{"slug": "yield-not", "name": "YIELD NOT", "meter_label": "6.5.6.5 D with Refrain", "meter": {"pattern": [6, 5, 6, 5], "doubled": true, "refrain": true, "alleluia": false, "irregular": false}, "composer": "horatio-r-palmer", "composer_date": "1868", "canonical_key": {"tonic": "bes", "mode": "major"}, "preferred_harmonization": "yield-not-1883", "provenance": {"layer": "tune", "source": "gospel-hymns-consolidated-1883", "source_kind": "hand_encoded", "pd_evidence": "Tune and harmony both by Horatio R. Palmer (1834-1907), written for his own words and first published in The Song Queen (Chicago, 1868). Read here off two dated pre-1931 printings of one set of plates: Gospel Hymns, Consolidated, Nos. 1-4 (Cincinnati: The John Church Co. / New York: Biglow & Main, 1884), no. 89, printed page 87 \u2014 archive.org gospelh00cinc, leaf 0087 \u2014 and the same collection as first issued in 1883, no. 89, printed page 87 \u2014 archive.org hymnsconso00blis, leaf 0093. Key of B-flat major, 6/8, no pickup, 24 bars: 16 of verse and 8 of chorus, laid out on the page in five systems. The plates letter no tune name; YIELD NOT is the name the hymnals and Hymnary carry for it. The soprano's first sixteen scale degrees, read notehead by notehead off the 1884 scan (3 3 4 3 2 2 1 1 2 1 7 7 6 6 7 1 in B-flat), are identical on the 1883 scan. A soprano interval fingerprint of this melody against all 285 corpus tunes returned a best match of 0.56 (WHITER THAN SNOW), far under the 0.80 duplicate threshold \u2014 the tune is new to the corpus. US public domain: published 1868, both witnesses pre-1931, composer died 1907.", "pd_basis": "pub_pre_1931", "pd_jurisdictions": {"US": true}}, "harmonization": {"slug": "yield-not-1883", "tune": "yield-not", "music_file": "yield-not-1883.ly", "harmonizer": "horatio-r-palmer", "arrangement_type": "harmonization", "canonical_key": {"tonic": "bes", "mode": "major"}, "time": "6/8", "pickup": null, "system_breaks": [12, 24, 36, 48, 62], "breaks_manual": true, "markings": [{"soprano_note": 49, "kind": "refrain"}], "final_barline": "|.", "amen": false, "verse_variable_melisma": false, "refrain": {"soprano_note_start": 49, "inline": true}, "source_line": "Setting: H. R. Palmer, 1868, as printed in Gospel Hymns Consolidated, 1883. Public domain.", "provenance": {"layer": "harmonization", "source": "gospel-hymns-consolidated-1883", "source_kind": "hand_encoded", "pd_evidence": "Horatio R. Palmer's own four-part setting of his own tune and words, first published in The Song Queen (Chicago: H. R. Palmer, 1868) and credited on the plates read here as \"H. R. PALMER, by per.\" Every chord of all four voices was measured notehead by notehead off the full-resolution page image of Gospel Hymns, Consolidated, Nos. 1 to 4 (Cincinnati: The John Church Co.; New York: Biglow & Main, 1884), no. 89, printed page 87 \u2014 archive.org gospelh00cinc, leaf 0087 \u2014 and then read END TO END A SECOND TIME, by the same rig but with its own binarisation and its own staff-line fits, on a different dated scan of the same plates: Gospel Hymns Consolidated (Biglow & Main / John Church, 1883), no. 89, printed page 87 \u2014 archive.org hymnsconso00blis, leaf 0093. Staff space 28.6 px (1884) and 28.9 px (1883), i.e. one diatonic step = 14.3 / 14.5 px. RESIDUALS, stated honestly: most heads land inside 0.15 step but NOT all \u2014 everything over 0.15 was re-read by eye at 3-10x on the 1884 plate before being accepted (the three cross-witness disagreements and both sharps were re-read on BOTH), and the worst accepted residual is 0.37 step, always on a low bass octave engraved on a ledger line below the staff, where the head fuses with its ledger (0.37 step is 5 px against a 14.3 px diatonic spacing, so the reading is not close to ambiguous, and each such head was confirmed by eye against its ledger). NO scaffold of any kind was used \u2014 no MIDI, no NoteWorthy file, no ABC, no Cyber Hymnal asset \u2014 so there is no circular check anywhere in this import. THE TWO WITNESSES DISAGREED IN EXACTLY THREE PLACES, all of them a blob-splitting artifact rather than a difference in the print, and all three were settled by eye: bar 8's second chord is D5 over B-flat4, not D5/A4 (zoomed to 7x on BOTH leaves); bar 14's second chord is B-flat4 over D4 (zoomed on the 1884 leaf, which is what the 1883 rig read cleanly); chorus bar 21's fourth chord is A4 over F4 (zoomed on the 1884 leaf, where the 1883 rig read it cleanly and the 1884 rig had refused it at residual 0.48). Everywhere else the two scans agree notehead for notehead. ACCIDENTALS WERE CHECKED SEPARATELY AND BY EYE at 9-10x on both leaves, because a notehead probe reads heads and not the signs in front of them: the automated figure above covers NO accidental. The page carries exactly two, and they are the same sign twice \u2014 a sharp at the head of bar 4 and a sharp at the head of bar 12, each centred on the F4 space, i.e. on the alto, giving f sharp' under the soprano's held A4 at \"sin\" and at \"sub-due\". Both were verified to be sharps and not naturals (two full-length verticals with two upward-slanting crossbars). Nothing else on the page is inflected; every other B and E takes the key signature's two flats. DEPARTURES FROM THE PAGE, stated in full: (1) THE VERSE'S BEAT-1 UNISON. In fourteen of the sixteen verse bars (not bars 8 and 16, which print two) the lower staff prints ONE notehead on beat 1 and TWO on beat 2, with a tie from the beat-1 head to whichever beat-2 head repeats its pitch, and NO rest is printed anywhere on either plate. This encoding reads that single head as tenor and bass in UNISON \u2014 the standard hymn-engraving convention \u2014 rather than as one voice resting unmarked for beat 1 of fourteen of the sixteen verse bars. That is an interpretation, not a transcription, and it is the only note in the file that the plate does not itself spell out. The same figure occurs at chorus bars 20 and 23 and is read the same way. (2) THE CASTING. The plates run five systems, two of whose breaks fall in the middle of a poetic line; this setting re-casts in six systems that break only at line ends. No note is changed by that. (3) The plates print no amen and none is added; the closing double barline is kept. Nothing else differs: the flagged (never beamed) eighths, the dotted quarters, the ties of bars 4/8/12/16/24, the restruck (UNtied) lower-staff repeats of chorus bars 18/20/22, and the printed seconds g-over-f in chorus bars 19 and 23 are all as engraved. US public domain: published 1868, both witnesses pre-1931, Palmer died 1907, and no modern arranger stands anywhere in the chain.", "pd_basis": "pub_pre_1931", "pd_jurisdictions": {"US": true}, "encoding_origin": "Hand-encoded from the 1884 Gospel Hymns Consolidated page image (archive.org gospelh00cinc, leaf 0087) and independently re-read on the 1883 printing of the same plates (archive.org hymnsconso00blis, leaf 0093). No MIDI, no NoteWorthy file and no Cyber Hymnal asset was used at any stage.", "encoding_license": "PD-dedication"}, "music_ly": "%% YIELD NOT (6.5.6.5 D with Refrain) \u2014 \"Yield not to temptation\"\n%% Horatio R. Palmer's own four-part setting of his own words, key B-flat major,\n%% 6/8, NO pickup, 24 bars: 16 of verse (bars 1-16) then 8 of chorus (17-24).\n%% RULES (CI-enforced): exactly the variables soprano/alto/tenor/bass/global;\n%% absolute octaves; each voice rhythmically complete; NO chords <>, NO <<>>,\n%% NO \\secondShift/\\unisonWholeShift/tightSpacing, NO \\break, NO \\bar, NO lyrics.\n%%\n%% SOURCE OF THE NOTES \u2014 TWO DATED PLATES, NO SCAFFOLD.\n%% Nothing was generated from a MIDI, a .nwc, an ABC file or any Cyber Hymnal\n%% asset. Every notehead was measured off page images:\n%%   (a) Gospel Hymns, Consolidated, Nos. 1 to 4 (Cincinnati: The John Church Co.;\n%%       New York: Biglow & Main, 1884), no. 89, printed page 87 \u2014\n%%       archive.org gospelh00cinc, leaf 0087, full-resolution jp2 (2348 x 3747);\n%%   (b) the same plates as first issued, Gospel Hymns Consolidated (Biglow &\n%%       Main / John Church, 1883), no. 89, printed page 87 \u2014\n%%       archive.org hymnsconso00blis, leaf 0093, full-resolution jp2 (2428 x 3756).\n%% Staff space measures 28.6 px on (a) and 28.9 px on (b) \u2014 one diatonic step is\n%% 14.3 / 14.5 px, comfortably above the ~12 px floor this campaign works to.\n%% Both scans were read END TO END by the same rig, each with its own local\n%% binarisation and its own staff-line fits: per notehead the five lines were\n%% re-found in a +-70 px window around that head's own x (from THIN ink runs\n%% only, so heads and accidentals cannot drag the fit), the five staff rows were\n%% erased and re-closed so a head touching a line is not pulled toward it, heads\n%% were isolated by elliptical erosion sized in staff spaces, and a merged\n%% soprano-over-alto blob was split at its two clean outer edges with the\n%% separation snapped to a whole staff step. Steps were taken against a table\n%% INTERPOLATED between the measured lines (1880s type is not evenly spaced).\n%% RESIDUALS, stated honestly rather than flatteringly: most heads land inside\n%% 0.15 step, but not all. Every reading over 0.15 was re-read BY EYE at 3-10x on\n%% the 1884 plate before it was accepted, and the three places where the two rigs\n%% disagreed (below) were re-read by eye on BOTH plates, as were both sharps.\n%% The worst accepted residual is 0.37 step, and it is always the same\n%% situation \u2014 a low bass octave engraved on a ledger\n%% line BELOW the staff (the d, e, f, and bes, heads of the verse), where the\n%% head fuses with its own ledger and the blob centre is pulled toward it. Those\n%% are nonetheless unambiguous: 0.37 step is 5 px where the next diatonic\n%% position is 14.3 px away, and each was confirmed by eye against its ledger.\n%%\n%% WHERE THE TWO RIGS DISAGREED (three places, all settled by eye):\n%%   * bar 8, second chord (\"win\", tied): (b) split the blob as D5/A4, (a) as\n%%     D5/B4. At 7x BOTH plates plainly print D5 on line 2 over B-flat4 on line 3\n%%     (this one was zoomed on both). D5/B-flat4 is encoded.\n%%   * bar 14, second chord (\"sus\"): (a) recovered only the lower head. At 3x the\n%%     1884 plate prints B-flat4 over D4, which is also what (b)'s rig read with a\n%%     clean residual. B-flat4/D4 is encoded.\n%%   * chorus bar 21, fourth chord (\"ing\"): (a) gave a contaminated blob\n%%     (residual 0.48); (b) read A4/F4 cleanly and the 1884 plate shows A4/F4 at\n%%     6x. A4/F4 is encoded.\n%% Everywhere else the two plates agree notehead for notehead.\n%%\n%% ACCIDENTALS WERE CHECKED SEPARATELY, BY EYE (the notehead probe reads HEADS,\n%% never the sharps and flats in front of them, so the automated figures above\n%% cover NO accidental). Every one of the ten staves of the 1884 leaf was swept\n%% by eye at 2.6-3.2x for accidental glyphs, and the two that turned up were then\n%% zoomed to 9-10x on BOTH plates. The whole page carries exactly TWO, and they\n%% are the same sign twice: a SHARP\n%% (two full-length verticals, two upward-slanting crossbars \u2014 not a natural)\n%% standing at the head of bar 4 and again at the head of bar 12, in each case\n%% centred on the F4 space, i.e. on the ALTO. So the alto sings f sharp' through\n%% bars 4 and 12 (the accidental holds for the bar; neither bar's second chord\n%% repeats it, and both chords of each bar are tied to each other anyway). No\n%% other accidental appears anywhere in any voice; every other B and E in this\n%% file takes the two flats of the key signature.\n%%\n%% RHYTHM, exactly as engraved.\n%%   VERSE, upper staff: odd-numbered bars are a dotted quarter then three\n%%   separately FLAGGED eighths (no beams anywhere on these plates); even bars are\n%%   two dotted quarters. Verse lines 2, 4, 6 and 8 end on a TIE across the two\n%%   dotted quarters of bars 4, 8, 12 and 16 (\"sin\", \"win\", \"sub-due\", \"through\").\n%%   VERSE, lower staff: two dotted quarters in every bar, so tenor and bass move\n%%   at half the soprano's rate through the verse.\n%%   CHORUS, all four voices together: odd bars are a QUARTER then four flagged\n%%   eighths; even bars are two dotted quarters. Only bar 24 ties.\n%%\n%% THE VERSE'S SINGLE NOTEHEAD ON BEAT 1 \u2014 THE ONE EDITORIAL READING IN THIS FILE.\n%% In FOURTEEN of the sixteen verse bars the lower staff prints ONE notehead on\n%% beat 1 and TWO on\n%% beat 2, and a tie joins the beat-1 head to whichever beat-2 head repeats its\n%% pitch (usually the upper; in bars 6 and 14 the LOWER). NO REST is printed\n%% anywhere on either plate. Read literally, one of the two lower voices would\n%% have to be silent for beat 1 of fourteen of the sixteen verse bars (bars 8 and\n%% 16 print TWO heads on beat 1) with no rest engraved,\n%% which these plates do not do; the standard hymn-engraving reading is that\n%% tenor and bass share one notehead when they are in UNISON. That is what is\n%% encoded: on beat 1 both lower voices take the printed pitch, and on beat 2 they\n%% split, the tied one holding. This is an interpretation of the page, not a\n%% transcription of it, and it is the only place in this file where the encoding\n%% supplies something the plate does not spell out. It puts tenor and bass in\n%% unison on the down-beat of bars 1-5, 7-13 and 15-16, and it makes the tenor of\n%% bars 6 and 14 leap b flat, -> b flat (the printed beat-1 note there sits an\n%% octave lower than in the other bars). Chorus bars 20 and 23 print the same\n%% one-head figure and are read the same way.\n%%\n%% SECONDS, NOT DIVISI. In chorus bars 19 and 23 the fourth eighth of the lower\n%% staff is engraved as two noteheads offset side by side on one stem, g over f \u2014\n%% a printed SECOND, tenor g and bass f. LilyPond's own collision resolver is left\n%% to place them; nothing is hand-shifted and nothing is divisi.\n%%\n%% UNISONS elsewhere: chorus bar 23's first three eighths print a single f in the\n%% lower staff (tenor and bass together), as does chorus bar 20's first dotted\n%% quarter (a single b flat). Again the default resolver is left alone.\n%%\n%% NO AMEN is printed. The hymn closes on a light-heavy double barline.\n\\version \"2.24.0\"\n\nglobal  = { \\key bes \\major \\time 6/8 }\n\nsoprano = {\n  d''4. d''8 ees''8 d''8 |\n  c''4. c''4. |\n  bes'4. bes'8 c''8 bes'8 |\n  a'4.~ a'4. |\n  g'4. g'8 a'8 bes'8 |\n  f'4. bes'4. |\n  c''4. c''8 bes'8 c''8 |\n  d''4.~ d''4. |\n  d''4. d''8 ees''8 d''8 |\n  c''4. c''4. |\n  bes'4. bes'8 c''8 bes'8 |\n  a'4.~ a'4. |\n  g'4. g'8 a'8 bes'8 |\n  f'4. bes'4. |\n  d''4. c''8 bes'8 c''8 |\n  bes'4.~ bes'4. |\n  f'4 f'8 f'8 g'8 f'8 |\n  f'4. bes'4. |\n  c''4 c''8 c''8 bes'8 c''8 |\n  d''4. bes'4. |\n  bes'4 bes'8 bes'8 a'8 g'8 |\n  f'4. bes'4. |\n  d''4 d''8 c''8 bes'8 c''8 |\n  bes'4.~ bes'4.\n}\n\nalto = {\n  bes'4. bes'8 c''8 bes'8 |\n  a'4. a'4. |\n  g'4. g'8 a'8 g'8 |\n  fis'4.~ fis'4. |\n  ees'4. ees'8 f'8 g'8 |\n  d'4. d'4. |\n  a'4. a'8 g'8 a'8 |\n  bes'4.~ bes'4. |\n  bes'4. bes'8 c''8 bes'8 |\n  a'4. a'4. |\n  g'4. g'8 a'8 g'8 |\n  fis'4.~ fis'4. |\n  ees'4. ees'8 f'8 g'8 |\n  d'4. d'4. |\n  f'4. ees'8 d'8 ees'8 |\n  d'4.~ d'4. |\n  d'4 d'8 d'8 ees'8 d'8 |\n  d'4. d'4. |\n  f'4 f'8 f'8 f'8 f'8 |\n  f'4. d'4. |\n  g'4 g'8 g'8 f'8 ees'8 |\n  d'4. d'4. |\n  f'4 f'8 ees'8 d'8 ees'8 |\n  d'4.~ d'4.\n}\n\ntenor = {\n  bes4.~ bes4. |\n  f4.~ f4. |\n  g4.~ g4. |\n  d4.~ d4. |\n  ees4.~ ees4. |\n  bes,4. bes4. |\n  f4.~ f4. |\n  f4.~ f4. |\n  bes4.~ bes4. |\n  f4.~ f4. |\n  g4.~ g4. |\n  d4.~ d4. |\n  ees4.~ ees4. |\n  bes,4. bes4. |\n  f4.~ f4. |\n  f4.~ f4. |\n  bes4 bes8 bes8 bes8 bes8 |\n  bes4. f4. |\n  a4 a8 a8 g8 a8 |\n  bes4. bes4. |\n  bes4 bes8 bes8 bes8 bes8 |\n  bes4. f4. |\n  f4 f8 f8 g8 a8 |\n  bes4.~ bes4.\n}\n\nbass = {\n  bes4. bes,4. |\n  f4. f,4. |\n  g4. g,4. |\n  d4. d,4. |\n  ees4. ees,4. |\n  bes,4.~ bes,4. |\n  f4. f,4. |\n  bes,4.~ bes,4. |\n  bes4. bes,4. |\n  f4. f,4. |\n  g4. g,4. |\n  d4. d,4. |\n  ees4. ees,4. |\n  bes,4.~ bes,4. |\n  f4. f,4. |\n  bes,4.~ bes,4. |\n  bes,4 bes,8 bes,8 bes,8 bes,8 |\n  bes,4. bes,4. |\n  f4 f8 f8 f8 f8 |\n  bes4. bes,4. |\n  ees4 ees8 ees8 ees8 ees8 |\n  bes,4. bes,4. |\n  f4 f8 f8 f8 f8 |\n  bes,4.~ bes,4.\n}\n"}, "harmonizations": [{"slug": "yield-not-1883", "tune": "yield-not", "music_file": "yield-not-1883.ly", "harmonizer": "horatio-r-palmer", "arrangement_type": "harmonization", "canonical_key": {"tonic": "bes", "mode": "major"}, "time": "6/8", "pickup": null, "system_breaks": [12, 24, 36, 48, 62], "breaks_manual": true, "markings": [{"soprano_note": 49, "kind": "refrain"}], "final_barline": "|.", "amen": false, "verse_variable_melisma": false, "refrain": {"soprano_note_start": 49, "inline": true}, "source_line": "Setting: H. R. Palmer, 1868, as printed in Gospel Hymns Consolidated, 1883. Public domain.", "provenance": {"layer": "harmonization", "source": "gospel-hymns-consolidated-1883", "source_kind": "hand_encoded", "pd_evidence": "Horatio R. Palmer's own four-part setting of his own tune and words, first published in The Song Queen (Chicago: H. R. Palmer, 1868) and credited on the plates read here as \"H. R. PALMER, by per.\" Every chord of all four voices was measured notehead by notehead off the full-resolution page image of Gospel Hymns, Consolidated, Nos. 1 to 4 (Cincinnati: The John Church Co.; New York: Biglow & Main, 1884), no. 89, printed page 87 \u2014 archive.org gospelh00cinc, leaf 0087 \u2014 and then read END TO END A SECOND TIME, by the same rig but with its own binarisation and its own staff-line fits, on a different dated scan of the same plates: Gospel Hymns Consolidated (Biglow & Main / John Church, 1883), no. 89, printed page 87 \u2014 archive.org hymnsconso00blis, leaf 0093. Staff space 28.6 px (1884) and 28.9 px (1883), i.e. one diatonic step = 14.3 / 14.5 px. RESIDUALS, stated honestly: most heads land inside 0.15 step but NOT all \u2014 everything over 0.15 was re-read by eye at 3-10x on the 1884 plate before being accepted (the three cross-witness disagreements and both sharps were re-read on BOTH), and the worst accepted residual is 0.37 step, always on a low bass octave engraved on a ledger line below the staff, where the head fuses with its ledger (0.37 step is 5 px against a 14.3 px diatonic spacing, so the reading is not close to ambiguous, and each such head was confirmed by eye against its ledger). NO scaffold of any kind was used \u2014 no MIDI, no NoteWorthy file, no ABC, no Cyber Hymnal asset \u2014 so there is no circular check anywhere in this import. THE TWO WITNESSES DISAGREED IN EXACTLY THREE PLACES, all of them a blob-splitting artifact rather than a difference in the print, and all three were settled by eye: bar 8's second chord is D5 over B-flat4, not D5/A4 (zoomed to 7x on BOTH leaves); bar 14's second chord is B-flat4 over D4 (zoomed on the 1884 leaf, which is what the 1883 rig read cleanly); chorus bar 21's fourth chord is A4 over F4 (zoomed on the 1884 leaf, where the 1883 rig read it cleanly and the 1884 rig had refused it at residual 0.48). Everywhere else the two scans agree notehead for notehead. ACCIDENTALS WERE CHECKED SEPARATELY AND BY EYE at 9-10x on both leaves, because a notehead probe reads heads and not the signs in front of them: the automated figure above covers NO accidental. The page carries exactly two, and they are the same sign twice \u2014 a sharp at the head of bar 4 and a sharp at the head of bar 12, each centred on the F4 space, i.e. on the alto, giving f sharp' under the soprano's held A4 at \"sin\" and at \"sub-due\". Both were verified to be sharps and not naturals (two full-length verticals with two upward-slanting crossbars). Nothing else on the page is inflected; every other B and E takes the key signature's two flats. DEPARTURES FROM THE PAGE, stated in full: (1) THE VERSE'S BEAT-1 UNISON. In fourteen of the sixteen verse bars (not bars 8 and 16, which print two) the lower staff prints ONE notehead on beat 1 and TWO on beat 2, with a tie from the beat-1 head to whichever beat-2 head repeats its pitch, and NO rest is printed anywhere on either plate. This encoding reads that single head as tenor and bass in UNISON \u2014 the standard hymn-engraving convention \u2014 rather than as one voice resting unmarked for beat 1 of fourteen of the sixteen verse bars. That is an interpretation, not a transcription, and it is the only note in the file that the plate does not itself spell out. The same figure occurs at chorus bars 20 and 23 and is read the same way. (2) THE CASTING. The plates run five systems, two of whose breaks fall in the middle of a poetic line; this setting re-casts in six systems that break only at line ends. No note is changed by that. (3) The plates print no amen and none is added; the closing double barline is kept. Nothing else differs: the flagged (never beamed) eighths, the dotted quarters, the ties of bars 4/8/12/16/24, the restruck (UNtied) lower-staff repeats of chorus bars 18/20/22, and the printed seconds g-over-f in chorus bars 19 and 23 are all as engraved. US public domain: published 1868, both witnesses pre-1931, Palmer died 1907, and no modern arranger stands anywhere in the chain.", "pd_basis": "pub_pre_1931", "pd_jurisdictions": {"US": true}, "encoding_origin": "Hand-encoded from the 1884 Gospel Hymns Consolidated page image (archive.org gospelh00cinc, leaf 0087) and independently re-read on the 1883 printing of the same plates (archive.org hymnsconso00blis, leaf 0093). No MIDI, no NoteWorthy file and no Cyber Hymnal asset was used at any stage.", "encoding_license": "PD-dedication"}, "music_ly": "%% YIELD NOT (6.5.6.5 D with Refrain) \u2014 \"Yield not to temptation\"\n%% Horatio R. Palmer's own four-part setting of his own words, key B-flat major,\n%% 6/8, NO pickup, 24 bars: 16 of verse (bars 1-16) then 8 of chorus (17-24).\n%% RULES (CI-enforced): exactly the variables soprano/alto/tenor/bass/global;\n%% absolute octaves; each voice rhythmically complete; NO chords <>, NO <<>>,\n%% NO \\secondShift/\\unisonWholeShift/tightSpacing, NO \\break, NO \\bar, NO lyrics.\n%%\n%% SOURCE OF THE NOTES \u2014 TWO DATED PLATES, NO SCAFFOLD.\n%% Nothing was generated from a MIDI, a .nwc, an ABC file or any Cyber Hymnal\n%% asset. Every notehead was measured off page images:\n%%   (a) Gospel Hymns, Consolidated, Nos. 1 to 4 (Cincinnati: The John Church Co.;\n%%       New York: Biglow & Main, 1884), no. 89, printed page 87 \u2014\n%%       archive.org gospelh00cinc, leaf 0087, full-resolution jp2 (2348 x 3747);\n%%   (b) the same plates as first issued, Gospel Hymns Consolidated (Biglow &\n%%       Main / John Church, 1883), no. 89, printed page 87 \u2014\n%%       archive.org hymnsconso00blis, leaf 0093, full-resolution jp2 (2428 x 3756).\n%% Staff space measures 28.6 px on (a) and 28.9 px on (b) \u2014 one diatonic step is\n%% 14.3 / 14.5 px, comfortably above the ~12 px floor this campaign works to.\n%% Both scans were read END TO END by the same rig, each with its own local\n%% binarisation and its own staff-line fits: per notehead the five lines were\n%% re-found in a +-70 px window around that head's own x (from THIN ink runs\n%% only, so heads and accidentals cannot drag the fit), the five staff rows were\n%% erased and re-closed so a head touching a line is not pulled toward it, heads\n%% were isolated by elliptical erosion sized in staff spaces, and a merged\n%% soprano-over-alto blob was split at its two clean outer edges with the\n%% separation snapped to a whole staff step. Steps were taken against a table\n%% INTERPOLATED between the measured lines (1880s type is not evenly spaced).\n%% RESIDUALS, stated honestly rather than flatteringly: most heads land inside\n%% 0.15 step, but not all. Every reading over 0.15 was re-read BY EYE at 3-10x on\n%% the 1884 plate before it was accepted, and the three places where the two rigs\n%% disagreed (below) were re-read by eye on BOTH plates, as were both sharps.\n%% The worst accepted residual is 0.37 step, and it is always the same\n%% situation \u2014 a low bass octave engraved on a ledger\n%% line BELOW the staff (the d, e, f, and bes, heads of the verse), where the\n%% head fuses with its own ledger and the blob centre is pulled toward it. Those\n%% are nonetheless unambiguous: 0.37 step is 5 px where the next diatonic\n%% position is 14.3 px away, and each was confirmed by eye against its ledger.\n%%\n%% WHERE THE TWO RIGS DISAGREED (three places, all settled by eye):\n%%   * bar 8, second chord (\"win\", tied): (b) split the blob as D5/A4, (a) as\n%%     D5/B4. At 7x BOTH plates plainly print D5 on line 2 over B-flat4 on line 3\n%%     (this one was zoomed on both). D5/B-flat4 is encoded.\n%%   * bar 14, second chord (\"sus\"): (a) recovered only the lower head. At 3x the\n%%     1884 plate prints B-flat4 over D4, which is also what (b)'s rig read with a\n%%     clean residual. B-flat4/D4 is encoded.\n%%   * chorus bar 21, fourth chord (\"ing\"): (a) gave a contaminated blob\n%%     (residual 0.48); (b) read A4/F4 cleanly and the 1884 plate shows A4/F4 at\n%%     6x. A4/F4 is encoded.\n%% Everywhere else the two plates agree notehead for notehead.\n%%\n%% ACCIDENTALS WERE CHECKED SEPARATELY, BY EYE (the notehead probe reads HEADS,\n%% never the sharps and flats in front of them, so the automated figures above\n%% cover NO accidental). Every one of the ten staves of the 1884 leaf was swept\n%% by eye at 2.6-3.2x for accidental glyphs, and the two that turned up were then\n%% zoomed to 9-10x on BOTH plates. The whole page carries exactly TWO, and they\n%% are the same sign twice: a SHARP\n%% (two full-length verticals, two upward-slanting crossbars \u2014 not a natural)\n%% standing at the head of bar 4 and again at the head of bar 12, in each case\n%% centred on the F4 space, i.e. on the ALTO. So the alto sings f sharp' through\n%% bars 4 and 12 (the accidental holds for the bar; neither bar's second chord\n%% repeats it, and both chords of each bar are tied to each other anyway). No\n%% other accidental appears anywhere in any voice; every other B and E in this\n%% file takes the two flats of the key signature.\n%%\n%% RHYTHM, exactly as engraved.\n%%   VERSE, upper staff: odd-numbered bars are a dotted quarter then three\n%%   separately FLAGGED eighths (no beams anywhere on these plates); even bars are\n%%   two dotted quarters. Verse lines 2, 4, 6 and 8 end on a TIE across the two\n%%   dotted quarters of bars 4, 8, 12 and 16 (\"sin\", \"win\", \"sub-due\", \"through\").\n%%   VERSE, lower staff: two dotted quarters in every bar, so tenor and bass move\n%%   at half the soprano's rate through the verse.\n%%   CHORUS, all four voices together: odd bars are a QUARTER then four flagged\n%%   eighths; even bars are two dotted quarters. Only bar 24 ties.\n%%\n%% THE VERSE'S SINGLE NOTEHEAD ON BEAT 1 \u2014 THE ONE EDITORIAL READING IN THIS FILE.\n%% In FOURTEEN of the sixteen verse bars the lower staff prints ONE notehead on\n%% beat 1 and TWO on\n%% beat 2, and a tie joins the beat-1 head to whichever beat-2 head repeats its\n%% pitch (usually the upper; in bars 6 and 14 the LOWER). NO REST is printed\n%% anywhere on either plate. Read literally, one of the two lower voices would\n%% have to be silent for beat 1 of fourteen of the sixteen verse bars (bars 8 and\n%% 16 print TWO heads on beat 1) with no rest engraved,\n%% which these plates do not do; the standard hymn-engraving reading is that\n%% tenor and bass share one notehead when they are in UNISON. That is what is\n%% encoded: on beat 1 both lower voices take the printed pitch, and on beat 2 they\n%% split, the tied one holding. This is an interpretation of the page, not a\n%% transcription of it, and it is the only place in this file where the encoding\n%% supplies something the plate does not spell out. It puts tenor and bass in\n%% unison on the down-beat of bars 1-5, 7-13 and 15-16, and it makes the tenor of\n%% bars 6 and 14 leap b flat, -> b flat (the printed beat-1 note there sits an\n%% octave lower than in the other bars). Chorus bars 20 and 23 print the same\n%% one-head figure and are read the same way.\n%%\n%% SECONDS, NOT DIVISI. In chorus bars 19 and 23 the fourth eighth of the lower\n%% staff is engraved as two noteheads offset side by side on one stem, g over f \u2014\n%% a printed SECOND, tenor g and bass f. LilyPond's own collision resolver is left\n%% to place them; nothing is hand-shifted and nothing is divisi.\n%%\n%% UNISONS elsewhere: chorus bar 23's first three eighths print a single f in the\n%% lower staff (tenor and bass together), as does chorus bar 20's first dotted\n%% quarter (a single b flat). Again the default resolver is left alone.\n%%\n%% NO AMEN is printed. The hymn closes on a light-heavy double barline.\n\\version \"2.24.0\"\n\nglobal  = { \\key bes \\major \\time 6/8 }\n\nsoprano = {\n  d''4. d''8 ees''8 d''8 |\n  c''4. c''4. |\n  bes'4. bes'8 c''8 bes'8 |\n  a'4.~ a'4. |\n  g'4. g'8 a'8 bes'8 |\n  f'4. bes'4. |\n  c''4. c''8 bes'8 c''8 |\n  d''4.~ d''4. |\n  d''4. d''8 ees''8 d''8 |\n  c''4. c''4. |\n  bes'4. bes'8 c''8 bes'8 |\n  a'4.~ a'4. |\n  g'4. g'8 a'8 bes'8 |\n  f'4. bes'4. |\n  d''4. c''8 bes'8 c''8 |\n  bes'4.~ bes'4. |\n  f'4 f'8 f'8 g'8 f'8 |\n  f'4. bes'4. |\n  c''4 c''8 c''8 bes'8 c''8 |\n  d''4. bes'4. |\n  bes'4 bes'8 bes'8 a'8 g'8 |\n  f'4. bes'4. |\n  d''4 d''8 c''8 bes'8 c''8 |\n  bes'4.~ bes'4.\n}\n\nalto = {\n  bes'4. bes'8 c''8 bes'8 |\n  a'4. a'4. |\n  g'4. g'8 a'8 g'8 |\n  fis'4.~ fis'4. |\n  ees'4. ees'8 f'8 g'8 |\n  d'4. d'4. |\n  a'4. a'8 g'8 a'8 |\n  bes'4.~ bes'4. |\n  bes'4. bes'8 c''8 bes'8 |\n  a'4. a'4. |\n  g'4. g'8 a'8 g'8 |\n  fis'4.~ fis'4. |\n  ees'4. ees'8 f'8 g'8 |\n  d'4. d'4. |\n  f'4. ees'8 d'8 ees'8 |\n  d'4.~ d'4. |\n  d'4 d'8 d'8 ees'8 d'8 |\n  d'4. d'4. |\n  f'4 f'8 f'8 f'8 f'8 |\n  f'4. d'4. |\n  g'4 g'8 g'8 f'8 ees'8 |\n  d'4. d'4. |\n  f'4 f'8 ees'8 d'8 ees'8 |\n  d'4.~ d'4.\n}\n\ntenor = {\n  bes4.~ bes4. |\n  f4.~ f4. |\n  g4.~ g4. |\n  d4.~ d4. |\n  ees4.~ ees4. |\n  bes,4. bes4. |\n  f4.~ f4. |\n  f4.~ f4. |\n  bes4.~ bes4. |\n  f4.~ f4. |\n  g4.~ g4. |\n  d4.~ d4. |\n  ees4.~ ees4. |\n  bes,4. bes4. |\n  f4.~ f4. |\n  f4.~ f4. |\n  bes4 bes8 bes8 bes8 bes8 |\n  bes4. f4. |\n  a4 a8 a8 g8 a8 |\n  bes4. bes4. |\n  bes4 bes8 bes8 bes8 bes8 |\n  bes4. f4. |\n  f4 f8 f8 g8 a8 |\n  bes4.~ bes4.\n}\n\nbass = {\n  bes4. bes,4. |\n  f4. f,4. |\n  g4. g,4. |\n  d4. d,4. |\n  ees4. ees,4. |\n  bes,4.~ bes,4. |\n  f4. f,4. |\n  bes,4.~ bes,4. |\n  bes4. bes,4. |\n  f4. f,4. |\n  g4. g,4. |\n  d4. d,4. |\n  ees4. ees,4. |\n  bes,4.~ bes,4. |\n  f4. f,4. |\n  bes,4.~ bes,4. |\n  bes,4 bes,8 bes,8 bes,8 bes,8 |\n  bes,4. bes,4. |\n  f4 f8 f8 f8 f8 |\n  bes4. bes,4. |\n  ees4 ees8 ees8 ees8 ees8 |\n  bes,4. bes,4. |\n  f4 f8 f8 f8 f8 |\n  bes,4.~ bes,4.\n}\n", "preferred": true}], "license": "The hymn texts and tunes are in the public domain; per-layer provenance is recorded on each record."}