审阅范围 · Review scope
限定范围独立审阅。请保留正文中的出处与限制；不代表全库最终验收、实际演奏或学习效果验证。
Independent review within the stated scope. Keep the evidence and limitations stated in the text in view. This is not final corpus acceptance or validation of performance or learning outcomes.
先定拍，再找起音：分组、速度单位与跨拍保持
本包通过限定独立文字／静态来源、双语含义、初学者说明与符号审阅，未发现实质问题。T2是指定静帧支持的作者归属，G是本例限定术语，O是本库原创纸面任务与练习规划。来源逐项布尔值保留作者冻结时快照，publication_review记录当前通过范围；不表示来源听校、连续动作、完整来源谱例、演奏或学习效果验收。
先别跟着每一个音重新数“一”。一首曲子可以在拍子到来时继续保持前音，也可以暂时没有音。先写清拍子走到哪里，再标新音何时开始，才能把节奏问题拆成可核对的小任务。
本包读完NC-077／080全部307段既有ASR，并检查16个静帧。80段非教学、术语误识候选、重复或未能核实的口数／概括未用于归属，不是听校确认的80处错误。继承视频账本827.478秒；同哈希容器元数据827.434秒，差异原因未定。二者不是观看或听校时长。
两项任务都可完全用纸笔完成；不需要钢琴、唱歌、听力测试或录音。E4是中央C4右边第二个白键，88键从A0=1并计黑键时为第44键。这里不指定手指、实际配手或身体动作。
G：拍是选定的规律时间单位；细分是把一拍继续分小；小节把若干拍组成一组。起音是新音开始，保持是该音继续，释放是本模型不再保持该音。拍号、拍单位与速度要一起读，数字60单独不能确定音符的秒数。
G：本包只用常见规则读法：3/4以四分音符为拍，6/8以附点四分音符为拍并把每拍分成三个八分音符。附点让时值增加原来的一半，因此附点四分等于三个八分。实际作品的分组可因速度、写法和语境而不同；不能只看上方数字自动定论。
G：拍子的层次不等于每个强拍都必须弹得更响。切分涉及起音、时长或其他强调与既定拍子层次的关系，不等于拍子消失或一定弹错；本包不把任何单个弱位置的音都自动命名为切分。
有边界的来源归属
B38-S01 · T2
Wiwi将拍的分组和细分联系起来，画面并列简单二／三／四拍与复合二／三／四拍，并以6/8、9/8等展示三个八分一组。我们借此提醒先辨拍单位，不把入门总结推广到所有拍号或所有演奏语境。
NC-077静帧211秒术语、375秒四分音符、452／468／481／498秒拍号及分组；只核标题和选定节奏结构。ASR第180–188段作定位，不作为已听到分组感的证据。
B38-S02 · T2
Wiwi用改变较长音与和声更换位置的例子说明切分，选定画面显示较长音移位及跨拍延音线。这里保留位置与时长的区别，不把作者对冲突感／趣味的描述当作每个人必然的听觉结果。
NC-080静帧44秒标题、113／125秒长音位置对照、147／163秒改变和声时机的选定结构。90秒仍保留原强拍低音，不能用邻近ASR证明该帧已删音；完整谱例、原声未验。
拿到乐谱后的节奏工作单
O：先记录拍号、明确的拍单位和目标速度；再画小节与拍的格线，最后在格线上分别放起音、保持、释放和休止。延音线跨到下一拍时，不再加一次起音。遇到实际版本不清楚的连线、拍号或速度，先标待确认，再向清晰版本或老师核实。
O：反馈先问错在单位、分组、起音位置、保持时长，还是多弹了一次延音。每次只改一项再核原谱。计时计算只能发现表格不一致；不能证明演奏者实际稳定、放松或已经听懂节奏。
O：先核一个小节，再把它和前后一小节连起来，保持原来的拍单位与格线。一次成功后，休息再核，次日可重做；这是已知材料的保持检查。真正的新谱迁移和实际演奏还需另验，本包没有证明练习效果。
六个八分音符，不一定有同样的分组或秒数
纸面任务：每个版本只有一个小节、六个连续E4八分音符，无伴奏、无踏板。表中的格号0–6是八分音符格的位置，不是主拍编号。相邻事件释放与起音在同一格界；这只规定书面时值，不规定实际连奏技术。先记三版速度：M34四分=60；M68-EQUAL附点四分=40；M68-60附点四分=60。
先只看“题目表”和共同事件表，遮住“答案表”。算每格、每主拍和整小节各几秒，并在六格上标主拍起点。不要把“同为60 BPM”当成同一个音符单位；所有答案随后公开，只是引导练习。
题目表
| ID | 拍号 | 拍单位 | BPM | 主拍数 | 每拍格数 |
|---|---|---|---|---|---|
| M34 | 3/4 | 四分音符 | 60 | 3 | 2 |
| M68-EQUAL | 6/8 | 附点四分音符 | 40 | 2 | 3 |
| M68-60 | 6/8 | 附点四分音符 | 60 | 2 | 3 |
共同事件表——八分音符格
| ID | 音高／键号 | 起始格 | 释放格 |
|---|---|---|---|
| M1 | E4 / 44 | 0 | 1 |
| M2 | E4 / 44 | 1 | 2 |
| M3 | E4 / 44 | 2 | 3 |
| M4 | E4 / 44 | 3 | 4 |
| M5 | E4 / 44 | 4 | 5 |
| M6 | E4 / 44 | 5 | 6 |
每小节秒数=主拍数×60÷BPM；每格秒数=每主拍秒数÷每拍格数。这里主拍数分别为3、2、2，每拍格数为2、3、3。相同总时长不代表相同拍感，相同BPM数字也不保证同一小节时长。
答案表
| ID | 秒／格 | 秒／拍 | 秒／小节 | 主拍起始格 |
|---|---|---|---|---|
| M34 | 1/2 | 1 | 3 | 0,2,4 |
| M68-EQUAL | 1/2 | 3/2 | 3 | 0,3 |
| M68-60 | 1/3 | 1 | 2 | 0,3 |
M34: 1 + | 2 + | 3 +
M68-EQUAL: 1 + a | 2 + a
M68-60: 1 + a | 2 + a
M34为每格1/2秒、每拍1秒、每小节3秒，分成2+2+2；M68-EQUAL也是每格1/2秒、每小节3秒，但每主拍3/2秒，分成3+3；M68-60为每格1/3秒、每拍1秒、每小节2秒，仍分成3+3。数字1／2／3标主拍起点，+与a只是细分标签，不是重音指令。
即时核对：每版都恰有六格，主拍起点分别为0／2／4或0／3。若写6/8却把六格全叫六个主拍，回到本题已指定的附点四分拍单位；慢速数细分可作为另一层计数，但不能悄悄替换本题的BPM含义。
跨过下一拍，为什么不重新起音？
纸面原创例：两小节4/4，四分=60 BPM，共8拍=8秒。所有音是E4；无伴奏、无踏板、未指定配手。起点包含在保持时段内，释放点不再保持。offset从0起，主拍位置为0到7，第二小节从4起。S4以同音延音线把3.5–4与4–4.5两个八分连接，合成一个3.5–4.5的保持事件。
先看事件表与S4的两段写法，遮住答案和时间条。在4、4.5、5三个位置，分别判断有没有新起音、有没有保持的音、拍子的格线是否继续。延音线不是连奏线：本例连接相同音高并合并时值，不产生第二次起音。
| ID | 音高／键号 | 起点 | 释放 |
|---|---|---|---|
| S1 | E4 / 44 | 0 | 0.5 |
| S2 | E4 / 44 | 1 | 1.5 |
| S3 | E4 / 44 | 2 | 2.5 |
| S4 | E4 / 44 | 3.5 | 4.5 |
| S5 | E4 / 44 | 5 | 5.5 |
| S6 | E4 / 44 | 6 | 6.5 |
| S7 | E4 / 44 | 7 | 7.5 |
S4a: E4 [3.5,4) → S4b: E4 [4,4.5)
在4：进入第二小节的主拍，S4仍保持，没有新起音。在4.5：S4释放且没有别的保持音，直到5才由S5重新起音；4.5是细分位置，格线没有停止。八个主拍格点不等于八次起音：本例七个事件、八个书面音符段，其中S4占两个同音相连段。
A . A . A . . A | H . A . A . A .
每格为0.5拍。A表示该格开始时新起音且保持，H表示从前格继续保持，.表示没有保持的旋律音。S4跨过小节线；小节线不要求释放或重弹。这里“没有保持音”不等于声学上绝对安静。
即时核对：总保持4拍、空档4拍；七个起点为0、1、2、3.5、5、6、7，没有4。若在4又记一次起音，就把延音误改成重弹；若把4.5–5填成保持，则改变了释放时间。只有在本例既定4/4层次下，3.5起音跨到4才作为切分示意；未声称听感或学习效果已验证。
范围与相关阅读
本包没有新增原声听校、连续动作、完整来源谱例、身体适用性或学习者验证。NC-077对钟声、知觉与比例的说明、NC-080的趣味评价不作为普遍科学结论；来源全部旋律、和弦配置及演奏未逐音验收。原始转录和静帧不在候选包内，历史全局转写快照也未修改。
本包来源覆盖
分组与延音的书面检查
Set the beat, then locate attacks: grouping, tempo units and holds across beats
This package passed bounded independent written/static-source, bilingual-meaning, beginner-instruction and symbolic review with no material findings. T2 marks attribution supported by specified stills, G scoped terms, and O original paper tasks and practice planning. Per-source booleans retain their author-time snapshot; publication_review records the current passed scope, not source audition, continuous motion, complete source scores, performance or learning-efficacy acceptance.
Do not restart the count at one for every note. A piece can continue holding an earlier note when a beat arrives, or contain a gap. First track where the beat is, then mark when each new note begins, so a rhythm problem becomes a checkable task.
This package reads all 307 existing ASR segments of NC-077/080 and inspects 16 stills. 80 noninstructional, suspected terminology, repeated, or unverified counting/generalization candidates are excluded from attribution, not 80 audition-confirmed errors. Inherited video-inventory duration is 827.478 seconds; same-hash container metadata is 827.434 seconds, with the difference unexplained. Neither measures viewing or audition time.
Both tasks can be completed entirely with paper and pencil; no piano, singing, hearing test or recording is required. E4 is the second white key to the right of middle C4, key 44 when all 88 black and white keys are counted from A0=1. No fingering, actual hand assignment or bodily motion is prescribed.
G: a beat is the chosen regular time unit; subdivision divides it further; a bar groups beats. An attack starts a new note, holding continues it, and release ends that holding in this model. Read meter, beat unit and tempo together: the number 60 alone does not determine note durations in seconds.
G: this package uses common regular readings: 3/4 with a quarter-note beat, and 6/8 with a dotted-quarter beat divided into three eighth notes. A dot adds half the original value, so a dotted quarter equals three eighths. Actual grouping can vary with tempo, notation and context; the upper number alone is not a universal decision rule.
G: metrical hierarchy does not require every strong beat to be played louder. Syncopation concerns attacks, durations or other emphasis in relation to an established meter; it does not mean that the beat disappears or that playing is necessarily wrong. This package does not automatically label every weak-position note syncopated.
Bounded source attributions
B38-S01 · T2
Wiwi connects beat grouping with subdivision, contrasting simple and compound duple/triple/quadruple meters and showing groups of three eighths in 6/8 and 9/8. We use this to prompt explicit beat-unit identification, without extending the introductory summary to every meter or performance context.
NC-077 stills: terminology at 211 seconds, quarter note at 375, signatures/grouping at 452/468/481/498. Only headings and selected rhythmic structure were checked. ASR segments 180–188 locate context, not evidence of hearing a grouping.
B38-S02 · T2
Wiwi illustrates syncopation by changing the placement of longer notes and harmonic changes; selected frames show shifted longer values and ties across beats. We retain the distinction between position and duration without treating the author’s descriptions of conflict or interest as inevitable perceptual results.
NC-080 stills: title at 44 seconds, longer-value comparison at 113/125, and selected harmonic-timing structure at 147/163. The 90-second frame still retains the original strong-beat bass; nearby ASR cannot prove an omission in that frame. Complete source scores and original audio remain unverified.
A rhythm work sheet for a new score
O: first record the signature, explicit beat unit and target tempo. Draw bars and beats, then place attacks, holds, releases and rests separately on that grid. A tie continuing into the next beat adds no new attack. Mark unclear ties, signatures or tempo in the actual edition as unresolved and check a clearer edition or teacher.
O: identify whether the mismatch concerns the unit, grouping, attack location, hold length or an extra attack inside a tie. Change one item at a time and compare with the score again. Timing arithmetic detects inconsistencies in a table; it does not establish actual steadiness, relaxation or rhythmic understanding.
O: check one bar, then join it to the preceding and following bars while preserving the beat unit and grid. After one success, recheck following a break and optionally the next day: this tests retention of known material. New-score transfer and actual performance need separate evaluation; this package does not establish learning efficacy.
Six eighth notes need not share grouping or elapsed time
Paper task: each version is one bar of six consecutive E4 eighth notes, with no accompaniment or pedal. Grid positions 0–6 count eighth-note slots, not main beats. Adjacent releases and attacks meet at a boundary; this specifies written duration, not physical legato technique. First record all tempos: M34 quarter=60; M68-EQUAL dotted quarter=40; M68-60 dotted quarter=60.
Use only the prompt table and shared event table first, covering the answer table. Calculate seconds per slot, main beat and bar, and mark main-beat starts on six slots. Do not treat equal BPM numbers as equal note units. Answers are disclosed below; this is guided practice.
Prompt table
| ID | Meter | Beat unit | BPM | Main beats | Slots per beat |
|---|---|---|---|---|---|
| M34 | 3/4 | quarter | 60 | 3 | 2 |
| M68-EQUAL | 6/8 | dotted quarter | 40 | 2 | 3 |
| M68-60 | 6/8 | dotted quarter | 60 | 2 | 3 |
Shared event table — eighth-note slots
| ID | Pitch / key | Start slot | Release slot |
|---|---|---|---|
| M1 | E4 / 44 | 0 | 1 |
| M2 | E4 / 44 | 1 | 2 |
| M3 | E4 / 44 | 2 | 3 |
| M4 | E4 / 44 | 3 | 4 |
| M5 | E4 / 44 | 4 | 5 |
| M6 | E4 / 44 | 5 | 6 |
Bar seconds = main beats × 60 ÷ BPM; slot seconds = seconds per main beat ÷ slots per beat. Here the main-beat counts are 3,2,2 and slots per beat are 2,3,3. Equal elapsed time does not establish equal metrical feel; equal BPM numbers do not guarantee equal bar duration.
Answer table
| ID | Seconds/slot | Seconds/beat | Seconds/bar | Main-beat start slots |
|---|---|---|---|---|
| M34 | 1/2 | 1 | 3 | 0,2,4 |
| M68-EQUAL | 1/2 | 3/2 | 3 | 0,3 |
| M68-60 | 1/3 | 1 | 2 | 0,3 |
M34: 1 + | 2 + | 3 +
M68-EQUAL: 1 + a | 2 + a
M68-60: 1 + a | 2 + a
M34 has 1/2 second per slot, 1 second per beat and 3 seconds per bar, grouped 2+2+2. M68-EQUAL also has 1/2 second per slot and 3 seconds per bar, but 3/2 seconds per main beat, grouped 3+3. M68-60 has 1/3 second per slot, 1 second per beat and 2 seconds per bar, still grouped 3+3. Numerals 1/2/3 mark main-beat starts; + and a are subdivision labels, not accent instructions.
Immediate check: each version has exactly six slots; main-beat starts are 0/2/4 or 0/3. If all six slots in 6/8 are called main beats, return to the dotted-quarter beat assigned here. Counting subdivisions slowly is a separate counting layer; it must not silently replace the stated meaning of BPM.
Why does crossing the next beat add no new attack?
Original paper example: two bars of 4/4 at quarter=60 BPM, totaling 8 beats=8 seconds. Every note is E4; no accompaniment, pedal or hand assignment. An onset is included in a hold; its release is excluded. Offsets begin at 0, main-beat positions are 0 through 7, and bar 2 begins at 4. S4 joins the eighth-note spans 3.5–4 and 4–4.5 with a same-pitch tie, creating one held event from 3.5 to 4.5.
Inspect the event table and the two written pieces of S4 before uncovering the answer and timeline. At positions 4,4.5,5, decide whether a new attack occurs, whether a note is held, and whether the beat grid continues. A tie is not a slur: here it joins equal pitches and combines durations without a second attack.
| ID | Pitch / key | Start | Release |
|---|---|---|---|
| S1 | E4 / 44 | 0 | 0.5 |
| S2 | E4 / 44 | 1 | 1.5 |
| S3 | E4 / 44 | 2 | 2.5 |
| S4 | E4 / 44 | 3.5 | 4.5 |
| S5 | E4 / 44 | 5 | 5.5 |
| S6 | E4 / 44 | 6 | 6.5 |
| S7 | E4 / 44 | 7 | 7.5 |
S4a: E4 [3.5,4) → S4b: E4 [4,4.5)
At 4, bar 2 begins on a main beat while S4 is held without a new attack. At 4.5, S4 releases and no other note is held until S5 attacks at 5; 4.5 is a subdivision position, and the grid has not stopped. Eight main-beat points do not imply eight attacks: there are seven events and eight written note pieces, with S4 occupying two tied same-pitch pieces.
A . A . A . . A | H . A . A . A .
Each cell is 0.5 beat. A starts and holds a new note at the cell boundary, H continues a previous hold, and . has no held melody note. S4 crosses the barline; a barline requires neither release nor reattack. No held note does not imply absolute acoustic silence.
Immediate check: holds total 4 beats and gaps total 4. The seven starts are 0,1,2,3.5,5,6,7, excluding 4. Adding an attack at 4 turns the tie into a reattack; filling 4.5–5 changes the release. Only within the assigned 4/4 hierarchy is the attack at 3.5 continuing through 4 used as a syncopation illustration; perceived effect and learning efficacy are untested.
Scope and related reading
This package adds no original-audio audition, continuous-motion review, complete source-score acceptance, physical-suitability or learner validation. NC-077’s clock, perception and prevalence explanations and NC-080’s interest judgments are not adopted as universal scientific findings. Complete source melodies, voicings and performances were not checked note by note. Raw transcripts and stills are excluded from the candidate; the historical global transcription snapshot is unchanged.
Separate tempo, density and holding
Source coverage for this package
Written grouping and tie checks