{"slug": "near-the-cross", "name": "NEAR THE CROSS", "meter_label": "7.6.7.6 with Refrain", "meter": {"pattern": [7, 6, 7, 6], "doubled": false, "refrain": true, "alleluia": false, "irregular": false}, "composer": "william-howard-doane", "composer_date": "1869", "canonical_key": {"tonic": "f", "mode": "major"}, "preferred_harmonization": "near-the-cross-1875", "provenance": {"layer": "tune", "source": "gospel-hymns-sacred-songs-1875", "source_kind": "hand_encoded", "pd_basis": "pub_pre_1931", "pd_evidence": "Tune by William Howard Doane (1832-1915), printed over \"W. H. DOANE, by per.\" on both dated printings read for this record: Gospel Hymns and Sacred Songs (New York: Biglow & Main, 1875), No. 45, printed page 46 \u2014 archive.org gospelhymnssacre00blis_0, leaf n49 \u2014 and Gospel Hymns Consolidated, Nos. 1 to 4 (New York: Biglow & Main; Cincinnati: John Church, 1883), No. 45, printed page 48 \u2014 archive.org hymnsconso00blis, leaf n49. Neither page prints a tune name; the name NEAR THE CROSS is the page's own title \"Near the Cross.\" and is the name the tune carries in later hymnals. US public domain: both printings read are pre-1931, the earlier being 1875, and the composer died in 1915. The tune is conventionally dated 1869 (its first appearance was in Doane's collections of 1869-1870); no printing earlier than the 1875 one cited here was opened for this record, so 1869 is given as the conventional date, not as something verified from a page. DUPLICATE CHECK: tools/tune_fingerprint.py scores this soprano 0.811 against corpus tune BETHANY, above the 0.80 flag, but they are NOT the same tune \u2014 the meters differ (6.4.6.4.6.6.6.4 against 7.6.7.6 with Refrain), the lengths differ (47 soprano notes against 51), and the openings differ (BETHANY B-A-G-G-E-E-D descending; NEAR THE CROSS A-B flat-A-G-F-D-D). No other corpus tune scores above 0.59.", "pd_jurisdictions": {"US": true}}, "harmonization": {"slug": "near-the-cross-1875", "tune": "near-the-cross", "music_file": "near-the-cross-1875.ly", "canonical_key": {"tonic": "f", "mode": "major"}, "time": "6/8", "pickup": null, "system_breaks": [13, 26, 38], "breaks_manual": true, "markings": [{"soprano_note": 27, "kind": "refrain"}], "final_barline": "|.", "amen": false, "verse_variable_melisma": false, "refrain": {"soprano_note_start": 27}, "provenance": {"layer": "harmonization", "source": "gospel-hymns-sacred-songs-1875", "source_kind": "hand_encoded", "pd_basis": "pub_pre_1931", "pd_evidence": "William H. Doane's four-part setting of his own tune, credited \"W. H. DOANE, by per.\" on the page it was read from: Gospel Hymns and Sacred Songs (New York: Biglow & Main, 1875), No. 45, printed page 46 \u2014 archive.org gospelhymnssacre00blis_0, leaf n49 \u2014 in F major and 6/8 with no pickup, sixteen bars cast as four systems of four. US public domain: the earlier printing read is 1875, well before 1931, and no modern arranger stands anywhere in the chain. TWO DATED PRINTINGS WERE READ AND REQUIRED TO AGREE, each with its own binarisation and its own staff-line fits: the 1875 page above (2244 px across, staff space 28.2 px, ink taken at 0.70 of the local paper level) and Gospel Hymns Consolidated, Nos. 1 to 4 (New York: Biglow & Main; Cincinnati: John Church, 1883), No. 45, printed page 46 \u2014 archive.org hymnsconso00blis, leaf n49 (2428 px across, staff space 28.9 px, ink taken at 0.82 of the local paper level). They are not one plate reprinted: the lyric type is reset between them (1875 \"fountain\" and \"Shed its beams\", 1883 \"fount-ain\" and \"Sheds its beams\") and the note spacing differs, so they are two independent engravings of the same setting. THE NOTEHEADS WERE MEASURED, NOT EYEBALLED. Per scan: adaptive binarisation against a local paper level from a 41x41 grey dilation plus a 41x41 box filter, then a 5x5 binary closing to seal the pale gaps the plate's ink leaves inside a head; staff-line removal by vertical run length; stems, barlines and ledger stubs removed by a horizontal erosion of 0.30 staff space; then each chord column's vertical ink profile fitted against a model of one or two notehead boxes placed at EXACT diatonic positions on that column's own freshly re-fitted five staff lines, the lines re-found from staff-line pixels alone in a +-2.6-staff-space window around the column's own x and fitted as a rigid equally spaced grid, so page curvature and heads sitting on lines cannot drag a reading. Of the 157 columns the two scans report, 110 agree exactly; the other 47 are the clef, key- and time-signature glyphs, the eighth rests, and eight note columns, and every one of those eight was settled BY EYE at 3x to 8x on crops with the fitted grid drawn over them. THREE MEASUREMENT ERRORS WERE FOUND AND CORRECTED THAT WAY, and they are named so the next reader can look for them: an eighth's DOWN-STEM FLAG can score as a second notehead (three columns first read \"F3 over A2\" are tenor/bass unisons on F3 \u2014 bar 13 note 2, bar 3 note 2, bar 11 note 2); the tenor rises to F4 on two ledger lines above the bass staff through the chorus, and a first pass that capped the search at D4 mis-read all five of those columns; and the tenor of bar 2 note 1 measured half-way between A3 and B flat 3 on the 1883 scan, where the 1875 scan prints the head plainly bisected by the A3 line and settles it as A3. NO SCAFFOLD WAS USED AND NOTHING WAS CHECKED AGAINST ONE: no MIDI, no Cyber Hymnal asset and no existing encoding was opened at any point. HOLLOW NOTEHEADS: there are none. Every notehead in this hymn is filled \u2014 a quarter, a dotted quarter or an eighth \u2014 and the hymn contains no half or whole note at all. Exactly one head reads hollow on the 1883 scan, the bass F3 of bar 9's third chord, and it is ink dropout: the 1875 plate prints it solid, and a dotted half could not fit a 6/8 bar that has already spent a quarter and an eighth. ACCIDENTALS WERE SWEPT SEPARATELY, BY EYE, AND THE AUTOMATED FIGURE ABOVE NEVER COVERED THEM \u2014 a notehead probe reads staff positions, not the sharps and flats in front of them. All four systems of both leaves were examined for accidental glyphs at system-level magnification and again on every 3x-8x crop. There is exactly ONE printed accidental in the hymn: a sharp before the LOWER (alto) head of the fourth chord of bar 13, on the syllable \"-tured\", centred on the first ledger line below the treble staff, making that note C sharp 4 and the chord G4 / C sharp 4 / E3 / A2, an A7. Both plates print it, in the same place. There are no naturals, no other sharps, and no flats beyond the one-flat key signature. DEPARTURES FROM THE PAGE, in full: (1) the plates draw a light double bar at each system end and a heavy double bar at the close, and only the closing barline is carried (`final_barline` \"|.\"), the engine drawing ordinary barlines and a system break at the same soprano ordinals; (2) nothing else \u2014 every pitch and every duration below is the plates', the phrase-end eighth rests of bars 2, 4, 6, 8, 14 and 16 are printed and are reproduced rather than collapsed (this is 6/8, not the 4/4 case the corpus collapses), the alto's two beamed eighths in bar 6 against a held quarter in the other three voices are the plates' own and are the only non-homorhythmic moment in the hymn, and the tenor/bass unisons of bars 3, 7 and 13 carry both voices on one printed head exactly as engraved.", "pd_jurisdictions": {"US": true}, "encoding_origin": "Hand-encoded from the page images of Gospel Hymns and Sacred Songs 1875, No. 45, p. 46 \u2014 archive.org gospelhymnssacre00blis_0 leaf n49 \u2014 collated chord by chord against the independent 1883 engraving in Gospel Hymns Consolidated Nos. 1 to 4, No. 45, p. 48 (archive.org hymnsconso00blis, leaf n49). No MIDI, no Cyber Hymnal asset and no existing encoding was used at any point.", "encoding_license": "PD-dedication"}, "source_line": "Setting: William H. Doane, as printed in Gospel Hymns and Sacred Songs, 1875. Public domain.", "music_ly": "%% NEAR THE CROSS (7.6.7.6 with Refrain) \u2014 \"Jesus, Keep Me Near the Cross\"\n%% William H. Doane's four-part setting, key F major, 6/8, no pickup, 16 bars.\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. Read off TWO dated pre-1931 printings of this setting,\n%% each scanned separately and each read with its own binarisation and its own\n%% staff-line fits:\n%%   (A) Gospel Hymns and Sacred Songs (New York: Biglow & Main, 1875), No. 45,\n%%       printed page 46 \u2014 archive.org gospelhymnssacre00blis_0, leaf n49\n%%       (2244 px across, staff space 28.2 px = 14.1 px per diatonic step);\n%%   (B) Gospel Hymns Consolidated, Nos. 1 to 4 (New York: Biglow & Main;\n%%       Cincinnati: John Church, 1883), No. 45, printed page 46 \u2014 archive.org\n%%       hymnsconso00blis, leaf n49 (2428 px across, staff space 28.9 px =\n%%       14.4 px per diatonic step).\n%% Both pages are headed \"Near the Cross.\" over the motto \"Peace through the\n%% blood of His cross.\"\u2014Coll. 1: 29, and credit \"FANNY J. CROSBY.\" at the left\n%% and \"W. H. DOANE, by per.\" at the right. They are NOT the same plate: the\n%% lyric type is reset (1875 \"fountain\" / \"Shed its beams\", 1883 \"fount-ain\" /\n%% \"Sheds its beams\") and the note spacing differs, so they are two independent\n%% engravings of one setting, which is why they were required to agree.\n%%\n%% HOW THE NOTEHEADS WERE READ. Per scan: adaptive binarisation (ink darker than\n%% a fraction of the LOCAL paper level taken from a 41x41 grey dilation plus a\n%% 41x41 box filter \u2014 0.82 on the 1883 scan, 0.70 on the darker 1875 scan, each\n%% chosen from that scan's own row-profile behaviour), then a 5x5 binary CLOSING\n%% to seal the pale gaps the plate's ink leaves inside a head; STAFF-LINE REMOVAL\n%% by vertical run length (an ink run no longer than 0.30 staff space sitting on\n%% a fitted line is deleted); stems, barlines and ledger stubs removed by a\n%% horizontal erosion of 0.30 staff space; then, for every surviving chord\n%% column, the observed vertical ink profile was FITTED against a model built\n%% from that column's own freshly re-fitted five staff lines \u2014 one or two\n%% notehead boxes 0.76 staff space tall placed at exact diatonic positions, the\n%% hypothesis with the smallest squared residual winning. The five lines are\n%% re-found in a +-2.6-staff-space window around each column's own x, from the\n%% staff-line pixels alone, and fitted as a RIGID equally spaced grid, so neither\n%% the page's curvature nor a notehead sitting on a line can drag a reading.\n%% 110 of the 157 columns the two scans report agree exactly; the remainder are\n%% the clef/key/time glyphs, the eighth rests, and eight note columns that were\n%% then settled BY EYE at 3x-8x on crops with the fitted grid drawn over them.\n%%\n%% WHERE THE MEASUREMENT WAS WRONG, AND HOW IT WAS CAUGHT. (1) An eighth note's\n%% DOWN-STEM FLAG can score as a second notehead: three columns first read\n%% \"F3 over A2\" are in fact a tenor/bass unison on F3 (system 4, bar 1, note 2;\n%% system 1, bar 3, note 2; system 3, bar 3, note 2). (2) The tenor rises to F4\n%% through the chorus, above the bass staff on two ledger lines; a first pass\n%% that capped the search at D4 mis-read all five of those columns. (3) The\n%% tenor at system 1, bar 2, note 1 measured half-way between A3 and B-flat 3 on\n%% the 1883 scan; the 1875 scan prints the head plainly bisected by the A3 line\n%% and settles it as A3. Each of the three was then confirmed by eye on a\n%% grid-overlay crop.\n%%\n%% NO HALF OR WHOLE NOTES. Every notehead in this hymn is filled: a quarter, a\n%% dotted quarter or an eighth. ONE head reads hollow on the 1883 scan \u2014 the bass\n%% F3 of the chorus's third chord (bar 9) \u2014 and it is ink dropout, not an outline\n%% head: the 1875 plate prints it solid, and a dotted half would not fit a 6/8 bar\n%% that already spends a quarter and an eighth before it.\n%%\n%% THE ONE PRINTED ACCIDENTAL. Both plates engrave a SHARP, and only one, in the\n%% whole hymn: before the LOWER (alto) head of the fourth chord of bar 13\n%% (\"-tured\"), centred on the first ledger line below the treble staff, making\n%% that note C sharp 4. The chord is then G4 / C#4 / E3 / A2 \u2014 an A7. Every other\n%% column on both leaves was swept separately by eye for accidentals at\n%% system-level magnification and at 3x-8x on the crops; there are no naturals,\n%% no other sharps and no flats beyond the one-flat key signature.\n%%\n%% THE ONE PLACE THE FOUR VOICES ARE NOT HOMORHYTHMIC. In bar 6 the soprano,\n%% tenor and bass hold a quarter on the third beat-group while the ALTO moves in\n%% two beamed eighths, D4 then B-flat 3. Everywhere else all four voices restrike\n%% together, exactly as both plates print.\n%%\n%% THE PHRASE-END EIGHTH RESTS ARE THE PAGE'S. Bars 2, 6, 10 (system 2's fourth\n%% bar), 14 and 16 close with a quarter plus an eighth rest in all four voices;\n%% bars 4 and 8 likewise. They are reproduced, not collapsed into a longer note:\n%% this is 6/8, not the 4/4 case the corpus collapses.\n%%\n%% STRUCTURE. Bars 1-8 are the verse (two systems of four bars, 13 soprano\n%% noteheads each); bars 9-16 are the CHORUS, which the plates label \"CHORUS.\"\n%% over bar 9 (12 soprano noteheads, then 13). 51 soprano noteheads in all \u2014\n%% 26 in the verse and 25 in the refrain \u2014 and the setting is strictly syllabic:\n%% there is not one slur or tie anywhere on either plate.\n\\version \"2.24.0\"\n\nglobal  = { \\key f \\major \\time 6/8 }\n\nsoprano = {\n  a'4 bes'8 a'4 g'8 |\n  f'4 d'8 d'4 r8 |\n  c'4 f'8 f'4 a'8 |\n  a'4. g'4 r8 |\n  a'4 bes'8 a'4 g'8 |\n  f'4 d'8 d'4 r8 |\n  c'4 f'8 f'4 e'8 |\n  g'4. f'4 r8 |\n  a'4 c''8 c''4. |\n  bes'4 d''8 d''4. |\n  c''4 d''8 c''4 a'8 |\n  a'4. g'4. |\n  a'4 bes'8 a'4 g'8 |\n  f'4 d'8 d'4 r8 |\n  c'4 f'8 f'4 e'8 |\n  g'4. f'4 r8\n}\n\nalto = {\n  f'4 f'8 f'4 e'8 |\n  d'4 d'8 bes4 r8 |\n  c'4 c'8 c'4 f'8 |\n  f'4. e'4 r8 |\n  f'4 f'8 f'4 e'8 |\n  d'4 d'8 d'8 bes8 r8 |\n  a4 c'8 c'4 c'8 |\n  c'4. c'4 r8 |\n  f'4 f'8 f'4. |\n  f'4 bes'8 bes'4. |\n  a'4 bes'8 a'4 f'8 |\n  f'4. e'4. |\n  c'4 d'8 c'4 cis'8 |\n  d'4 bes8 bes4 r8 |\n  a4 c'8 c'4 c'8 |\n  c'4. c'4 r8\n}\n\ntenor = {\n  c'4 d'8 c'4 bes8 |\n  a4 bes8 f4 r8 |\n  a4 a8 c'4 c'8 |\n  c'4. c'4 r8 |\n  c'4 d'8 c'4 bes8 |\n  a4 bes8 bes4 r8 |\n  f4 a8 a4 g8 |\n  bes4. a4 r8 |\n  c'4 a8 a4. |\n  d'4 f'8 f'4. |\n  f'4 f'8 f'4 c'8 |\n  c'4. c'4. |\n  f4 f8 f4 e8 |\n  d4 f8 f4 r8 |\n  f4 a8 a4 g8 |\n  bes4. a4 r8\n}\n\nbass = {\n  f4 f8 f4 c8 |\n  d4 bes,8 bes,4 r8 |\n  f4 f8 a4 f8 |\n  c4. c4 r8 |\n  f4 f8 f4 c8 |\n  d4 bes,8 bes,4 r8 |\n  c4 c8 c4 c8 |\n  c4. f4 r8 |\n  f4 f8 f4. |\n  bes4 bes8 bes4. |\n  f4 f8 f4 f8 |\n  c4. c4. |\n  f4 f8 f4 a,8 |\n  bes,4 bes,8 bes,4 r8 |\n  c4 c8 c4 c8 |\n  c4. f4 r8\n}\n"}, "harmonizations": [{"slug": "near-the-cross-1875", "tune": "near-the-cross", "music_file": "near-the-cross-1875.ly", "canonical_key": {"tonic": "f", "mode": "major"}, "time": "6/8", "pickup": null, "system_breaks": [13, 26, 38], "breaks_manual": true, "markings": [{"soprano_note": 27, "kind": "refrain"}], "final_barline": "|.", "amen": false, "verse_variable_melisma": false, "refrain": {"soprano_note_start": 27}, "provenance": {"layer": "harmonization", "source": "gospel-hymns-sacred-songs-1875", "source_kind": "hand_encoded", "pd_basis": "pub_pre_1931", "pd_evidence": "William H. Doane's four-part setting of his own tune, credited \"W. H. DOANE, by per.\" on the page it was read from: Gospel Hymns and Sacred Songs (New York: Biglow & Main, 1875), No. 45, printed page 46 \u2014 archive.org gospelhymnssacre00blis_0, leaf n49 \u2014 in F major and 6/8 with no pickup, sixteen bars cast as four systems of four. US public domain: the earlier printing read is 1875, well before 1931, and no modern arranger stands anywhere in the chain. TWO DATED PRINTINGS WERE READ AND REQUIRED TO AGREE, each with its own binarisation and its own staff-line fits: the 1875 page above (2244 px across, staff space 28.2 px, ink taken at 0.70 of the local paper level) and Gospel Hymns Consolidated, Nos. 1 to 4 (New York: Biglow & Main; Cincinnati: John Church, 1883), No. 45, printed page 46 \u2014 archive.org hymnsconso00blis, leaf n49 (2428 px across, staff space 28.9 px, ink taken at 0.82 of the local paper level). They are not one plate reprinted: the lyric type is reset between them (1875 \"fountain\" and \"Shed its beams\", 1883 \"fount-ain\" and \"Sheds its beams\") and the note spacing differs, so they are two independent engravings of the same setting. THE NOTEHEADS WERE MEASURED, NOT EYEBALLED. Per scan: adaptive binarisation against a local paper level from a 41x41 grey dilation plus a 41x41 box filter, then a 5x5 binary closing to seal the pale gaps the plate's ink leaves inside a head; staff-line removal by vertical run length; stems, barlines and ledger stubs removed by a horizontal erosion of 0.30 staff space; then each chord column's vertical ink profile fitted against a model of one or two notehead boxes placed at EXACT diatonic positions on that column's own freshly re-fitted five staff lines, the lines re-found from staff-line pixels alone in a +-2.6-staff-space window around the column's own x and fitted as a rigid equally spaced grid, so page curvature and heads sitting on lines cannot drag a reading. Of the 157 columns the two scans report, 110 agree exactly; the other 47 are the clef, key- and time-signature glyphs, the eighth rests, and eight note columns, and every one of those eight was settled BY EYE at 3x to 8x on crops with the fitted grid drawn over them. THREE MEASUREMENT ERRORS WERE FOUND AND CORRECTED THAT WAY, and they are named so the next reader can look for them: an eighth's DOWN-STEM FLAG can score as a second notehead (three columns first read \"F3 over A2\" are tenor/bass unisons on F3 \u2014 bar 13 note 2, bar 3 note 2, bar 11 note 2); the tenor rises to F4 on two ledger lines above the bass staff through the chorus, and a first pass that capped the search at D4 mis-read all five of those columns; and the tenor of bar 2 note 1 measured half-way between A3 and B flat 3 on the 1883 scan, where the 1875 scan prints the head plainly bisected by the A3 line and settles it as A3. NO SCAFFOLD WAS USED AND NOTHING WAS CHECKED AGAINST ONE: no MIDI, no Cyber Hymnal asset and no existing encoding was opened at any point. HOLLOW NOTEHEADS: there are none. Every notehead in this hymn is filled \u2014 a quarter, a dotted quarter or an eighth \u2014 and the hymn contains no half or whole note at all. Exactly one head reads hollow on the 1883 scan, the bass F3 of bar 9's third chord, and it is ink dropout: the 1875 plate prints it solid, and a dotted half could not fit a 6/8 bar that has already spent a quarter and an eighth. ACCIDENTALS WERE SWEPT SEPARATELY, BY EYE, AND THE AUTOMATED FIGURE ABOVE NEVER COVERED THEM \u2014 a notehead probe reads staff positions, not the sharps and flats in front of them. All four systems of both leaves were examined for accidental glyphs at system-level magnification and again on every 3x-8x crop. There is exactly ONE printed accidental in the hymn: a sharp before the LOWER (alto) head of the fourth chord of bar 13, on the syllable \"-tured\", centred on the first ledger line below the treble staff, making that note C sharp 4 and the chord G4 / C sharp 4 / E3 / A2, an A7. Both plates print it, in the same place. There are no naturals, no other sharps, and no flats beyond the one-flat key signature. DEPARTURES FROM THE PAGE, in full: (1) the plates draw a light double bar at each system end and a heavy double bar at the close, and only the closing barline is carried (`final_barline` \"|.\"), the engine drawing ordinary barlines and a system break at the same soprano ordinals; (2) nothing else \u2014 every pitch and every duration below is the plates', the phrase-end eighth rests of bars 2, 4, 6, 8, 14 and 16 are printed and are reproduced rather than collapsed (this is 6/8, not the 4/4 case the corpus collapses), the alto's two beamed eighths in bar 6 against a held quarter in the other three voices are the plates' own and are the only non-homorhythmic moment in the hymn, and the tenor/bass unisons of bars 3, 7 and 13 carry both voices on one printed head exactly as engraved.", "pd_jurisdictions": {"US": true}, "encoding_origin": "Hand-encoded from the page images of Gospel Hymns and Sacred Songs 1875, No. 45, p. 46 \u2014 archive.org gospelhymnssacre00blis_0 leaf n49 \u2014 collated chord by chord against the independent 1883 engraving in Gospel Hymns Consolidated Nos. 1 to 4, No. 45, p. 48 (archive.org hymnsconso00blis, leaf n49). No MIDI, no Cyber Hymnal asset and no existing encoding was used at any point.", "encoding_license": "PD-dedication"}, "source_line": "Setting: William H. Doane, as printed in Gospel Hymns and Sacred Songs, 1875. Public domain.", "music_ly": "%% NEAR THE CROSS (7.6.7.6 with Refrain) \u2014 \"Jesus, Keep Me Near the Cross\"\n%% William H. Doane's four-part setting, key F major, 6/8, no pickup, 16 bars.\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. Read off TWO dated pre-1931 printings of this setting,\n%% each scanned separately and each read with its own binarisation and its own\n%% staff-line fits:\n%%   (A) Gospel Hymns and Sacred Songs (New York: Biglow & Main, 1875), No. 45,\n%%       printed page 46 \u2014 archive.org gospelhymnssacre00blis_0, leaf n49\n%%       (2244 px across, staff space 28.2 px = 14.1 px per diatonic step);\n%%   (B) Gospel Hymns Consolidated, Nos. 1 to 4 (New York: Biglow & Main;\n%%       Cincinnati: John Church, 1883), No. 45, printed page 46 \u2014 archive.org\n%%       hymnsconso00blis, leaf n49 (2428 px across, staff space 28.9 px =\n%%       14.4 px per diatonic step).\n%% Both pages are headed \"Near the Cross.\" over the motto \"Peace through the\n%% blood of His cross.\"\u2014Coll. 1: 29, and credit \"FANNY J. CROSBY.\" at the left\n%% and \"W. H. DOANE, by per.\" at the right. They are NOT the same plate: the\n%% lyric type is reset (1875 \"fountain\" / \"Shed its beams\", 1883 \"fount-ain\" /\n%% \"Sheds its beams\") and the note spacing differs, so they are two independent\n%% engravings of one setting, which is why they were required to agree.\n%%\n%% HOW THE NOTEHEADS WERE READ. Per scan: adaptive binarisation (ink darker than\n%% a fraction of the LOCAL paper level taken from a 41x41 grey dilation plus a\n%% 41x41 box filter \u2014 0.82 on the 1883 scan, 0.70 on the darker 1875 scan, each\n%% chosen from that scan's own row-profile behaviour), then a 5x5 binary CLOSING\n%% to seal the pale gaps the plate's ink leaves inside a head; STAFF-LINE REMOVAL\n%% by vertical run length (an ink run no longer than 0.30 staff space sitting on\n%% a fitted line is deleted); stems, barlines and ledger stubs removed by a\n%% horizontal erosion of 0.30 staff space; then, for every surviving chord\n%% column, the observed vertical ink profile was FITTED against a model built\n%% from that column's own freshly re-fitted five staff lines \u2014 one or two\n%% notehead boxes 0.76 staff space tall placed at exact diatonic positions, the\n%% hypothesis with the smallest squared residual winning. The five lines are\n%% re-found in a +-2.6-staff-space window around each column's own x, from the\n%% staff-line pixels alone, and fitted as a RIGID equally spaced grid, so neither\n%% the page's curvature nor a notehead sitting on a line can drag a reading.\n%% 110 of the 157 columns the two scans report agree exactly; the remainder are\n%% the clef/key/time glyphs, the eighth rests, and eight note columns that were\n%% then settled BY EYE at 3x-8x on crops with the fitted grid drawn over them.\n%%\n%% WHERE THE MEASUREMENT WAS WRONG, AND HOW IT WAS CAUGHT. (1) An eighth note's\n%% DOWN-STEM FLAG can score as a second notehead: three columns first read\n%% \"F3 over A2\" are in fact a tenor/bass unison on F3 (system 4, bar 1, note 2;\n%% system 1, bar 3, note 2; system 3, bar 3, note 2). (2) The tenor rises to F4\n%% through the chorus, above the bass staff on two ledger lines; a first pass\n%% that capped the search at D4 mis-read all five of those columns. (3) The\n%% tenor at system 1, bar 2, note 1 measured half-way between A3 and B-flat 3 on\n%% the 1883 scan; the 1875 scan prints the head plainly bisected by the A3 line\n%% and settles it as A3. Each of the three was then confirmed by eye on a\n%% grid-overlay crop.\n%%\n%% NO HALF OR WHOLE NOTES. Every notehead in this hymn is filled: a quarter, a\n%% dotted quarter or an eighth. ONE head reads hollow on the 1883 scan \u2014 the bass\n%% F3 of the chorus's third chord (bar 9) \u2014 and it is ink dropout, not an outline\n%% head: the 1875 plate prints it solid, and a dotted half would not fit a 6/8 bar\n%% that already spends a quarter and an eighth before it.\n%%\n%% THE ONE PRINTED ACCIDENTAL. Both plates engrave a SHARP, and only one, in the\n%% whole hymn: before the LOWER (alto) head of the fourth chord of bar 13\n%% (\"-tured\"), centred on the first ledger line below the treble staff, making\n%% that note C sharp 4. The chord is then G4 / C#4 / E3 / A2 \u2014 an A7. Every other\n%% column on both leaves was swept separately by eye for accidentals at\n%% system-level magnification and at 3x-8x on the crops; there are no naturals,\n%% no other sharps and no flats beyond the one-flat key signature.\n%%\n%% THE ONE PLACE THE FOUR VOICES ARE NOT HOMORHYTHMIC. In bar 6 the soprano,\n%% tenor and bass hold a quarter on the third beat-group while the ALTO moves in\n%% two beamed eighths, D4 then B-flat 3. Everywhere else all four voices restrike\n%% together, exactly as both plates print.\n%%\n%% THE PHRASE-END EIGHTH RESTS ARE THE PAGE'S. Bars 2, 6, 10 (system 2's fourth\n%% bar), 14 and 16 close with a quarter plus an eighth rest in all four voices;\n%% bars 4 and 8 likewise. They are reproduced, not collapsed into a longer note:\n%% this is 6/8, not the 4/4 case the corpus collapses.\n%%\n%% STRUCTURE. Bars 1-8 are the verse (two systems of four bars, 13 soprano\n%% noteheads each); bars 9-16 are the CHORUS, which the plates label \"CHORUS.\"\n%% over bar 9 (12 soprano noteheads, then 13). 51 soprano noteheads in all \u2014\n%% 26 in the verse and 25 in the refrain \u2014 and the setting is strictly syllabic:\n%% there is not one slur or tie anywhere on either plate.\n\\version \"2.24.0\"\n\nglobal  = { \\key f \\major \\time 6/8 }\n\nsoprano = {\n  a'4 bes'8 a'4 g'8 |\n  f'4 d'8 d'4 r8 |\n  c'4 f'8 f'4 a'8 |\n  a'4. g'4 r8 |\n  a'4 bes'8 a'4 g'8 |\n  f'4 d'8 d'4 r8 |\n  c'4 f'8 f'4 e'8 |\n  g'4. f'4 r8 |\n  a'4 c''8 c''4. |\n  bes'4 d''8 d''4. |\n  c''4 d''8 c''4 a'8 |\n  a'4. g'4. |\n  a'4 bes'8 a'4 g'8 |\n  f'4 d'8 d'4 r8 |\n  c'4 f'8 f'4 e'8 |\n  g'4. f'4 r8\n}\n\nalto = {\n  f'4 f'8 f'4 e'8 |\n  d'4 d'8 bes4 r8 |\n  c'4 c'8 c'4 f'8 |\n  f'4. e'4 r8 |\n  f'4 f'8 f'4 e'8 |\n  d'4 d'8 d'8 bes8 r8 |\n  a4 c'8 c'4 c'8 |\n  c'4. c'4 r8 |\n  f'4 f'8 f'4. |\n  f'4 bes'8 bes'4. |\n  a'4 bes'8 a'4 f'8 |\n  f'4. e'4. |\n  c'4 d'8 c'4 cis'8 |\n  d'4 bes8 bes4 r8 |\n  a4 c'8 c'4 c'8 |\n  c'4. c'4 r8\n}\n\ntenor = {\n  c'4 d'8 c'4 bes8 |\n  a4 bes8 f4 r8 |\n  a4 a8 c'4 c'8 |\n  c'4. c'4 r8 |\n  c'4 d'8 c'4 bes8 |\n  a4 bes8 bes4 r8 |\n  f4 a8 a4 g8 |\n  bes4. a4 r8 |\n  c'4 a8 a4. |\n  d'4 f'8 f'4. |\n  f'4 f'8 f'4 c'8 |\n  c'4. c'4. |\n  f4 f8 f4 e8 |\n  d4 f8 f4 r8 |\n  f4 a8 a4 g8 |\n  bes4. a4 r8\n}\n\nbass = {\n  f4 f8 f4 c8 |\n  d4 bes,8 bes,4 r8 |\n  f4 f8 a4 f8 |\n  c4. c4 r8 |\n  f4 f8 f4 c8 |\n  d4 bes,8 bes,4 r8 |\n  c4 c8 c4 c8 |\n  c4. f4 r8 |\n  f4 f8 f4. |\n  bes4 bes8 bes4. |\n  f4 f8 f4 f8 |\n  c4. c4. |\n  f4 f8 f4 a,8 |\n  bes,4 bes,8 bes,4 r8 |\n  c4 c8 c4 c8 |\n  c4. f4 r8\n}\n", "preferred": true}], "license": "The hymn texts and tunes are in the public domain; per-layer provenance is recorded on each record."}