudness differences between segments are the ones the conversion produced, not a per-clip reset.
Did it unify the voice? On the 250-dimensional Speaker-wavLM-id verification embedding, the worst similarity between any segment and segment 1 was 0.812 before conversion and 0.826 after — it rose by 0.014. Neighbour-to-neighbour the worst pair went 0.812 → 0.826. (The earlier render, with segment 1 left raw, scores 0.728 here.) This chain started at or above 0.80 — already effectively one voice. Across the whole build that is the band where conversion tends to cost identity agreement rather than add it, and a hard identity cut would keep this chain without converting it at all. Judge it by ear against the original above.
Did the emotion survive? Re-scored end to end through the same emotion stack and the same corpus-percentile scale the chain was mined on, Sadness moved +0.562 in the original and +0.527 after conversion — 94 % of the delta retained, which is essentially all of it.
Quality. Mean predicted overall quality across the segments went 2.92 → 3.11 (+0.19) on the Empathic-Insight head. The corrected render is 48 kHz because SIDON outputs 48 kHz; the original is the 24 kHz source. Some of what you hear as “cleaner” is that bandwidth, not the conversion.
This is not a strict-rule trajectory. It comes from rescue rule S3, which exists because the strict rule returns nothing at all for Sadness.
The raw scorer output across the chain is -0.00, 1.20. In the first clip the scorer found no Sadness whatsoever (-0.00); by the last it is at 1.20.
On the corpus-wide percentile scale those become 0.43, 0.99 — a total move of +0.56.
That is why the strict rule cannot build this chain. Roughly 90 % of the corpus scores exactly zero on Sadness, and tied values all collapse onto one point (about 0.45). The clips that genuinely carry Sadness sit above 0.85. There is simply nothing in between to step onto, so the only move available is one jump far wider than the 0.25 per-step cap.
What this rule changes: Absent to present. The chain must start at or below 0.05 (the emotion is absent) and end at or above 1.0 (it is clearly present). The most literal reading of 'from not sad to sad'.
What it costs: Says nothing about the SHAPE of the path -- only that it begins absent and ends present. Any intermediate clip is unconstrained.
Same speaker? No similarity score is available here — the emolia clips in this chain are not covered by either speaker-embedding store. The chain therefore rests on the corpus's own speaker/track labelling, which is not the same as a measured check.
Voice consistency: these clips are separate recordings joined together, and no voice-similarity check could be run for this sample, so there is no measurement of how closely the voices match. You may hear the voice shift between segments. Voice conversion has not been applied yet in this build. A planned pass will re-render every segment onto the first segment's voice, which removes this effect entirely.
What was done to this chain. All 2 segments were re-synthesised with ChatterboxVC onto segment 1's voice — including segment 1 itself, converted with itself as the target — and then restored with SIDON. Converting the anchor too is what keeps the room and the reverb the same across the whole chain; leaving it raw put a change of acoustic at the first join. The words, timing and delivery still come from each original clip. The joins are 150–150 ms equal-power crossfades. The chain is normalised as one signal, so the loudness differences between segments are the ones the conversion produced, not a per-clip reset.
Did it unify the voice? On the 250-dimensional Speaker-wavLM-id verification embedding, the worst similarity between any segment and segment 1 was 0.086 before conversion and 0.374 after — it rose by 0.288. Neighbour-to-neighbour the worst pair went 0.086 → 0.374. (The earlier render, with segment 1 left raw, scores 0.419 here.) This chain started below 0.50 — the segments really were different people, the band the conversion helps most: chains starting below 0.50 improve on this measure almost without exception.
Did the emotion survive? Re-scored end to end through the same emotion stack and the same corpus-percentile scale the chain was mined on, Sadness moved +0.557 in the original and +0.048 after conversion — 9 % of the delta retained, so a meaningful part of the trajectory was flattened.
Quality. Mean predicted overall quality across the segments went 2.56 → 2.77 (+0.21) on the Empathic-Insight head. The corrected render is 48 kHz because SIDON outputs 48 kHz; the original is the 24 kHz source. Some of what you hear as “cleaner” is that bandwidth, not the conversion.
This is not a strict-rule trajectory. It comes from rescue rule S3, which exists because the strict rule returns nothing at all for Sadness.
The raw scorer output across the chain is 0.00, 1.15. In the first clip the scorer found no Sadness whatsoever (0.00); by the last it is at 1.15.
On the corpus-wide percentile scale those become 0.43, 0.98 — a total move of +0.55.
That is why the strict rule cannot build this chain. Roughly 90 % of the corpus scores exactly zero on Sadness, and tied values all collapse onto one point (about 0.45). The clips that genuinely carry Sadness sit above 0.85. There is simply nothing in between to step onto, so the only move available is one jump far wider than the 0.25 per-step cap.
What this rule changes: Absent to present. The chain must start at or below 0.05 (the emotion is absent) and end at or above 1.0 (it is clearly present). The most literal reading of 'from not sad to sad'.
What it costs: Says nothing about the SHAPE of the path -- only that it begins absent and ends present. Any intermediate clip is unconstrained.
Same speaker? No similarity score is available here — the emolia clips in this chain are not covered by either speaker-embedding store. The chain therefore rests on the corpus's own speaker/track labelling, which is not the same as a measured check.
Voice consistency: these clips are separate recordings joined together, and no voice-similarity check could be run for this sample, so there is no measurement of how closely the voices match. You may hear the voice shift between segments. Voice conversion has not been applied yet in this build. A planned pass will re-render every segment onto the first segment's voice, which removes this effect entirely.
What was done to this chain. All 2 segments were re-synthesised with ChatterboxVC onto segment 1's voice — including segment 1 itself, converted with itself as the target — and then restored with SIDON. Converting the anchor too is what keeps the room and the reverb the same across the whole chain; leaving it raw put a change of acoustic at the first join. The words, timing and delivery still come from each original clip. The joins are 150–150 ms equal-power crossfades. The chain is normalised as one signal, so the loudness differences between segments are the ones the conversion produced, not a per-clip reset.
Did it unify the voice? On the 250-dimensional Speaker-wavLM-id verification embedding, the worst similarity between any segment and segment 1 was 0.316 before conversion and 0.489 after — it rose by 0.173. Neighbour-to-neighbour the worst pair went 0.316 → 0.489. (The earlier render, with segment 1 left raw, scores 0.452 here.) This chain started below 0.50 — the segments really were different people, the band the conversion helps most: chains starting below 0.50 improve on this measure almost without exception.
Did the emotion survive? Re-scored end to end through the same emotion stack and the same corpus-percentile scale the chain was mined on, Sadness moved +0.555 in the original and +0.535 after conversion — 96 % of the delta retained, which is essentially all of it.
Quality. Mean predicted overall quality across the segments went 2.86 → 3.05 (+0.19) on the Empathic-Insight head. The corrected render is 48 kHz because SIDON outputs 48 kHz; the original is the 24 kHz source. Some of what you hear as “cleaner” is that bandwidth, not the conversion.
This is not a strict-rule trajectory. It comes from rescue rule S3, which exists because the strict rule returns nothing at all for Sadness.
The raw scorer output across the chain is -0.00, 1.05. In the first clip the scorer found no Sadness whatsoever (-0.00); by the last it is at 1.05.
On the corpus-wide percentile scale those become 0.43, 0.98 — a total move of +0.55.
That is why the strict rule cannot build this chain. Roughly 90 % of the corpus scores exactly zero on Sadness, and tied values all collapse onto one point (about 0.45). The clips that genuinely carry Sadness sit above 0.85. There is simply nothing in between to step onto, so the only move available is one jump far wider than the 0.25 per-step cap.
What this rule changes: Absent to present. The chain must start at or below 0.05 (the emotion is absent) and end at or above 1.0 (it is clearly present). The most literal reading of 'from not sad to sad'.
What it costs: Says nothing about the SHAPE of the path -- only that it begins absent and ends present. Any intermediate clip is unconstrained.
Same speaker? No similarity score is available here — the emolia clips in this chain are not covered by either speaker-embedding store. The chain therefore rests on the corpus's own speaker/track labelling, which is not the same as a measured check.
Voice consistency: these clips are separate recordings joined together, and no voice-similarity check could be run for this sample, so there is no measurement of how closely the voices match. You may hear the voice shift between segments. Voice conversion has not been applied yet in this build. A planned pass will re-render every segment onto the first segment's voice, which removes this effect entirely.
What was done to this chain. All 2 segments were re-synthesised with ChatterboxVC onto segment 1's voice — including segment 1 itself, converted with itself as the target — and then restored with SIDON. Converting the anchor too is what keeps the room and the reverb the same across the whole chain; leaving it raw put a change of acoustic at the first join. The words, timing and delivery still come from each original clip. The joins are 100–100 ms equal-power crossfades. The chain is normalised as one signal, so the loudness differences between segments are the ones the conversion produced, not a per-clip reset.
Did it unify the voice? On the 250-dimensional Speaker-wavLM-id verification embedding, the worst similarity between any segment and segment 1 was 0.390 before conversion and 0.790 after — it rose by 0.400. Neighbour-to-neighbour the worst pair went 0.390 → 0.790. (The earlier render, with segment 1 left raw, scores 0.655 here.) This chain started below 0.50 — the segments really were different people, the band the conversion helps most: chains starting below 0.50 improve on this measure almost without exception.
Did the emotion survive? Re-scored end to end through the same emotion stack and the same corpus-percentile scale the chain was mined on, Sadness moved +0.549 in the original and +0.542 after conversion — 99 % of the delta retained, which is essentially all of it.
Quality. Mean predicted overall quality across the segments went 2.79 → 3.09 (+0.30) on the Empathic-Insight head. The corrected render is 48 kHz because SIDON outputs 48 kHz; the original is the 24 kHz source. Some of what you hear as “cleaner” is that bandwidth, not the conversion.
This is not a strict-rule trajectory. It comes from rescue rule S3, which exists because the strict rule returns nothing at all for Sadness.
The raw scorer output across the chain is 0.00, 1.85. In the first clip the scorer found no Sadness whatsoever (0.00); by the last it is at 1.85.
On the corpus-wide percentile scale those become 0.44, 1.00 — a total move of +0.56.
That is why the strict rule cannot build this chain. Roughly 90 % of the corpus scores exactly zero on Sadness, and tied values all collapse onto one point (about 0.45). The clips that genuinely carry Sadness sit above 0.85. There is simply nothing in between to step onto, so the only move available is one jump far wider than the 0.25 per-step cap.
What this rule changes: Absent to present. The chain must start at or below 0.05 (the emotion is absent) and end at or above 1.0 (it is clearly present). The most literal reading of 'from not sad to sad'.
What it costs: Says nothing about the SHAPE of the path -- only that it begins absent and ends present. Any intermediate clip is unconstrained.
Same speaker? No similarity score is available here — the emolia clips in this chain are not covered by either speaker-embedding store. The chain therefore rests on the corpus's own speaker/track labelling, which is not the same as a measured check.
Voice consistency: these clips are separate recordings joined together, and no voice-similarity check could be run for this sample, so there is no measurement of how closely the voices match. You may hear the voice shift between segments. Voice conversion has not been applied yet in this build. A planned pass will re-render every segment onto the first segment's voice, which removes this effect entirely.
What was done to this chain. All 2 segments were re-synthesised with ChatterboxVC onto segment 1's voice — including segment 1 itself, converted with itself as the target — and then restored with SIDON. Converting the anchor too is what keeps the room and the reverb the same across the whole chain; leaving it raw put a change of acoustic at the first join. The words, timing and delivery still come from each original clip. The joins are 150–150 ms equal-power crossfades. The chain is normalised as one signal, so the loudness differences between segments are the ones the conversion produced, not a per-clip reset.
Did it unify the voice? On the 250-dimensional Speaker-wavLM-id verification embedding, the worst similarity between any segment and segment 1 was 0.267 before conversion and 0.757 after — it rose by 0.490. Neighbour-to-neighbour the worst pair went 0.267 → 0.757. (The earlier render, with segment 1 left raw, scores 0.662 here.) This chain started below 0.50 — the segments really were different people, the band the conversion helps most: chains starting below 0.50 improve on this measure almost without exception.
Did the emotion survive? Re-scored end to end through the same emotion stack and the same corpus-percentile scale the chain was mined on, Sadness moved +0.568 in the original and +0.444 after conversion — 78 % of the delta retained, which is most of it.
Quality. Mean predicted overall quality across the segments went 2.88 → 3.05 (+0.18) on the Empathic-Insight head. The corrected render is 48 kHz because SIDON outputs 48 kHz; the original is the 24 kHz source. Some of what you hear as “cleaner” is that bandwidth, not the conversion.
This is not a strict-rule trajectory. It comes from rescue rule S3, which exists because the strict rule returns nothing at all for Sadness.
The raw scorer output across the chain is 0.00, 1.15. In the first clip the scorer found no Sadness whatsoever (0.00); by the last it is at 1.15.
On the corpus-wide percentile scale those become 0.43, 0.98 — a total move of +0.55.
That is why the strict rule cannot build this chain. Roughly 90 % of the corpus scores exactly zero on Sadness, and tied values all collapse onto one point (about 0.45). The clips that genuinely carry Sadness sit above 0.85. There is simply nothing in between to step onto, so the only move available is one jump far wider than the 0.25 per-step cap.
What this rule changes: Absent to present. The chain must start at or below 0.05 (the emotion is absent) and end at or above 1.0 (it is clearly present). The most literal reading of 'from not sad to sad'.
What it costs: Says nothing about the SHAPE of the path -- only that it begins absent and ends present. Any intermediate clip is unconstrained.
Same speaker? No similarity score is available here — the emolia clips in this chain are not covered by either speaker-embedding store. The chain therefore rests on the corpus's own speaker/track labelling, which is not the same as a measured check.
Voice consistency: these clips are separate recordings joined together, and no voice-similarity check could be run for this sample, so there is no measurement of how closely the voices match. You may hear the voice shift between segments. Voice conversion has not been applied yet in this build. A planned pass will re-render every segment onto the first segment's voice, which removes this effect entirely.
What was done to this chain. All 2 segments were re-synthesised with ChatterboxVC onto segment 1's voice — including segment 1 itself, converted with itself as the target — and then restored with SIDON. Converting the anchor too is what keeps the room and the reverb the same across the whole chain; leaving it raw put a change of acoustic at the first join. The words, timing and delivery still come from each original clip. The joins are 100–100 ms equal-power crossfades. The chain is normalised as one signal, so the loudness differences between segments are the ones the conversion produced, not a per-clip reset.
Did it unify the voice? On the 250-dimensional Speaker-wavLM-id verification embedding, the worst similarity between any segment and segment 1 was 0.712 before conversion and 0.734 after — it rose by 0.021. Neighbour-to-neighbour the worst pair went 0.712 → 0.734. (The earlier render, with segment 1 left raw, scores 0.721 here.) This chain started 0.70-0.80 — close, but under the identity threshold, a band where the conversion is close to a wash on this measure.
Did the emotion survive? Re-scored end to end through the same emotion stack and the same corpus-percentile scale the chain was mined on, Sadness moved +0.555 in the original and +0.550 after conversion — 99 % of the delta retained, which is essentially all of it.
Quality. Mean predicted overall quality across the segments went 2.87 → 3.11 (+0.24) on the Empathic-Insight head. The corrected render is 48 kHz because SIDON outputs 48 kHz; the original is the 24 kHz source. Some of what you hear as “cleaner” is that bandwidth, not the conversion.
This is not a strict-rule trajectory. It comes from rescue rule S3, which exists because the strict rule returns nothing at all for Sadness.
The raw scorer output across the chain is -0.00, 1.11. In the first clip the scorer found no Sadness whatsoever (-0.00); by the last it is at 1.11.
On the corpus-wide percentile scale those become 0.43, 0.98 — a total move of +0.55.
That is why the strict rule cannot build this chain. Roughly 90 % of the corpus scores exactly zero on Sadness, and tied values all collapse onto one point (about 0.45). The clips that genuinely carry Sadness sit above 0.85. There is simply nothing in between to step onto, so the only move available is one jump far wider than the 0.25 per-step cap.
What this rule changes: Absent to present. The chain must start at or below 0.05 (the emotion is absent) and end at or above 1.0 (it is clearly present). The most literal reading of 'from not sad to sad'.
What it costs: Says nothing about the SHAPE of the path -- only that it begins absent and ends present. Any intermediate clip is unconstrained.
Same speaker? No similarity score is available here — the emolia clips in this chain are not covered by either speaker-embedding store. The chain therefore rests on the corpus's own speaker/track labelling, which is not the same as a measured check.
Voice consistency: these clips are separate recordings joined together, and no voice-similarity check could be run for this sample, so there is no measurement of how closely the voices match. You may hear the voice shift between segments. Voice conversion has not been applied yet in this build. A planned pass will re-render every segment onto the first segment's voice, which removes this effect entirely.
What was done to this chain. All 2 segments were re-synthesised with ChatterboxVC onto segment 1's voice — including segment 1 itself, converted with itself as the target — and then restored with SIDON. Converting the anchor too is what keeps the room and the reverb the same across the whole chain; leaving it raw put a change of acoustic at the first join. The words, timing and delivery still come from each original clip. The joins are 100–100 ms equal-power crossfades. The chain is normalised as one signal, so the loudness differences between segments are the ones the conversion produced, not a per-clip reset.
Did it unify the voice? On the 250-dimensional Speaker-wavLM-id verification embedding, the worst similarity between any segment and segment 1 was 0.834 before conversion and 0.849 after — it rose by 0.015. Neighbour-to-neighbour the worst pair went 0.834 → 0.849. (The earlier render, with segment 1 left raw, scores 0.655 here.) This chain started at or above 0.80 — already effectively one voice. Across the whole build that is the band where conversion tends to cost identity agreement rather than add it, and a hard identity cut would keep this chain without converting it at all. Judge it by ear against the original above.
Did the emotion survive? Re-scored end to end through the same emotion stack and the same corpus-percentile scale the chain was mined on, Sadness moved +0.553 in the original and +0.104 after conversion — 19 % of the delta retained, so a meaningful part of the trajectory was flattened.
Quality. Mean predicted overall quality across the segments went 2.93 → 3.24 (+0.31) on the Empathic-Insight head. The corrected render is 48 kHz because SIDON outputs 48 kHz; the original is the 24 kHz source. Some of what you hear as “cleaner” is that bandwidth, not the conversion.