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This page preserves the author’s frozen review bytes, so prose or data may still contain candidate, pending-review or unpublished labels.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.
半音下行：分清声部、低音与根音
先找下行的是哪条线
看见G、F#、F、E依次出现，能直接说“低音在半音下行”吗？还不能。必须知道每个音属于哪条声部，以及同一时刻有没有更低的音。本课的两份原创配置保留相同的和弦音高类，却把这条线放在不同音区；你会分别填写旋律线、实际低音与给定和弦根音。
前置是字母音名、半音、三和弦与拍数；文末有相关笔记。b表示降号，#表示升号；F是不带升降号的还原F。C4是中央C，末尾数字表示八度，不是指法。半音距离按十二平均律琴键计算。先做纸笔；本页没有必须一次弹出的四声部手位要求。
先遮住结果表预测：若G2、Bb3、D4一直保持，而最高声部从G4走到F#4、F4、E4，低音是否移动？若把变化放在最低声部，G2变F#2、F2、E2，G和弦的根音是否必须一起下降？把“不确定”和理由也写下来。
线条、低音、根音是三个问题
声部（voice）是我们持续追踪的一条音线。低音（bass）是在当前配置中实际最低的音；它不一定是和弦根音。根音（root）是所采用和弦名称的组织参照。分析时不能从四个和弦各挑一个音拼成下行路线，再声称那个路线已经是同一个声部。
本课固定采用G根音读法，作为G小调参照下的原创和弦拼写题。Gm三和弦是G–Bb–D；Bb保持小三度。变化的F#、F与E分别是G的大七度、小七度与大六度，得到Gm(maj7)、Gm7、Gm6。m指小三和弦，不会使附加的6自动变成小六度；Gm6用E，不用Eb。
Gm(maj7)也不是Gmaj7：前者含Bb，后者含B。这里F#到F是一半音下行，虽然字母仍为F；F到E也是一半音下行。按G4作起点，整条G4–F#4–F4–E4路线位移为0、−1、−2、−3半音。它是四个位置的半音线，不是一套完整的小调音阶。
后面用斜线标出不同于根音的低音：Gm7/F表示G–Bb–D–F这组音以F作最低音，根音仍按题意读G。斜线右侧不表示下一和弦，也不是“某调的属和弦”中的罗马数字斜线。本文不以最低音自动重命名根音。
采用的和弦拼写（无独立时值）
| 名称 | 音名 | 从G起的半音类 |
|---|---|---|
| Gm | G Bb D | 0,3,7 |
| Gm(maj7) | G Bb D F# | 0,3,7,11 |
| Gm7 | G Bb D F | 0,3,7,10 |
| Gm6 | G Bb D E | 0,3,7,9 |
两版完整时间表
U版把下行线放在最高声部V4；B版把它放在低音声部V1。两版都是四小节3/4，总计12q；q是四分音符单位，小节开头在0、3、6、9。前三个和弦区间各3q，最后一个有声区间只有2q，随后全体休止1q。因此第四小节仍满3q。
音名[a,b)表示a起音、b结束，含起点不含终点。一个保持事件跨过小节线时继续保持，没有再次起音；若改写成五线谱，应将跨小节时值拆开并用延音线连接。本表已经列全四条声部，包括[11,12)的四个休止，没有隐藏伴奏或默认重复。
每版有7个音符起音事件：移动声部4个，其余三个声部各1个。q=0有四个声部同时起音，q=3、6、9各只有一个新起音。两版共14个音符事件、8个休止事件；同时声部的时长不相加，所以每版长度是12q，不是48q。四分音符=60 BPM时每版谱写长度12秒；这是计算例，不是推荐速度或演奏实测。
每条声部的全部事件
| 版本 | 声部 | 全部音符事件 | 休止 |
|---|---|---|---|
| U | V1 | G2[0,11) | [11,12) |
| U | V2 | Bb3[0,11) | [11,12) |
| U | V3 | D4[0,11) | [11,12) |
| U | V4 | G4[0,3); F#4[3,6); F4[6,9); E4[9,11) | [11,12) |
| B | V1 | G2[0,3); F#2[3,6); F2[6,9); E2[9,11) | [11,12) |
| B | V2 | Bb3[0,11) | [11,12) |
| B | V3 | D4[0,11) | [11,12) |
| B | V4 | G4[0,11) | [11,12) |
同组音高类，不同的最低音
U版V1始终是G2；V4下降到E4时仍在D4上方，所以四个区间的最低音均为G2。B版V1的G2–F#2–F2–E2始终低于Bb3，因此那条下行线确实就是低音线。两版的内声部V2=Bb3与V3=D4都不动。
两版第一和弦的四个实际音完全相同。此后U保留G2、改变最高音，B保留G4、改变最低音；所以它们不只是把同一个声音换个名字，也不是整个和弦统一移高或移低。每个对应区间的音高类集合相同，音区分配与声部路线不同；相同集合不保证相同听感。
采用本题G根音读法，U依次标Gm、Gm(maj7)、Gm7、Gm6；B依次标Gm、Gm(maj7)/F#、Gm7/F、Gm6/E。B的低音下降3半音，但给定根音仍是G。于是“半音下行低音”与“半音下行根音进行”在本例不是同一句话。
在B的两个七和弦中，低音F#或F是各自的七音；按七和弦转位名称可称第三转位。最后的Gm6/E却不能因“E在下面”就自动套用同一个七和弦转位名称：6在这里是大六度附加音，不是七音。下一节还会给出这组音的另一种根音读法。
逐区间核对
| q区间 | U实际音（低到高） | U低音／根音 | U标签 | B实际音（低到高） | B低音／根音 | B标签 |
|---|---|---|---|---|---|---|
| [0,3) | G2 Bb3 D4 G4 | G2 / G | Gm | G2 Bb3 D4 G4 | G2 / G | Gm |
| [3,6) | G2 Bb3 D4 F#4 | G2 / G | Gm(maj7) | F#2 Bb3 D4 G4 | F#2 / G | Gm(maj7)/F# |
| [6,9) | G2 Bb3 D4 F4 | G2 / G | Gm7 | F2 Bb3 D4 G4 | F2 / G | Gm7/F |
| [9,11) | G2 Bb3 D4 E4 | G2 / G | Gm6 | E2 Bb3 D4 G4 | E2 / G | Gm6/E |
标签的边界：集合不唯一决定功能
G–Bb–D–E也可按E–G–Bb–D组织成Em7b5，即E半减七和弦：从E起为0、3、6、10半音。因此U最后一组可写Em7b5/G，B最后一组可写Em7b5。两种命名保留相同实际音，却选择了不同根音。这是给定材料的同一音高类集合分析，不是宣告哪一个功能必然正确。
要在真实作品中选择更有解释力的名称，还需前后和弦、乐句重心、节奏位置与更长范围的语境。本题预先固定G根音，目的是避免把“最低音”误当成“根音”的定义；它不是通过四个集合证明作品一直在G小调，也不是证明最后已经转到E调。
F#或E的出现本身不证明转调、某一种旋律小调方向规则，或某种必然的情绪。若把最后E改成Eb，下行会变成F到Eb的2半音，和弦也不再符合这里的Gm6拼写。能算出变化，不等于已经听出“更顺”“更悲伤”或得到通用和弦替换规则。
纸笔练习：先修一段，再恢复全部
先遮住B版与结果表，复制U的四条完整声部。要构造B，只把V1改为G2[0,3)、F#2[3,6)、F2[6,9)、E2[9,11)，V4改为G4[0,11)；V2、V3及四个末尾休止不改。按0、3、6、9分别写出四个同时音、最低音与采用的根音，再核对结果表。
纠错题：有人把U在q=3的最低音写F#4，因为这是新起音；又把B在q=9写成“根音必是E，因为E2最低”。找出各自漏掉的证据。答案：U还有持续的G2；B的最低音确为E2，但本题Gm6/E约定根音G，另读Em7b5时才采用E根音。新起音、实际最低音与分析选择不能互相代替。
若在F#和F之间出错，先截取原q=3到9的6q片段。局部从0计，移动声部F#在[0,3)、F在[3,6)，另外三条声部作为原先已保持的音带入整个[0,6)。纸上注明“带入保持，不是原谱q=3的新起音”。若单独弹这个片段，另外三音的开头按键只是片段重建，不应加进原版的7次起音计数。
局部核对后立即恢复所选版本全部四小节：保留q=0的起音、q=9到11的E及q=11到12的全体休止。用五项核对整句：声部身份、音名与八度、最低音与根音分列、起点及保持、末尾休止。只完成中间6q不能记录整句通过。
钢琴路线可选：先只弹U的V4单线G4–F#4–F4–E4，时值3、3、3、2q，然后休止1q，不踩右踏板。对应88键号（A0=1）为47、46、45、44；B的V1对应23、22、21、20。单线弹奏能记录所弹音名、起点和休止，不能代替完整四声部配置的听感比较。若未弹，演奏观察保持空白。
完成整句后，计划在下一次练习开始、重新练习本题前，用同一版本作延迟复测。保留相同路线与可见提示；若用了节拍器，保留同一设定。先遮答案重建或单线弹完整12q，再按五项核对，记录实际间隔。这是熟悉材料的保留记录，不是新材料迁移或长期效果的证明。
记录中分开填写预测、实际错误、局部修复、整句恢复、复测和听感。若把新起音误作低音，下一步检查仍在保持的声部；若漏休止，下一步只核时间轴，再恢复整句。听感可以写“尚不能判断”，未实施项目不填成成功。
待填写的记录
| 项目 | 初始值 |
|---|---|
| 日期、版本、路线、提示、实际速度 | 未知 |
| 预测、错误与局部修复 | 未知 |
| 五项核对与整句恢复 | 未知 |
| 计划复测、实际间隔、同材料结果 | 未知 |
| 实际弹奏及听感（可选） | 未知 |
来源边界与本课增加的比较
NC-075既有ASR共95段已全部阅读，并查看65、112、134、150、168秒的5张静帧。65秒可见“半音階下行”红箭头与带斜线和弦标签的八格谱面；其他所选帧可见不同位置的指示箭头，168秒另有C→D7→G卡片。这里只核可见文字与布局，不逐音验收整份源谱。
转写候选讨论最低声部半音下行、为给定低音选择不同和弦，以及最后转向属和弦的示例；这些归属尚未听校。模糊唱名、不明确音名、和弦听感形容与简化的小调改法不作为本课定音或通用规则。五张图不能验证原声、连续动作或ASR与画面的精确对齐。
来源主题是下行低音，不是本课的G小调上声部四音模型。U/B音区、Gm(maj7)到Gm6的配置、相同集合不同低音的对照、替代根音读法、时间表与练习流程均为本库原创。此处把来源提出的“低音与和弦选择”问题延伸为可核对的声部配置题，不宣称原片教了这两个版本。
库存时长365.568秒，容器时长365.546秒，相差0.022秒且原因未明；两者不是本次实际听看时长。本稿等待独立审阅，没有原声听校、连续动作、完整源谱、实奏或学习效果验收。公开文件不含源媒体、源图、全文转写或私人路径；有限ASR文字与选定静帧覆盖不能称作全视听验收。
相关阅读
结构化数据
来源记录
固定例题检查
Chromatic descent: distinguish voice, bass and root
First locate the descending line
If G, F#, F and E appear in succession, can we immediately say that the bass descends chromatically? We still need to know which voice carries each note and whether another note lies below it. Two original voicings retain the same chord pitch classes while placing this line in different registers; you will separately identify the moving line, actual bass and stipulated chord root.
Prerequisites are letter names, semitones, triads and beat counting; related notes are linked below. A b means flat and # means sharp; F without an accidental means F-natural. C4 is middle C and the final numeral is an octave, not a fingering. Semitone distances use an equal-tempered keyboard. Start on paper; no four-voice hand position must be played all at once.
Before revealing the results, predict: if G2, Bb3 and D4 remain while the top voice moves from G4 to F#4, F4 and E4, does the bass move? If instead the lowest voice changes from G2 to F#2, F2 and E2, must the root of the named G chord descend too? Record uncertainty and your reason as well.
Line, bass and root are three questions
A voice is one musical line we follow through time. The bass is the actual lowest note in the current voicing; it need not be the chord root. The root is the organizing reference of the adopted chord name. We cannot choose a convenient note from each chord, join them into a descent, and thereby establish that they belonged to one voice.
This original chord-spelling task stipulates a G-root reading with G minor as reference. The Gm triad is G–Bb–D, retaining Bb as its minor third. The changing F#, F and E are respectively the major seventh, minor seventh and major sixth above G, giving Gm(maj7), Gm7 and Gm6. The m indicates the minor triad; it does not automatically make the added sixth minor. Gm6 uses E, not Eb.
Gm(maj7) is also different from Gmaj7: the former contains Bb and the latter B. F# to F descends one semitone even though the letter stays F; F to E also descends one semitone. Relative to G4, the whole G4–F#4–F4–E4 path has offsets 0,−1,−2,−3 semitones. This is a four-position chromatic line, not a complete minor scale.
A slash will specify a bass different from the root: Gm7/F places F lowest in the set G–Bb–D–F, while G remains the stipulated root. The right side of this slash is neither the next chord nor the Roman-numeral slash used to identify a local dominant target. We do not automatically rename the root after the lowest note.
Adopted chord spellings (no separate durations)
| Name | Spelling | Semitone classes from G |
|---|---|---|
| Gm | G Bb D | 0,3,7 |
| Gm(maj7) | G Bb D F# | 0,3,7,11 |
| Gm7 | G Bb D F | 0,3,7,10 |
| Gm6 | G Bb D E | 0,3,7,9 |
Two complete timelines
U places the descent in the top voice V4; B places it in bass voice V1. Both versions are four bars of 3/4, totaling 12q, where q is a quarter-note unit and bars start at 0,3,6,9. The first three chord windows each last 3q; the last sounding window lasts only 2q, followed by a 1q rest in every voice. The fourth bar therefore still contains 3q.
Pitch[a,b) means an attack at a and an end at b, including the start but excluding the end. A held event continues across a barline without a new attack; staff notation would split such a duration at barlines and connect it with ties. The table specifies all four voices, including four rests at [11,12), with no hidden accompaniment or default repetition.
Each version has seven note-attack events: four in the moving voice and one in each of the other three. All four voices attack at q=0, with only one new attack at each of q=3,6,9. The two versions contain fourteen note events and eight rest events. Concurrent voice durations do not add, so one version lasts 12q, not 48q. At quarter note=60 BPM its notated duration is twelve seconds, a calculation rather than a recommended tempo or measured performance.
Every event in every voice
| Version | Voice | All note events | Rest |
|---|---|---|---|
| U | V1 | G2[0,11) | [11,12) |
| U | V2 | Bb3[0,11) | [11,12) |
| U | V3 | D4[0,11) | [11,12) |
| U | V4 | G4[0,3); F#4[3,6); F4[6,9); E4[9,11) | [11,12) |
| B | V1 | G2[0,3); F#2[3,6); F2[6,9); E2[9,11) | [11,12) |
| B | V2 | Bb3[0,11) | [11,12) |
| B | V3 | D4[0,11) | [11,12) |
| B | V4 | G4[0,11) | [11,12) |
Same pitch classes, different lowest notes
In U, V1 remains G2. Even when V4 reaches E4 it stays above D4, so G2 is lowest in every window. In B, the G2–F#2–F2–E2 path in V1 remains below Bb3, so the descent really is the bass line. The inner voices V2=Bb3 and V3=D4 remain unchanged in both versions.
The four actual pitches of the opening chord match exactly. Thereafter U keeps G2 and changes its top note, whereas B keeps G4 and changes its lowest note. This is neither merely renaming one sound nor uniformly transposing the whole chord. Corresponding windows share pitch-class sets while register allocation and voice paths differ; matching sets do not guarantee matching impressions.
Under the stipulated G-root reading, U is labeled Gm, Gm(maj7), Gm7, Gm6; B is labeled Gm, Gm(maj7)/F#, Gm7/F, Gm6/E. B’s bass descends three semitones while the stipulated root remains G. Thus a chromatically descending bass and a chromatically descending root progression are different descriptions in this example.
In B’s two seventh chords, bass F# or F is the chordal seventh, so the seventh-chord inversion name is third inversion. The final Gm6/E does not automatically take the same seventh-chord inversion name simply because E is underneath: here 6 is an added major sixth, not a seventh. The next section gives another root reading of that collection.
Window-by-window check
| q interval | U actual pitches (low to high) | U bass / root | U label | B actual pitches (low to high) | B bass / root | B label |
|---|---|---|---|---|---|---|
| [0,3) | G2 Bb3 D4 G4 | G2 / G | Gm | G2 Bb3 D4 G4 | G2 / G | Gm |
| [3,6) | G2 Bb3 D4 F#4 | G2 / G | Gm(maj7) | F#2 Bb3 D4 G4 | F#2 / G | Gm(maj7)/F# |
| [6,9) | G2 Bb3 D4 F4 | G2 / G | Gm7 | F2 Bb3 D4 G4 | F2 / G | Gm7/F |
| [9,11) | G2 Bb3 D4 E4 | G2 / G | Gm6 | E2 Bb3 D4 G4 | E2 / G | Gm6/E |
Limits of labels: a set does not uniquely determine function
G–Bb–D–E can also be organized as E–G–Bb–D, Em7b5 or an E half-diminished seventh chord, with offsets 0,3,6,10 semitones from E. U’s last sonority could therefore be labeled Em7b5/G and B’s Em7b5. The two naming approaches retain the actual pitches but choose different roots. This analyzes the same pitch-class collection, without declaring one function necessarily correct.
Choosing the more explanatory name in a real work also requires surrounding chords, phrase emphasis, rhythmic placement and a broader context. This task fixes G as root to prevent treating “lowest note” as the definition of “root.” Four collections neither prove that a work remains in G minor nor establish a final modulation to E.
The presence of F# or E alone proves neither modulation, a particular melodic-minor direction rule nor an inevitable emotion. Changing final E to Eb makes F to Eb a two-semitone step and no longer matches the Gm6 spelling used here. Computing a change does not mean that we have heard greater smoothness or sadness, or established a general substitution rule.
Paper task: repair a span, then restore the whole
Hide B and the results table, then copy all four complete voices of U. To construct B, change only V1 to G2[0,3), F#2[3,6), F2[6,9), E2[9,11) and V4 to G4[0,11), retaining V2, V3 and all four closing rests. At 0,3,6,9 list the four simultaneous pitches, lowest note and adopted root, then check the results table.
Repair these statements: someone calls F#4 the bass of U at q=3 because it is the new attack, and writes “the root of B at q=9 must be E because E2 is lowest.” What evidence is missing? U still contains the held G2. B does have E2 as bass, but the stipulated Gm6/E reading uses root G; the alternative Em7b5 reading adopts E. A new attack, actual lowest note and analytical choice cannot substitute for one another.
If F# and F are confused, isolate the six-q span from original q=3 to 9. Starting local time at zero, the moving voice has F# at [0,3) and F at [3,6), while the other three voices enter as previously held notes through [0,6). Mark “carried in, not new attacks at original q=3.” If you play the fragment alone, attacks used to reconstruct those held notes at its start do not increase the original version’s seven-attack count.
After checking the fragment, immediately restore all four bars of the chosen version, retaining the q=0 attacks, E from q=9 to 11 and the full rest from q=11 to 12. Check five items: voice identity; spelling and octave; separate bass and root entries; onsets and holds; closing rest. Completing only the middle six q cannot count as completing the whole phrase.
Optional piano route: first play only U’s V4, G4–F#4–F4–E4, lasting 3,3,3,2q followed by a 1q rest, without damper pedal. On an 88-key piano numbered from A0=1, the keys are 47,46,45,44; B’s V1 uses 23,22,21,20. Single-line playing can document its pitches, onsets and rest but cannot substitute for listening to the full four-voice comparison. If you do not play, leave performance observations blank.
After restoring the whole phrase, plan a delayed retest of the same version at the start of the next session, before practicing this task again. Keep the same route and visible prompts; retain the metronome setting if used. Hide answers, reconstruct or play the single line for all 12q, then check the five items and record the actual interval. This is a retention record for familiar material, not transfer to new material or proof of long-term effects.
Record prediction, actual errors, local repair, whole-phrase restoration, retest and listening impressions separately. If a new attack is mistaken for the bass, next inspect voices that remain held. If the rest is omitted, next check the timeline and restore the whole phrase again. “Cannot yet judge” is valid for an impression; do not mark unperformed tasks successful.
Record to complete
| Item | Initial value |
|---|---|
| Date, version, route, prompts, actual tempo | unknown |
| Prediction, errors and local repair | unknown |
| Five checks and whole-phrase restoration | unknown |
| Planned retest, actual interval, same-material result | unknown |
| Actual playing and impressions (optional) | unknown |
Source boundaries and this lesson’s added comparison
All 95 segments of the existing NC-075 ASR were read, and five stills at 65,112,134,150,168 seconds inspected. At 65, a red “chromatic descent” arrow and an eight-cell score with slash-chord labels are visible. Other selected frames show pointers at different positions, and 168 also contains a C→D7→G card. This checks visible text and layout, not every note of the entire source score.
Transcript candidates discuss a chromatically descending lowest voice, different chord choices over a given bass, and an example turning toward a dominant at the end. These attributions have not been checked against the audio. Ambiguous solfege and pitch names, chord-impression descriptions and a simplified minor adaptation do not determine this lesson’s pitches or general rules. Five images cannot verify source audio, continuous movement or exact ASR-to-image alignment.
The source topic is descending bass, not this lesson’s four-note upper-voice model on G minor. The U/B registers, Gm(maj7)-to-Gm6 voicings, same-set/different-bass comparison, alternative root reading, timelines and task sequence are original to this repository. They extend the source’s bass-and-chord-choice question into a checkable voicing task without claiming that the video taught either version.
The inventory duration is 365.568 seconds and the container duration 365.546 seconds, an unexplained difference of 0.022 seconds; neither is the duration actually heard or watched here. This candidate awaits independent review, with no source-audio audition, continuous-motion review, full-source-score audit, performance or learning-effect acceptance. Public files contain no source media, source images, full transcript or private paths. Limited ASR-text and selected-still coverage is not full audiovisual acceptance.
Related reading
Moving lines and chromatic approaches
Structured data
Source record
Fixed-fixture checks