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.
本页保留作者送审冻结字节，因此正文或数据可能仍有“候选／等待审阅／未发布”标记。当前版本已通过限定独立复核并发布；旧标记只记录冻结时状态。通过范围是双语含义、限定来源文字／静帧及原创固定例题检查，不含原声听校、连续动作、完整源谱、实际演奏、身体适用性或学习效果验证。
This page preserves the author’s frozen review bytes, so prose or data may still contain candidate, pending-review or unpublished labels. This version has passed bounded independent review and repository publication; those labels record author-time status only. Acceptance covers bilingual meaning, bounded source text/stills and original fixed-fixture checks, not source audition, continuous motion, full source scores, actual performance, physical suitability or learning outcomes.
跨过延音线，继续数小节
小节换了，音一定重弹吗？
试想这样一个读谱问题：你正在数3/4拍的“1、2、3”，下一小节的“1”到了，但同一个音有延音线连过来。拍子要继续，手却不应因此多弹一次。本课用一条原创四小节单音旋律，把拍点、小节开头与起音点分开核对。
先会辨认四分、二分、附点二分音符、四分休止和字母音名；需要时先看文末的节奏与读谱链接。C4是中央C，音名后的4是八度，不是指法。可以全程用纸笔或说数完成，钢琴只是可选路线。
先不看答案：第一小节写D4二分音符、E4四分音符；第二小节写E4四分音符、F#4四分音符、四分休止；两个相邻E4之间有延音线。第二小节第1拍要重新弹E4吗？两个E4共占几个四分单位？先写预测或“不确定”。
本例的延音线连接同一声部、同一音高、时间连续的两个E4，把它们合成一次持续2q的音符事件，只在前一个E4起点起音。这里q表示四分音符单位。延音线与用于乐句或奏法、也可连接不同音高的连线不同，也不等于使用右踏板。
两种计数，共用一条时间轴
四小节都保持3/4，每小节3q。另加一行本课指定的标注，把这四小节依次记为大跨度中的第1、2、3、4小节；这只是一种明确给定的阅读组织，不是在证明任意作品都按四小节分句。小节内计数仍是1、2、3，没有改成4/4。
全局q从0开始，小节拍次从1开始。四个小节开头在q=0、3、6、9，全段在q=12结束。四小节跨度包含12个均匀q单位；不能把两行计数相加成16个单位，也不能把大跨度的一次编号误当成一个四分音符。
表中的“起音”指本例要求开始一个新的单音事件，“持续”表示前音尚占谱写时值，“休止”占有自己的时间。第二小节第1拍q=3是E4延续，第三小节第1拍q=6则是G4新起音；同样是小节开头，事件类型可以不同。
大跨度、小节、拍的层级标注帮助定位，不自动要求重弹、加速或更用力。本文没有给音量、重音或乐句强弱分配一套固定公式，也没有从这张时间表推断你真正听到的节拍感。
| 全局q | 小节:拍 | 跨度小节号 | 音 | 事件状态 |
|---|---|---|---|---|
| 0 | 1:1 | 1 | D4 | 起音 |
| 1 | 1:2 | 1 | D4 | 持续 |
| 2 | 1:3 | 1 | E4 | 起音 |
| 3 | 2:1 | 2 | E4 | 持续 |
| 4 | 2:2 | 2 | F#4 | 起音 |
| 5 | 2:3 | 2 | — | 休止 |
| 6 | 3:1 | 3 | G4 | 起音 |
| 7 | 3:2 | 3 | G4 | 持续 |
| 8 | 3:3 | 3 | G4 | 持续 |
| 9 | 4:1 | 4 | F#4 | 起音 |
| 10 | 4:2 | 4 | E4 | 起音 |
| 11 | 4:3 | 4 | D4 | 起音 |
完整旋律与延音线的核对
第三小节是G4附点二分音符，占3q；第四小节依次是F#4、E4、D4三个四分音符。只有一个声部，无伴奏、踏板、反复或额外的速度变化。D4到G4的音区只是本例选择，不限制其他曲子的音区或指法。
谱面对象表保留两个写出的E4音符；演奏事件表则将它们按T1延音线合并为一个E4事件[2,4)。[a,b)包含a、不包含b。共有8个写出的有音高音符、1个休止对象，但只有7次规定起音：q=0、2、4、6、9、10、11。不要把写出的音头数量当成起音数量。
E4只在q=2开始一次，在q=3继续，到q=4换F#4。q=5开始完整1q休止，q=6才开始G4。可选钢琴路线意图在休止开始时放开F#4，在q=12结束最后的D4；这里列的是记谱与练习意图，不是已经测得的放键、持续音量或声音消失时刻。
钢琴发声会衰减，也可能有余音；一个持续音经过拍点时，不应为了每拍都听到新的声音而重复敲键。休止也不保证绝对声学静音。本例没有要求用踏板延续音，更没有用音符时值证明身体动作或学习效果。
| 谱面对象 | 小节:拍 | 音／休止 | q区间 | 延音线角色 |
|---|---|---|---|---|
| N1 | 1:1 | D4 | [0,2) | — |
| N2 | 1:3 | E4 | [2,3) | 起点 |
| N3 | 2:1 | E4 | [3,4) | 终点 |
| N4 | 2:2 | F#4 | [4,5) | — |
| R1 | 2:3 | 休止 | [5,6) | — |
| N5 | 3:1 | G4 | [6,9) | — |
| N6 | 4:1 | F#4 | [9,10) | — |
| N7 | 4:2 | E4 | [10,11) | — |
| N8 | 4:3 | D4 | [11,12) | — |
| 事件 | 音／休止 | q区间 | 谱面对象 |
|---|---|---|---|
| E1 | D4 | [0,2) | N1 |
| E2 | E4 | [2,4) | N2+N3 |
| E3 | F#4 | [4,5) | N4 |
| R1 | 休止 | [5,6) | R1 |
| E4 | G4 | [6,9) | N5 |
| E5 | F#4 | [9,10) | N6 |
| E6 | E4 | [10,11) | N7 |
| E7 | D4 | [11,12) | N8 |
从延音线中间截取，不等于能直接从那里开始
只查看第二、第三小节时，窗口是全局[3,9)，共6q，局部时间以全局q=3为0。窗口开头的E4来自全局q=2的起音，是延续进入；窗口中只有局部q=1的F#4和q=3的G4两次原有起音。不要在数据裁切时悄悄给局部0补一次起音。
这是一张查看窗口，不是从静止直接开始的完整演奏指令。要保留原来的延音线，给练习加上前导：从第一小节第3拍的E4开始，走全局[2,9)，共7q。局部q=0起E4，局部q=1跨小节继续，局部q=2起F#4，局部q=3休止，局部q=4起G4并占3q。
前导版的小节拍次读作“3 | 1、2、3 | 1、2、3”；它跨越已有小节位置，不是新写的7/4小节。这个版本不包含第一小节前两拍，所以它是带前导的局部练习，不可冒充完整四小节已弹过。
如果你选择在第二小节第1拍重新弹E4，那是另一个经过改写的起步练习；必须明确标记新起音，不能将它登记为保留原延音线的同一版本。本课提供的可执行局部路线采用前导版，随后回到q=0的完整四小节检查。
| 窗口／事件 | 全局q | 局部q | 起音 | 延续进入 |
|---|---|---|---|---|
| view/E2 | [3,4) | [0,1) | 否 | 是 |
| view/E3 | [4,5) | [1,2) | 是 | 否 |
| view/R1 | [5,6) | [2,3) | 否 | 否 |
| view/E4 | [6,9) | [3,6) | 是 | 否 |
| lead/E2 | [2,4) | [0,2) | 是 | 否 |
| lead/E3 | [4,5) | [2,3) | 是 | 否 |
| lead/R1 | [5,6) | [3,4) | 否 | 否 |
| lead/E4 | [6,9) | [4,7) | 是 | 否 |
可执行的小检查与复测
纸笔路线：先遮住事件答案，按每小节1、2、3指点十二格，再在格上写“起音、持续或休止”。第二遍给四个小节开头写跨度编号1到4。最后核对q=3没有新增起音，q=5没有被删掉；两行都正确才说明这份纸面任务完成，不代表实际演奏已经稳定。
钢琴路线：先分别找D4、E4、F#4、G4。只在舒适的条件下选择能读清的速度，单音试前导版；数到下一小节的1时保留E4，不重弹。若不能同时数与弹，先纸上指拍或只说数，再试单音。无需强制唱歌、合手或固定指法。
检查局部后，先把第一到第三小节连起来，再恢复完整四小节。分别记录第二小节开头的E4是否误重弹、休止是否完整、第三小节长音是否被多敲，以及完整四小节的结果。出现问题先退回相关纸面位置，不把同一个失误自动归咎于手部动作。
时间换算只示范单位：四分音符=60 BPM时，每小节3秒，全段12秒，带前导局部7秒；四分音符=48 BPM时，分别为3.75、15、8.75秒。计数层级改变不自动改变这些时长。示范速度不是必须达到的目标或实际演奏记录。
记录路线、实际速度及拍单位、提示、预测是否先于答案、实际结果、下一步及约定的复测时间；没观察就写未知。下一次先不看答案重做熟悉材料，并记录真实间隔，再练习。这检查保留；公开过的本例不能再算未见材料迁移，程序验算也不能替代听觉、动作或学习证据。
| 音 | MIDI | 88键序号（A0=1） |
|---|---|---|
| D4 | 62 | 42 |
| E4 | 64 | 44 |
| F#4 | 66 | 46 |
| G4 | 67 | 47 |
来源范围与原创边界
NC-079既有ASR的128段已全部阅读；在126、150、180、208、220、266秒检查6张静帧。126、150、180秒画面可见2/4例子；208秒可见3/4；220秒可见四个框出的3/4小节，分别标1、2、3、4；266秒是带1、2、3标注的华尔兹谱页。这里只核这些标记与布局，不逐音验收源谱。
未经听校的转写把强拍描述为分组参照而非必须更响，也提出同一材料可在不同层级分组。这些是ASR候选归因，不是已经核实的原声或感知实验。源作品的具体分析、必须采用某种解释的强语气，以及普遍的强弱排序，都不作为本课的演奏规则。
原始音频、连续动作、完整源谱和ASR／静帧精确对齐均未验证。清单时长435.584秒，容器时长435.562秒，相差0.022秒尚未解释；这些不是实际听看时长。没有新增ASR或实际学习者演奏证据。
四小节旋律、跨小节E4延音线、两行计数、裁切窗口、带前导练习及记录单均为本库原创，不是对源曲的复刻。公开材料不含原声、源图、完整转写或私人路径。ASR文字加选帧只构成有限模态覆盖，不等于完整视听审阅。
Count the bars through a tie
Does a new bar always require a new attack?
Imagine counting “1,2,3” in 3/4. The next bar’s “1” arrives, but a tie carries the same note across the barline. The count continues without requiring another attack. This lesson uses an original four-bar single-note melody to distinguish beat positions, bar starts and note attacks.
Start with quarter, half and dotted-half notes, quarter rests and letter names; use the rhythm and reading links at the end if needed. C4 is middle C, and the 4 after a note name is an octave number, not a fingering. Paper or spoken counting is sufficient; piano is optional.
Before viewing the answers: bar 1 contains D4 half note and E4 quarter note; bar 2 contains E4 quarter note, F#4 quarter note and a quarter rest, with a tie between the adjacent E4 notes. Should E4 be attacked again on beat 1 of bar 2? How many quarter units do the two written E4 notes occupy together? Record a prediction or “uncertain” first.
The tie here joins two temporally contiguous E4 notes of the same pitch in the same voice into one note event lasting 2q, with an attack only at the first E4’s start. Here q is a quarter-note unit. A tie differs from a slur used for phrasing or articulation, which may join different pitches, and does not mean using the damper pedal.
Two counts sharing one timeline
All four bars remain in 3/4, with 3q per bar. A separate annotation numbers them 1,2,3,4 within the span stipulated for this lesson. This is a chosen reading organization, not evidence that every work forms four-bar phrases. Within-bar counting stays 1,2,3; the time signature has not become 4/4.
Global q starts at 0, while beat numbers within a bar start at 1. Bar starts are q=0,3,6,9, and the passage ends at q=12. The four-bar span contains twelve uniform q units; do not add the two counting rows into sixteen units or treat one span-level number as one quarter note.
In the table, “attack” begins a new single-note event, “continue” means a previous note still occupies its notated duration, and “rest” occupies time of its own. Beat 1 of bar 2, q=3, continues E4, whereas beat 1 of bar 3, q=6, starts G4. Bar starts can therefore have different event types.
Annotations for span, bar and beat help locate events; they do not automatically demand rearticulation, acceleration or greater force. This lesson supplies no fixed formula for loudness, accents or phrase emphasis and does not infer your actual perceived meter from the timeline.
| Global q | Bar:beat | Span bar number | Pitch | Event state |
|---|---|---|---|---|
| 0 | 1:1 | 1 | D4 | attack |
| 1 | 1:2 | 1 | D4 | continue |
| 2 | 1:3 | 1 | E4 | attack |
| 3 | 2:1 | 2 | E4 | continue |
| 4 | 2:2 | 2 | F#4 | attack |
| 5 | 2:3 | 2 | — | rest |
| 6 | 3:1 | 3 | G4 | attack |
| 7 | 3:2 | 3 | G4 | continue |
| 8 | 3:3 | 3 | G4 | continue |
| 9 | 4:1 | 4 | F#4 | attack |
| 10 | 4:2 | 4 | E4 | attack |
| 11 | 4:3 | 4 | D4 | attack |
Check the complete melody and tie
Bar 3 is a G4 dotted-half note lasting 3q; bar 4 contains F#4, E4 and D4 as three quarter notes. There is one voice, with no accompaniment, pedal, repeats or extra tempo changes. The D4-to-G4 register is chosen for this example and does not constrain registers or fingerings in other music.
The notation-object table retains the two written E4 notes; the performance-event table merges them via tie T1 into one E4 event on [2,4). [a,b) includes a and excludes b. There are eight written pitched notes and one rest object, but only seven stipulated attacks: q=0,2,4,6,9,10,11. Written notehead count is not attack count.
E4 begins once at q=2, continues at q=3 and changes to F#4 at q=4. A full 1q rest starts at q=5, and G4 begins only at q=6. In the optional piano route, intend to release F#4 at the rest’s start and end the final D4 at q=12. These are notation and practice intentions, not measurements of key release, sustained loudness or sound cessation.
A piano tone decays and resonance may remain. Do not repeat a key merely to hear a new sound at every beat of a sustained note. A rest also does not guarantee absolute acoustic silence. This example does not ask for pedal sustain, and notated durations establish neither physical movement nor learning effectiveness.
| Notation object | Bar:beat | Pitch / rest | q span | Tie role |
|---|---|---|---|---|
| N1 | 1:1 | D4 | [0,2) | — |
| N2 | 1:3 | E4 | [2,3) | start |
| N3 | 2:1 | E4 | [3,4) | stop |
| N4 | 2:2 | F#4 | [4,5) | — |
| R1 | 2:3 | rest | [5,6) | — |
| N5 | 3:1 | G4 | [6,9) | — |
| N6 | 4:1 | F#4 | [9,10) | — |
| N7 | 4:2 | E4 | [10,11) | — |
| N8 | 4:3 | D4 | [11,12) | — |
| Event | Pitch / rest | q span | Notation objects |
|---|---|---|---|
| E1 | D4 | [0,2) | N1 |
| E2 | E4 | [2,4) | N2+N3 |
| E3 | F#4 | [4,5) | N4 |
| R1 | rest | [5,6) | R1 |
| E4 | G4 | [6,9) | N5 |
| E5 | F#4 | [9,10) | N6 |
| E6 | E4 | [10,11) | N7 |
| E7 | D4 | [11,12) | N8 |
A crop through a tie is not automatically a playable new beginning
Viewing only bars 2–3 creates the global window [3,9), lasting 6q, with global q=3 as local 0. Its opening E4 carries in from an attack at global q=2; the window contains only the original F#4 attack at local q=1 and G4 attack at local q=3. Do not silently insert an attack at local 0 when cropping the data.
This is an inspection window, not a complete instruction to start from silence. To preserve the original tie, add a lead-in: begin E4 on beat 3 of bar 1 and traverse global [2,9), lasting 7q. Attack E4 at local q=0, continue across the barline at local q=1, attack F#4 at local q=2, rest at local q=3, and attack G4 at local q=4 for 3q.
Read the lead-in as “3 | 1,2,3 | 1,2,3” using the original beat positions. It crosses existing bar locations and is not a newly written 7/4 bar. It omits the first two beats of bar 1, so it is a local exercise with a lead-in, not evidence that all four bars were performed.
Choosing to reattack E4 on beat 1 of bar 2 creates a different adapted starting exercise. That new attack must be labeled explicitly rather than recorded as the same version preserving the original tie. The executable local route supplied here uses the lead-in and then returns to all four bars from q=0.
| Window / event | Global q | Local q | Attack | Carry-in |
|---|---|---|---|---|
| view/E2 | [3,4) | [0,1) | false | true |
| view/E3 | [4,5) | [1,2) | true | false |
| view/R1 | [5,6) | [2,3) | false | false |
| view/E4 | [6,9) | [3,6) | true | false |
| lead/E2 | [2,4) | [0,2) | true | false |
| lead/E3 | [4,5) | [2,3) | true | false |
| lead/R1 | [5,6) | [3,4) | false | false |
| lead/E4 | [6,9) | [4,7) | true | false |
A small check you can perform and revisit
Paper route: cover the event answers, point through twelve slots counting 1,2,3 per bar and mark each slot “attack, continue or rest.” On a second pass, mark span numbers 1–4 at the four bar starts. Finally check that q=3 gained no attack and q=5 was not removed. Correctly completing both rows establishes this paper task, not stable performance.
Piano route: locate D4, E4, F#4 and G4 separately. Under comfortable conditions, choose a tempo that permits clear reading and try the single-note lead-in; keep E4 at the next bar’s 1 without reattacking. If counting and playing together is too much, first point to the beats on paper or speak the count, then try single notes. Singing, two-hand playing and a fixed fingering are not required.
After checking the local section, connect bars 1–3 and then restore all four bars. Record separately whether E4 was mistakenly reattacked at bar 2, whether the rest remained complete, whether the long G4 received extra attacks, and the full four-bar result. If a problem occurs, return to the relevant paper position first rather than automatically attributing it to hand movement.
Tempo conversion only illustrates units: at quarter note=60 BPM, a bar lasts 3 seconds, the passage 12 seconds and the lead-in section 7 seconds; at quarter note=48 BPM, the values are 3.75,15 and 8.75 seconds. Changing the counting level does not itself change these durations. These tempi are neither required targets nor actual performance records.
Record the route, actual tempo and beat unit, cues, whether prediction preceded the answer, actual result, next action and planned retest time; leave unobserved fields unknown. Next time, retry the familiar material without answers and record the actual interval before practicing. This checks retention; this published example cannot serve as unseen-material transfer, and arithmetic validation cannot replace auditory, movement or learning evidence.
| Pitch | MIDI | 88-key index (A0=1) |
|---|---|---|
| D4 | 62 | 42 |
| E4 | 64 | 44 |
| F#4 | 66 | 46 |
| G4 | 67 | 47 |
Source scope and originality
All 128 segments of the existing NC-079 ASR were read, and six stills at 126,150,180,208,220 and 266 seconds were inspected. The first three show 2/4 examples; 208 shows 3/4; 220 shows four boxed 3/4 bars labeled 1,2,3,4; 266 shows a waltz score page with 1,2,3 annotations. Inspection covers these labels and layouts, not a note-by-note audit of the source score.
The unaudited transcript describes strong beats as grouping references rather than necessarily louder events and proposes grouping material at different levels. These are ASR-candidate attributions, not verified source audio or perceptual experiments. Specific analyses of source works, emphatic demands for one interpretation and universal strength rankings are not used as performance rules here.
Source audio, continuous motion, the complete source score and precise ASR/still alignment remain unverified. Inventory duration is 435.584 seconds and container duration 435.562 seconds; the 0.022-second difference remains unexplained, and these are not actual listening/viewing times. There was no new ASR or actual learner-performance evidence.
The four-bar melody, E4 tie across the barline, two counting rows, crop window, lead-in exercise and record are original to this repository, not a reconstruction of source music. Public files contain no source audio, images, complete transcript or private paths. ASR text plus selected stills provides limited modality coverage, not full audiovisual review.
Prerequisites and review: rhythm, reading, interval spelling, practice project. Checkable data, source limits.