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.
练到哪里算完成：句尾预期与实际落点
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。
你以为下一和弦应该“回家”，谱上却写了另一个落点。先别改谱，也别把期待落空当弹错：分开记录“我预测什么”“实际写什么”“我真的听到什么”。用这三栏决定下一次练习要核对什么。
新读NC-014／015全部133段既有ASR，检查14个选定静帧；24段片头／片尾、重复或可疑专名候选排除出归属。535.297秒是继承的视频账本时长合计；同一哈希的MP4容器元数据合计535.253秒，差异原因未查明。二者均不是听校时长。没有新增ASR、原音听校、连续动作或完整来源谱例验收。
本包可完全纸面完成，不需要琴、唱歌、录音或麦克风。先认出音名和升降号：Bb是降B，在B白键左边紧邻的黑键；这里用国际音名B表示B本音。数字是八度，不是指法。键号从88键最左A0=1起，黑白键一起数，因此C4=40。纸面核对不要求同时按住任何和弦。
G：乐句是在上下文中有组织的一段音乐；终止式cadence涉及句尾停顿或结束的组织。罗马数字在这里表示指定调内的和弦级数，不是指法或拍数。只有两个和弦标签，还不足以判断完整终止类型、作品调性或作曲家的全部意图。
G：本例指定F大调：F G A Bb C D E。I是F大三和弦F–A–C，vi是D小三和弦D–F–A，V7是C7的C–E–G–Bb。根音用于命名，最低音由具体配置决定。V7中的7表示这个属和弦含七音，不是第7级根音；本例C7的七音是Bb。这里只计算已给定调中的身份，不从两个和弦“证明”调性。
有边界的作者归属
B36-S01 · T2
Wiwi的入门图把V→I和停在V作句尾比较，并分别标注较像结束与仍待继续的感觉。我们保留“观察句尾”的问题，不把这个简化图推广成任何V→I都是完整终止，或每个人都必然产生同样听感。
NC-014静帧62、80、146、154、217秒；111与196秒只查末端和弦标签和布局。ASR第18–29、39–58段仅作上下文定位，未逐音验收完整曲例。
B36-S02 · T2
Wiwi用预期落点被替换来解释“假终止”，图中写V到非I、通常为vi，并给出G7→Am标签。这个入门说明不等于所有非I连接都可不看上下文而判作假终止；本文的音符任务另行原创。
NC-015静帧41、98、137秒；ASR第9–32段只作上下文定位。其他色彩、情绪与脑反应断言未独立验证，未成为本库的感知规律。
先写判断依据，再分配练习任务
O：读自己的合法版本时，把候选句尾标成“位置待定／有上下文支持”，并核对旋律、低音、拍位和前后连接。小节线、休止或换行都不能单独证明句尾。先区分真正的乐句边界与为了练习临时切出的区间。
O：在揭示后续前先写预期与理由，允许“没有预期”。揭示后另记实际内容，不把自己的猜测改写成谱面事实。熟悉录音、已看过的答案和自己推想要分栏，不能都叫作首次由谱想象声音。
O：把“练好句尾”改成能检查的一句话，例如“保留节拍，准确到达谱写的目标音组，并保持到约定结束”。若只练低音，就只验低音定位，不能填成完整和声或双手通过。若缩短区间，仍需明确所有起音、保持与结束，不能漏掉连接音后就宣布完整乐句完成。
O：立即反馈分三项：①音与时间是否照任务；②自己的预期是否与所写落点相同；③实际听感是否有证据。第②项不同不自动表示第①项错误。没有实际演奏／声音证据时，第③项保持“未验证”，不推断手指问题。
O：休息后遮住标注，从谱面重建目标与理由；次日用同例检查是否还记得目标与做法。之后放回前句、目标与后续共同核对，保留作品要求的速度单位、力度、奏法与乐句停连提示。新材料迁移需要另准备未见过的任务，本包尚未做这项验证。
同一个出发和弦，两种已给定的落点
只做纸面。原创两小节，每小节4/4，四分音符=60 BPM，因此总长8拍=8秒。offset从0计：第一音组在0开始、3.5释放；第二音组在4开始、7.5释放。0与4分别是两小节的第1拍。各音同时开始，不能把同时音符的时长相加当经过时间。事件表是目标区间，不是来源五线谱。
先读共同出发组C3/E4/G4/Bb4。遮住从下一段开始的落点说明、图、表与答案，在纸上写：在指定F大调中，我预测根音去哪、为何这样猜，或“未知”。再揭示A/B并核对。两个答案都会公开，这是一项引导练习，不能据此声称盲听或未见材料迁移。
用原创两栏落点图与事件表解释：出发C7→A的F，或出发C7→B的Dm。箭头只表示这个任务的顺序，不代表必须这样解决或已经听到结束感。没有固定指法、配手、踏板或奏法处方。
C7 ── A: F
└─ B: Dm| 版本 | 事件 | 音高／键号 | 起点 | 释放 |
|---|---|---|---|---|
| A | A1 | C3 / 28 | 0 | 3.5 |
| A | A2 | E4 / 44 | 0 | 3.5 |
| A | A3 | G4 / 47 | 0 | 3.5 |
| A | A4 | Bb4 / 50 | 0 | 3.5 |
| A | A5 | F3 / 33 | 4 | 7.5 |
| A | A6 | A3 / 37 | 4 | 7.5 |
| A | A7 | C4 / 40 | 4 | 7.5 |
| A | A8 | F4 / 45 | 4 | 7.5 |
| B | B1 | C3 / 28 | 0 | 3.5 |
| B | B2 | E4 / 44 | 0 | 3.5 |
| B | B3 | G4 / 47 | 0 | 3.5 |
| B | B4 | Bb4 / 50 | 0 | 3.5 |
| B | B5 | D3 / 30 | 4 | 7.5 |
| B | B6 | A3 / 37 | 4 | 7.5 |
| B | B7 | D4 / 42 | 4 | 7.5 |
| B | B8 | F4 / 45 | 4 | 7.5 |
预测后核对
A目标的实际音为F3/A3/C4/F4，是F大三和弦，F重复一个八度；B为D3/A3/D4/F4，是D小三和弦，D重复一个八度。两者的最高音都是F4，但根音、最低音与中间音不同。故“最高音相同”不等于“和声身份相同”。这里A是V7→I，B是V7→vi；完整终止类型与实际听感仍未确定。
每版只有两个音组起点，间隔4拍=4秒；16个音符事件是两版各8个，不是16次先后起音。每版3.5–4与7.5–8没有保持音，都是0.5拍。第二音组仍占据小节2的目标区间，不因猜错和弦而提前结束。
写出两版各自的音组身份、最低音、最高音、起点与释放点，再将结果与答案逐项比较。若只猜中F4，不填写“和弦预测正确”；若知道B是Dm而不会想象声音，就记“结构正确／内听未验证”。纸面通过不表示能同时弹出这些配置。
稍后只看出发组，重写两种指定目标的实际音，不改拍数。次日同例检查后，再读自己的完整乐句，说明为何选那个练习落点；若谱面／版本还未核对，保持“目标待定”。本包没有要求在未核版本的《Kiss the Rain》或《River Flows in You》上套用任何小节分析。
范围与相关阅读
ASR反复把“终止”写成同音词；本包术语依据实际标题／说明帧，不按机器字面教错词。来源的情绪形容、历史教材例子、脑自动补和弦以及其他课程链接没有因此获得独立验证。14静帧只是选点，不是完整观看；来源曲谱、音视频与转录不随包发布。
NC-015的231／250秒画面带运动模糊，只记录可辨标题／标签，不支持精确音符读数。主要归属使用41／98／137秒。ASR术语以已检查标题帧定位，未把未听到的数值、作品或情绪标签补成事实。
本包来源范围
原创事件与和弦身份检查
What counts as finishing: phrase expectations and actual arrivals
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.
You expect the next chord to return “home,” but the score specifies another arrival. Do not change the score or equate a frustrated expectation with a wrong note. Record your prediction, what is actually written and what you actually heard separately. Use these three columns to choose what the next practice attempt should check.
All 133 existing ASR segments of NC-014/015 were newly read and 14 selected stills inspected; 24 opening/closing, repeated or suspect-name candidates are excluded from attributions. The inherited video inventory totals 535.297 seconds; metadata of the same-hash MP4 containers totals 535.253 seconds, with the cause of the difference unestablished. Neither is audition time. No new ASR, source audition, continuous-motion review or complete source-score acceptance occurred.
Complete this package entirely on paper without a piano, singing, recording or microphone. First identify note names and accidentals: Bb is B-flat, the black key immediately left of the B white key; B here means B-natural in international pitch naming. Numerals mark octaves, not fingers. Key numbers count black and white keys from leftmost A0=1 on an 88-key piano, so C4=40. Paper checking does not require holding any chord.
G: a phrase is a contextually organized musical span; a cadence concerns how a phrase pause or close is articulated. Roman numerals here label chord degrees in an assigned key, not fingers or beat counts. Two chord labels alone do not establish a complete cadence type, a work’s key or the composer’s complete intention.
G: this example assigns F major: F G A Bb C D E. I is the F-major triad F–A–C, vi the D-minor triad D–F–A, and V7 is C7 with C–E–G–Bb. The root names the chord; the actual bass depends on its arrangement. The 7 in V7 indicates that this dominant chord contains a seventh, not that its root is scale degree 7; C7’s seventh here is Bb. We calculate identities in an assigned key rather than proving a key from two chords.
Bounded author attributions
B36-S01 · T2
Wiwi’s introductory diagrams contrast phrase endings moving V→I with endings on V, labeling them as more closed or still continuing. We retain the question of examining a phrase ending, without treating every V→I as a complete cadence or predicting identical perceptions for every listener.
NC-014 stills at 62, 80, 146, 154, 217 seconds; 111/196 seconds were checked only for ending chord labels and layout. ASR segments 18–29, 39–58 locate context only; the full examples were not accepted note by note.
B36-S02 · T2
Wiwi explains deceptive cadence through a replaced expected arrival, displaying V to non-I, usually vi, and G7→Am labels. This introduction does not classify every non-I continuation as deceptive without context; our note task is separately original.
NC-015 stills at 41, 98, 137 seconds; ASR segments 9–32 locate context only. Other color, emotion and brain-response claims are not independently verified or adopted as perceptual laws.
Record the basis for a judgment before assigning practice
O: in your lawfully obtained edition, label a candidate ending as location uncertain or context supported, then inspect melody, bass, metrical position and surrounding connections. A barline, rest or system break alone does not prove a phrase ending. Separate a musical boundary from a temporary practice cut.
O: before revealing a continuation, record an expectation and reason; no expectation is an acceptable response. After revealing it, record the actual content separately. Keep a familiar recording, an already-seen answer and your own prediction in different columns rather than calling all of them first-time hearing from notation.
O: replace “improve the ending” with an observable statement, such as preserving the pulse, arriving at the written target pitches and holding to the agreed end. A bass-only task checks bass location, not complete harmony or both hands. A shorter interval must still specify every attack, hold and ending; omitting a connecting note does not complete the phrase.
O: immediate feedback has three parts: whether notes/timing match the task, whether your expectation matches the written arrival, and whether actual listening has evidence. A mismatch in the second does not automatically make the first wrong. Without actual performance/sound evidence, keep the third unverified rather than diagnosing fingers.
O: after a break, hide annotations and reconstruct the target and reason from the score; a next-day same-example retest checks retention. Then check the preceding phrase, arrival and continuation together, preserving the work’s tempo unit, dynamics, articulation and phrasing cues. Unseen transfer requires a separately prepared unfamiliar task and has not been validated here.
One departure chord, two specified arrivals
Paper only. Two original bars each use 4/4 at quarter note=60 BPM, totaling 8 beats=8 seconds. Offsets start at 0: the first pitch group starts at 0 and releases at 3.5; the second starts at 4 and releases at 7.5. 0 and 4 are beat 1 of their respective bars. Notes within a group start together; their simultaneous durations do not sum to elapsed time. This event table is a target-interval representation, not a source score.
First read the shared departure C3/E4/G4/Bb4. Cover all following arrival explanations, diagram, table and answers and write where you expect the root to move in assigned F major, why, or unknown. Then reveal A/B and check. Both answers are disclosed; this guided task establishes neither blind listening nor unseen-material transfer.
Use an original two-arrival diagram and event table: departure C7→A’s F, or C7→B’s Dm. Arrows show this task’s order, not a required resolution or an established feeling of closure. No fixed fingering, hand allocation, pedal or articulation prescription is supplied.
C7 ── A: F
└─ B: Dm| Version | Event | Pitch / key | Start | Release |
|---|---|---|---|---|
| A | A1 | C3 / 28 | 0 | 3.5 |
| A | A2 | E4 / 44 | 0 | 3.5 |
| A | A3 | G4 / 47 | 0 | 3.5 |
| A | A4 | Bb4 / 50 | 0 | 3.5 |
| A | A5 | F3 / 33 | 4 | 7.5 |
| A | A6 | A3 / 37 | 4 | 7.5 |
| A | A7 | C4 / 40 | 4 | 7.5 |
| A | A8 | F4 / 45 | 4 | 7.5 |
| B | B1 | C3 / 28 | 0 | 3.5 |
| B | B2 | E4 / 44 | 0 | 3.5 |
| B | B3 | G4 / 47 | 0 | 3.5 |
| B | B4 | Bb4 / 50 | 0 | 3.5 |
| B | B5 | D3 / 30 | 4 | 7.5 |
| B | B6 | A3 / 37 | 4 | 7.5 |
| B | B7 | D4 / 42 | 4 | 7.5 |
| B | B8 | F4 / 45 | 4 | 7.5 |
Check after prediction
A’s target is F3/A3/C4/F4: an F-major triad with F doubled at the octave. B’s D3/A3/D4/F4 is a D-minor triad with D doubled at the octave. Both have F4 highest, but their roots, basses and internal tones differ. The same highest pitch does not establish the same harmony. Here A is V7→I and B V7→vi; complete cadence type and actual listening remain undetermined.
Each version has only two pitch-group starts, 4 beats=4 seconds apart. Sixteen note events mean 8 in each version, not 16 successive attacks. Each version has no held notes at 3.5–4 and 7.5–8, both 0.5 beat gaps. The second group retains its bar 2 target interval even if your harmonic prediction was wrong.
Write each version’s chord identity, bass, highest pitch, starts and releases, then compare each item with the answers. Guessing F4 alone does not mean a correct chord prediction. Recognizing Dm without imagining its sound should be recorded as structure correct/inner hearing unverified. Passing on paper does not establish the ability to play these voicings.
Later, view only the departure and rewrite the actual pitches of both specified targets without changing the beat count. After a next-day same-example check, read your complete phrase and explain the chosen practice arrival; keep the target undecided if the score/edition is unverified. This package assigns no bar analysis to an unverified edition of Kiss the Rain or River Flows in You.
Scope and related reading
ASR repeatedly renders the cadence term as homophones; this package uses inspected title/explanation frames rather than teaching machine spellings. Source emotional labels, historical method-book examples, automatic-brain-completion claims and other course links are not thereby independently verified. 14 stills are selected points, not complete viewing; source scores, media and transcripts are excluded.
NC-015’s 231/250-second frames have motion blur; only legible headings/labels are recorded, not precise note readings. Main attribution uses 41/98/137 seconds. Inspected title frames locate ASR terminology; unheard numbers, works or emotional labels are not filled in as facts.
Separate root, bass and voicing first
Preserve voices and complete short-segment joins
This package’s source scope
Original-event and chord-identity checks