{"slug": "lower-lights", "name": "LOWER LIGHTS", "meter_label": "8.7.8.7 with Refrain", "meter": {"pattern": [8, 7, 8, 7], "doubled": false, "refrain": true, "alleluia": false, "irregular": false}, "composer": "philip-paul-bliss", "composer_date": "1871", "canonical_key": {"tonic": "bes", "mode": "major"}, "preferred_harmonization": "lower-lights-1875", "provenance": {"layer": "tune", "source": "gospel-hymns-sacred-songs-1875", "source_kind": "hand_encoded", "pd_basis": "pub_pre_1931", "pd_evidence": "Tune by Philip Paul Bliss (1838-1876), written for his own words and first published in The Charm: A Collection of Sunday School Music, ed. P. P. Bliss (Chicago: Root & Cady, 1871). Read here from the dated printing in Gospel Hymns and Sacred Songs, ed. P. P. Bliss & Ira D. Sankey (New York: Biglow & Main, 1875) \u2014 preface leaf engraved \"[COPYRIGHT 1875, BY BIGLOW & MAIN\" \u2014 no. 65, printed page 65, archive.org item gospelhymnssacre00blis_0, leaf n68, which credits \"P. P. Bliss.\" (words) and \"P. P. Bliss, by per.\" (music). Collated against a second dated printing in a DIFFERENT typesetting: Gospel Hymns and Sacred Songs, Consolidated Nos. 1 to 4 (1883), no. 65, printed page 65, archive.org gospelhymnsconso00blis_202606, leaf n64. US public domain by publication before 1929. Melody in B-flat, 3/4, one-quarter pickup, 60 soprano noteheads (30 verse + 30 refrain). tools/tune_fingerprint.py finds no corpus melody closer than 0.52 (LAUDA ANIMA), so this is not a duplicate. NEITHER dated page names the tune on the staff \u2014 the Gospel Hymns series rarely does \u2014 so the name LOWER LIGHTS is the one later hymnals and Hymnary print for it, taken from the chorus; no printed alt_name is recorded on either leaf.", "pd_jurisdictions": {"US": true}}, "harmonization": {"slug": "lower-lights-1875", "tune": "lower-lights", "music_file": "lower-lights-1875.ly", "canonical_key": {"tonic": "bes", "mode": "major"}, "time": "3/4", "pickup": 4, "system_breaks": [15, 30, 45], "breaks_manual": true, "markings": [{"soprano_note": 31, "kind": "refrain"}], "final_barline": "|.", "amen": false, "verse_variable_melisma": false, "refrain": {"soprano_note_start": 31}, "provenance": {"layer": "harmonization", "source": "gospel-hymns-sacred-songs-1875", "source_kind": "hand_encoded", "pd_basis": "pub_pre_1931", "pd_evidence": "P. P. Bliss's own four-part setting, read notehead by notehead off the dated page image of Gospel Hymns and Sacred Songs, ed. P. P. Bliss & Ira D. Sankey (New York: Biglow & Main, 1875) \u2014 the volume whose preface leaf engraves \"[COPYRIGHT 1875, BY BIGLOW & MAIN\" \u2014 no. 65, printed page 65, archive.org item gospelhymnssacre00blis_0, leaf n68. Every chord of all four voices was then re-read, end to end, on a DIFFERENT typesetting of the same setting: Gospel Hymns and Sacred Songs, Consolidated Nos. 1 to 4 (Biglow & Main / John Church, 1883), no. 65, printed page 65, archive.org item gospelhymnsconso00blis_202606, leaf n64. The two plates are demonstrably independent engravings, though they cast the same four systems: the 1875 underlay hyphenates \"Fa-ther's mer-cy\" where the 1883 sets \"Father's mer- cy\", the 1883 adds a semicolon after \"sea-man\" that the 1875 lacks, and their chord positions within a system differ by up to 2.7 per cent of the system width (measured on the normalised x of all fourteen chords of system 1) \u2014 a reprint from one plate could not do any of that. What they do NOT differ in is the notes: they agree at every position: 242 note positions are encoded here (60 soprano, 62 alto, 60 tenor, 60 bass) and NOT ONE of them differs between the two leaves. Both are pre-1931 US public domain and no modern arranger stands anywhere in the chain. NO Cyber Hymnal asset, no MIDI and no modern hymnal were consulted at any point. HOW THE PITCHES WERE MEASURED. Not by eye. Per staff the five staff lines were re-fitted LOCALLY as a function of x in 48-px blocks (both scans skew), the page binarised against its LOCAL paper level, the staff lines stripped column by column (any dark run no thicker than twice the measured line thickness), hollow half-note heads closed by hole filling, and heads isolated by correlation with an ellipse sized in staff spaces followed by non-maximum suppression at 0.85 staff space \u2014 which separates a soprano head stacked over an alto head without midpoint guesswork. Each peak was converted to a diatonic step offset from the LOCAL bottom-line fit and accepted only within 0.15 step of an exact staff position. Staff space measured 28.7 px on the 1875 leaf and 28.3 px on the 1883 leaf (one diatonic step = 14.4 and 14.2 px respectively), both far above the ~12 px floor at which readings stop being trustworthy. WHERE THE PROBE DID NOT REACH, and it is a small list. (a) The four HOLLOW half-note chords \u2014 verse \"more,\" and \"shore.\", refrain \"wave!\" and \"save.\" \u2014 fell below the ellipse threshold on one or both leaves, so all four were read BY EYE at 3.2-4.5x on BOTH leaves against a drawn staff-line grid; they read F3/A3/f'/c'' for \"more,\" and \"wave!\" and B-flat2/B-flat3/d'/b-flat' for \"shore.\" and \"save.\" on each. (b) The alto's first quarter under the slur at \"more,\" and \"wave!\" measured 0.58-1.08 step above the bottom line depending on the leaf \u2014 the slur curving under the head pulls its centroid down and two of those four readings sit outside the 0.15-step tolerance \u2014 so that note too was settled by eye on both leaves: f', a full staff space above the alto's e-flat' that follows it. No other reading on either leaf was outside tolerance. THE AUTOMATED FIGURE DOES NOT COVER ACCIDENTALS. A notehead probe reads heads, not the sharps, flats and naturals in front of them, so every accidental was checked SEPARATELY, by eye, at 3-5x magnification, on BOTH leaves. There are exactly four accidental glyphs on the page and all four are in the ALTO: a NATURAL before the alto's e' at the head of verse bar 3 (\"light-house ev-er\", making that bar a C7 into the half cadence on F); a courtesy FLAT before the alto's e-flat' in the second half of verse bar 4 (\"more,\"), cancelling it; and the same pair again in refrain bars 3 and 4 (\"gleam a-cross the\", \"wave!\"). Neither soprano, tenor nor bass carries a single accidental on either leaf. DEPARTURES FROM THE PAGE, stated in full. (1) The two courtesy FLATS are NOT reprinted in the engraving: LilyPond's default accidental style does not restate a key-signature flat in the bar after a cancelled natural and no accidental was forced. The PITCHES are the page's e-flat; only the cautionary glyph is missing. (2) The systems are cast four notes-per-two-poetic-lines as described under system_breaks, which is NOT the plates' break, so that no break falls inside \"ev-er-more\". (3) Nothing else. No note is restruck where the page holds and none is held where the page restrikes; no rest is added or removed (the page prints no rest anywhere); the page's one slur \u2014 the alto's two slurred quarters f'-e-flat' under the held half note on \"more,\" and on \"wave!\" \u2014 is reproduced and no slur or tie is invented; no voice is octave-shifted; no amen is added (the page prints none). UNISONS ARE THE PAGE'S, not an encoding artefact. Four places print ONE notehead in the lower staff where two voices sound, and both plates print one there: the refrain pickup \"Let the\" and the \"Send a\" figure of refrain bar 2 (tenor and bass both b-flat), and the second note of verse bar 7 and refrain bar 7 (\"a\" of \"a-long\", \"cue,\" of \"res-cue,\") where tenor and bass are both f. The last quarter of verse bar 7 and refrain bar 7 (\"the\" of \"the shore\", \"may\" of \"may save\") puts alto and tenor on the same c' \u2014 a ledger note under the treble staff and a ledger note over the bass staff. LilyPond's default collision resolver is left to handle all of them. THE RHYTHM IS NOT SYMMETRICAL and was read off flags and dots, then checked against the proportional chord spacing on both plates. Refrain bars 1 and 3 print a dotted eighth and a sixteenth on their third and fourth notes (\"lights be\", \"cross the\") where verse bars 1 and 3 print two plain eighths (\"Fa-ther's\", \"ev-er\"); bar 7 of both halves prints two eighths before the dotted quarter, the mirror of bar 3. Every distinct figure \u2014 the pickup and bars 1, 2, 3 and 7 of both halves; bars 5 and 6 repeat bars 1 and 2 note value for note value \u2014 was read off its flags and dots at 2.2-4.5x on the 1875 leaf AND again at 3.6-4.5x on the 1883 leaf, and the two agree everywhere. On the 1875 leaf the chord gaps are 146/84/114 px in refrain bar 1 against 145/114/116 px in verse bar 1, and 86/87/143 px in bar 7; the 1883 leaf agrees once its narrower measures are normalised.", "pd_jurisdictions": {"US": true}, "encoding_origin": "Hand-encoded from the 1875 Gospel Hymns and Sacred Songs page image (archive.org gospelhymnssacre00blis_0, leaf n68), collated end to end against the independently typeset 1883 Consolidated printing (archive.org gospelhymnsconso00blis_202606, leaf n64). No MIDI and no Cyber Hymnal asset was used.", "encoding_license": "PD-dedication"}, "music_ly": "%% LOWER LIGHTS (8.7.8.7 with Refrain) \u2014 \"Let the Lower Lights Be Burning\"\n%% P. P. Bliss's own four-part setting, key B-flat major, 3/4, one-quarter pickup,\n%% 16 bars (8 verse + 8 refrain, the pickup and the final half note together\n%% making one bar at each end).\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 notehead by notehead off the dated page image of\n%% Gospel Hymns and Sacred Songs, ed. P. P. Bliss & Ira D. Sankey (New York:\n%% Biglow & Main, 1875) \u2014 the volume whose preface leaf engraves \"[COPYRIGHT\n%% 1875, BY BIGLOW & MAIN\" \u2014 no. 65, printed page 65, archive.org item\n%% gospelhymnssacre00blis_0, leaf n68. Every chord of all four voices was then\n%% re-read, end to end, on a DIFFERENT typesetting of the same setting: Gospel\n%% Hymns and Sacred Songs, Consolidated Nos. 1 to 4 (Biglow & Main / John\n%% Church, 1883), no. 65, printed page 65, archive.org item\n%% gospelhymnsconso00blis_202606, leaf n64. The two dated witnesses agree at\n%% every one of the 60 soprano positions and in all four voices. They cast the\n%% same four systems, but they are demonstrably different engravings: the 1875\n%% underlay hyphenates \"Fa-ther's mer-cy\" where the 1883 sets \"Father's mer- cy\",\n%% the 1883 puts a semicolon after \"sea-man\" that the 1875 lacks, and their chord\n%% positions within a system differ by up to 2.7 per cent of the system width. So\n%% the agreement is not one plate reprinted from the other.\n%%\n%% HOW THE PITCHES WERE MEASURED (not eyeballed). For each staff the five staff\n%% lines were re-fitted LOCALLY as a function of x in 48-px blocks (both scans\n%% skew), the page binarised against its LOCAL paper level, the staff lines\n%% stripped column by column (any run no thicker than twice the measured line\n%% thickness), hollow half-note heads closed by hole filling, and heads located\n%% by correlation with an ellipse sized in staff spaces followed by non-maximum\n%% suppression at 0.85 staff space \u2014 which separates a soprano head stacked over\n%% an alto head without any midpoint guesswork. Each peak was converted to a\n%% diatonic step offset from the LOCAL bottom-line fit. Staff space measured\n%% 28.7 px on the 1875 plate and 28.3 px on the 1883 plate (one diatonic step =\n%% 14.4 / 14.2 px), and accepted readings sat within 0.15 step of an exact staff\n%% position. TWO PLACES WERE NOT SETTLED BY THAT PROBE and were read by eye at\n%% 3.2-4.5x on BOTH leaves instead: the four hollow half-note chords (\"more,\",\n%% \"shore.\", \"wave!\", \"save.\") which fell below the ellipse threshold, and the\n%% alto's first quarter under the slur at \"more,\" and \"wave!\", whose centroid the\n%% slur pulls down to 0.58-1.08 step. Everything else read inside tolerance on\n%% both leaves. 242 note positions are encoded here (60 soprano, 62 alto, 60\n%% tenor, 60 bass) and not one of them differs between the two dated leaves.\n%%\n%% ACCIDENTALS WERE NOT COVERED BY THAT MEASUREMENT. A notehead probe reads\n%% heads, not the sharps, flats and naturals in front of them, so every\n%% accidental on both leaves was checked SEPARATELY, by eye, at 3-5x\n%% magnification. There are exactly four accidental glyphs on the page and all\n%% four sit in the ALTO:\n%%   * a NATURAL before the alto's e' at the start of verse bar 3 (\"light-house\n%%     ev-er\"), holding for that whole bar \u2014 the bar is a C7 (c/g/e'/bes'-c'')\n%%     leading to the half cadence on F;\n%%   * a courtesy FLAT before the alto's ees' on the second half of verse bar 4\n%%     (\"more,\"), cancelling that natural;\n%%   * the same NATURAL before the alto's e' at the start of refrain bar 3\n%%     (\"gleam a-cross the\"), again a C7;\n%%   * the same courtesy FLAT before the alto's ees' in refrain bar 4 (\"wave!\").\n%% No accidental appears anywhere in the soprano, tenor or bass on either leaf.\n%%\n%% DEPARTURES FROM THE PAGE (declared, and repeated in pd_evidence):\n%%  1. The two courtesy FLATS above are NOT reprinted here. LilyPond's default\n%%     accidental style does not restate a key-signature flat in the bar after a\n%%     cancelled natural, and no accidental was forced. The PITCHES are the\n%%     page's (e-flat); only the cautionary glyph is absent.\n%%  2. The page's SLUR over the alto's two-quarter \"more,\" and \"wave!\" figure is\n%%     kept (see below), but the page draws no other slur or tie anywhere and\n%%     none is added.\n%%  3. Nothing else. No note is restruck where the page holds and none is held\n%%     where the page restrikes; no rest is added or removed (the page prints no\n%%     rest at all); no voice is transposed; no amen is added.\n%%\n%% CLOSE SCORE. The page is a two-staff close score: soprano over alto on the\n%% treble staff and tenor over bass on the bass staff, normally sharing one\n%% stem, so the two voices of a staff share a rhythm. The one exception is the\n%% phrase-end \"more,\" (verse bar 4) and \"wave!\" (refrain bar 4): there the\n%% soprano, tenor and bass hold a HALF note while the ALTO moves f' -> ees' in\n%% two quarters under a printed slur, its stem turned down against the soprano's\n%% up-stem. That is engraved identically on both dated leaves and is reproduced\n%% here exactly.\n%%\n%% UNISONS ARE THE PAGE'S. Four places print ONE notehead in the lower staff\n%% where two voices sound: the refrain pickup \"Let the\" and the \"Send a\" figure\n%% of refrain bar 2 (tenor and bass both bes), and the second note of verse bar\n%% 7 / refrain bar 7 (\"a\" of \"a-long\", \"cue,\" of \"res-cue,\") where tenor and\n%% bass are both f. The treble staff prints a unison c' for alto and tenor on\n%% \"the\" of \"the shore\" / \"may\" of \"may save\" (verse bar 7, refrain bar 7 last\n%% quarter) \u2014 the alto's c' is a ledger note under the treble staff and the\n%% tenor's c' a ledger note over the bass staff. Both voices carry the pitch here\n%% and LilyPond's default collision resolver is left to kiss the noteheads.\n%%\n%% RHYTHM, bar by bar (3/4 throughout; the page's gospel snap is real and was\n%% read off flags and dots at high magnification, then cross-checked against the\n%% horizontal spacing of the chords, which is proportional on both plates):\n%%            VERSE                             REFRAIN\n%%   pickup   8. 16   \"Bright-ly\"               8. 16   \"Let the\"\n%%   bar 1    4. 8 8 8   \"beams our Fa-ther's\"  4. 8 8. 16  \"low-er lights be\"\n%%   bar 2    8 4. 8. 16 \"mer-cy From His\"      8 4. 8. 16  \"burn-ing! Send a\"\n%%   bar 3    4. 8 8 8   \"light-house ev-er\"    4. 8 8. 16  \"gleam a-cross the\"\n%%   bar 4    2 + 8. 16  \"more, But to\"         2 + 8. 16   \"wave! Some poor\"\n%%   bar 5    4. 8 8 8   \"us He gives the\"      4. 8 8 8    \"faint-ing, strug-gling\"\n%%   bar 6    8 4. 8. 16 \"keep-ing Of the\"      8 4. 8. 16  \"sea-man You may\"\n%%   bar 7    8 8 4. 8   \"lights a-long the\"    8 8 4. 8    \"res-cue, you may\"\n%%   bar 8    2          \"shore.\"               2           \"save.\"\n%% Note the two places where the refrain does NOT copy the verse's figure:\n%% refrain bars 1 and 3 print a dotted eighth and a sixteenth on their third and\n%% fourth notes (\"lights be\", \"cross the\") where verse bars 1 and 3 print two\n%% plain eighths (\"Fa-ther's\", \"ev-er\"). The third note of refrain bars 1 and 3\n%% carries an augmentation dot AND a single flag and the fourth carries two\n%% flags, while the third and fourth notes of verse bars 1 and 3 carry a single\n%% flag and no dot. EVERY distinct figure above \u2014 the pickup and bars 1, 2, 3 and\n%% 7 of both halves \u2014 was read off its flags and dots at 2.2-4.5x on the 1875\n%% leaf AND again at 3.6-4.5x on the 1883 leaf (bars 5 and 6 repeat the figures\n%% of bars 1 and 2 note value for note value). The proportional chord spacing\n%% agrees on both plates as well (1875: 146/84/114 px in refrain bar 1 against\n%% 145/114/116 px in verse bar 1, and 86/87/143 px in bar 7).\n\\version \"2.24.0\"\n\nglobal  = { \\key bes \\major \\time 3/4 \\partial 4 }\n\nsoprano = {\n  f'8. g'16 |\n  f'4. d''8 c''8 bes'8 |\n  a'8 g'4. g'8. g'16 |\n  c''4. bes'8 a'8 bes'8 |\n  c''2 f'8. g'16 |\n  f'4. bes'8 c''8 d''8 |\n  ees''8 g'4. bes'8. c''16 |\n  d''8 bes'8 c''4. a'8 |\n  bes'2 bes'8. bes'16 |\n  ees''4. ees''8 ees''8. ees''16 |\n  ees''8 d''4. d''8. d''16 |\n  d''4. c''8 g'8. c''16 |\n  c''2 f'8. g'16 |\n  f'4. bes'8 c''8 d''8 |\n  ees''8 g'4. bes'8. c''16 |\n  d''8 bes'8 c''4. a'8 |\n  bes'2\n}\n\nalto = {\n  d'8. ees'16 |\n  d'4. f'8 ees'8 d'8 |\n  f'8 ees'4. ees'8. ees'16 |\n  e'4. e'8 e'8 e'8 |\n  f'4( ees'4) d'8. ees'16 |\n  d'4. d'8 ees'8 f'8 |\n  g'8 ees'4. g'8. g'16 |\n  f'8 d'8 ees'4. c'8 |\n  d'2 d'8. d'16 |\n  g'4. g'8 g'8. g'16 |\n  g'8 f'4. f'8. f'16 |\n  e'4. e'8 e'8. e'16 |\n  f'4( ees'4) d'8. ees'16 |\n  d'4. d'8 ees'8 f'8 |\n  g'8 ees'4. g'8. g'16 |\n  f'8 d'8 ees'4. c'8 |\n  d'2\n}\n\ntenor = {\n  bes8. bes16 |\n  bes4. bes8 bes8 bes8 |\n  bes8 bes4. bes8. g16 |\n  g4. c'8 c'8 bes8 |\n  a2 bes8. bes16 |\n  bes4. bes8 bes8 bes8 |\n  bes8 bes4. bes8. bes16 |\n  bes8 f8 a4. c'8 |\n  bes2 bes8. bes16 |\n  bes4. bes8 bes8. bes16 |\n  bes8 bes4. bes8. bes16 |\n  c'4. c'8 c'8. bes16 |\n  a2 bes8. bes16 |\n  bes4. bes8 bes8 bes8 |\n  bes8 bes4. bes8. bes16 |\n  bes8 f8 a4. c'8 |\n  bes2\n}\n\nbass = {\n  bes,8. bes,16 |\n  bes,4. bes,8 bes,8 bes,8 |\n  ees8 ees4. ees8. ees16 |\n  c4. c8 c8 c8 |\n  f2 bes,8. bes,16 |\n  bes,4. bes,8 bes,8 bes,8 |\n  ees8 ees4. ees8. ees16 |\n  f8 f8 f4. f8 |\n  bes,2 bes8. bes16 |\n  ees4. ees8 ees8. ees16 |\n  bes,8 bes,4. bes8. bes16 |\n  g4. g8 g8. g16 |\n  f2 bes,8. bes,16 |\n  bes,4. bes,8 bes,8 bes,8 |\n  ees8 ees4. ees8. ees16 |\n  f8 f8 f4. f8 |\n  bes,2\n}\n"}, "harmonizations": [{"slug": "lower-lights-1875", "tune": "lower-lights", "music_file": "lower-lights-1875.ly", "canonical_key": {"tonic": "bes", "mode": "major"}, "time": "3/4", "pickup": 4, "system_breaks": [15, 30, 45], "breaks_manual": true, "markings": [{"soprano_note": 31, "kind": "refrain"}], "final_barline": "|.", "amen": false, "verse_variable_melisma": false, "refrain": {"soprano_note_start": 31}, "provenance": {"layer": "harmonization", "source": "gospel-hymns-sacred-songs-1875", "source_kind": "hand_encoded", "pd_basis": "pub_pre_1931", "pd_evidence": "P. P. Bliss's own four-part setting, read notehead by notehead off the dated page image of Gospel Hymns and Sacred Songs, ed. P. P. Bliss & Ira D. Sankey (New York: Biglow & Main, 1875) \u2014 the volume whose preface leaf engraves \"[COPYRIGHT 1875, BY BIGLOW & MAIN\" \u2014 no. 65, printed page 65, archive.org item gospelhymnssacre00blis_0, leaf n68. Every chord of all four voices was then re-read, end to end, on a DIFFERENT typesetting of the same setting: Gospel Hymns and Sacred Songs, Consolidated Nos. 1 to 4 (Biglow & Main / John Church, 1883), no. 65, printed page 65, archive.org item gospelhymnsconso00blis_202606, leaf n64. The two plates are demonstrably independent engravings, though they cast the same four systems: the 1875 underlay hyphenates \"Fa-ther's mer-cy\" where the 1883 sets \"Father's mer- cy\", the 1883 adds a semicolon after \"sea-man\" that the 1875 lacks, and their chord positions within a system differ by up to 2.7 per cent of the system width (measured on the normalised x of all fourteen chords of system 1) \u2014 a reprint from one plate could not do any of that. What they do NOT differ in is the notes: they agree at every position: 242 note positions are encoded here (60 soprano, 62 alto, 60 tenor, 60 bass) and NOT ONE of them differs between the two leaves. Both are pre-1931 US public domain and no modern arranger stands anywhere in the chain. NO Cyber Hymnal asset, no MIDI and no modern hymnal were consulted at any point. HOW THE PITCHES WERE MEASURED. Not by eye. Per staff the five staff lines were re-fitted LOCALLY as a function of x in 48-px blocks (both scans skew), the page binarised against its LOCAL paper level, the staff lines stripped column by column (any dark run no thicker than twice the measured line thickness), hollow half-note heads closed by hole filling, and heads isolated by correlation with an ellipse sized in staff spaces followed by non-maximum suppression at 0.85 staff space \u2014 which separates a soprano head stacked over an alto head without midpoint guesswork. Each peak was converted to a diatonic step offset from the LOCAL bottom-line fit and accepted only within 0.15 step of an exact staff position. Staff space measured 28.7 px on the 1875 leaf and 28.3 px on the 1883 leaf (one diatonic step = 14.4 and 14.2 px respectively), both far above the ~12 px floor at which readings stop being trustworthy. WHERE THE PROBE DID NOT REACH, and it is a small list. (a) The four HOLLOW half-note chords \u2014 verse \"more,\" and \"shore.\", refrain \"wave!\" and \"save.\" \u2014 fell below the ellipse threshold on one or both leaves, so all four were read BY EYE at 3.2-4.5x on BOTH leaves against a drawn staff-line grid; they read F3/A3/f'/c'' for \"more,\" and \"wave!\" and B-flat2/B-flat3/d'/b-flat' for \"shore.\" and \"save.\" on each. (b) The alto's first quarter under the slur at \"more,\" and \"wave!\" measured 0.58-1.08 step above the bottom line depending on the leaf \u2014 the slur curving under the head pulls its centroid down and two of those four readings sit outside the 0.15-step tolerance \u2014 so that note too was settled by eye on both leaves: f', a full staff space above the alto's e-flat' that follows it. No other reading on either leaf was outside tolerance. THE AUTOMATED FIGURE DOES NOT COVER ACCIDENTALS. A notehead probe reads heads, not the sharps, flats and naturals in front of them, so every accidental was checked SEPARATELY, by eye, at 3-5x magnification, on BOTH leaves. There are exactly four accidental glyphs on the page and all four are in the ALTO: a NATURAL before the alto's e' at the head of verse bar 3 (\"light-house ev-er\", making that bar a C7 into the half cadence on F); a courtesy FLAT before the alto's e-flat' in the second half of verse bar 4 (\"more,\"), cancelling it; and the same pair again in refrain bars 3 and 4 (\"gleam a-cross the\", \"wave!\"). Neither soprano, tenor nor bass carries a single accidental on either leaf. DEPARTURES FROM THE PAGE, stated in full. (1) The two courtesy FLATS are NOT reprinted in the engraving: LilyPond's default accidental style does not restate a key-signature flat in the bar after a cancelled natural and no accidental was forced. The PITCHES are the page's e-flat; only the cautionary glyph is missing. (2) The systems are cast four notes-per-two-poetic-lines as described under system_breaks, which is NOT the plates' break, so that no break falls inside \"ev-er-more\". (3) Nothing else. No note is restruck where the page holds and none is held where the page restrikes; no rest is added or removed (the page prints no rest anywhere); the page's one slur \u2014 the alto's two slurred quarters f'-e-flat' under the held half note on \"more,\" and on \"wave!\" \u2014 is reproduced and no slur or tie is invented; no voice is octave-shifted; no amen is added (the page prints none). UNISONS ARE THE PAGE'S, not an encoding artefact. Four places print ONE notehead in the lower staff where two voices sound, and both plates print one there: the refrain pickup \"Let the\" and the \"Send a\" figure of refrain bar 2 (tenor and bass both b-flat), and the second note of verse bar 7 and refrain bar 7 (\"a\" of \"a-long\", \"cue,\" of \"res-cue,\") where tenor and bass are both f. The last quarter of verse bar 7 and refrain bar 7 (\"the\" of \"the shore\", \"may\" of \"may save\") puts alto and tenor on the same c' \u2014 a ledger note under the treble staff and a ledger note over the bass staff. LilyPond's default collision resolver is left to handle all of them. THE RHYTHM IS NOT SYMMETRICAL and was read off flags and dots, then checked against the proportional chord spacing on both plates. Refrain bars 1 and 3 print a dotted eighth and a sixteenth on their third and fourth notes (\"lights be\", \"cross the\") where verse bars 1 and 3 print two plain eighths (\"Fa-ther's\", \"ev-er\"); bar 7 of both halves prints two eighths before the dotted quarter, the mirror of bar 3. Every distinct figure \u2014 the pickup and bars 1, 2, 3 and 7 of both halves; bars 5 and 6 repeat bars 1 and 2 note value for note value \u2014 was read off its flags and dots at 2.2-4.5x on the 1875 leaf AND again at 3.6-4.5x on the 1883 leaf, and the two agree everywhere. On the 1875 leaf the chord gaps are 146/84/114 px in refrain bar 1 against 145/114/116 px in verse bar 1, and 86/87/143 px in bar 7; the 1883 leaf agrees once its narrower measures are normalised.", "pd_jurisdictions": {"US": true}, "encoding_origin": "Hand-encoded from the 1875 Gospel Hymns and Sacred Songs page image (archive.org gospelhymnssacre00blis_0, leaf n68), collated end to end against the independently typeset 1883 Consolidated printing (archive.org gospelhymnsconso00blis_202606, leaf n64). No MIDI and no Cyber Hymnal asset was used.", "encoding_license": "PD-dedication"}, "music_ly": "%% LOWER LIGHTS (8.7.8.7 with Refrain) \u2014 \"Let the Lower Lights Be Burning\"\n%% P. P. Bliss's own four-part setting, key B-flat major, 3/4, one-quarter pickup,\n%% 16 bars (8 verse + 8 refrain, the pickup and the final half note together\n%% making one bar at each end).\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 notehead by notehead off the dated page image of\n%% Gospel Hymns and Sacred Songs, ed. P. P. Bliss & Ira D. Sankey (New York:\n%% Biglow & Main, 1875) \u2014 the volume whose preface leaf engraves \"[COPYRIGHT\n%% 1875, BY BIGLOW & MAIN\" \u2014 no. 65, printed page 65, archive.org item\n%% gospelhymnssacre00blis_0, leaf n68. Every chord of all four voices was then\n%% re-read, end to end, on a DIFFERENT typesetting of the same setting: Gospel\n%% Hymns and Sacred Songs, Consolidated Nos. 1 to 4 (Biglow & Main / John\n%% Church, 1883), no. 65, printed page 65, archive.org item\n%% gospelhymnsconso00blis_202606, leaf n64. The two dated witnesses agree at\n%% every one of the 60 soprano positions and in all four voices. They cast the\n%% same four systems, but they are demonstrably different engravings: the 1875\n%% underlay hyphenates \"Fa-ther's mer-cy\" where the 1883 sets \"Father's mer- cy\",\n%% the 1883 puts a semicolon after \"sea-man\" that the 1875 lacks, and their chord\n%% positions within a system differ by up to 2.7 per cent of the system width. So\n%% the agreement is not one plate reprinted from the other.\n%%\n%% HOW THE PITCHES WERE MEASURED (not eyeballed). For each staff the five staff\n%% lines were re-fitted LOCALLY as a function of x in 48-px blocks (both scans\n%% skew), the page binarised against its LOCAL paper level, the staff lines\n%% stripped column by column (any run no thicker than twice the measured line\n%% thickness), hollow half-note heads closed by hole filling, and heads located\n%% by correlation with an ellipse sized in staff spaces followed by non-maximum\n%% suppression at 0.85 staff space \u2014 which separates a soprano head stacked over\n%% an alto head without any midpoint guesswork. Each peak was converted to a\n%% diatonic step offset from the LOCAL bottom-line fit. Staff space measured\n%% 28.7 px on the 1875 plate and 28.3 px on the 1883 plate (one diatonic step =\n%% 14.4 / 14.2 px), and accepted readings sat within 0.15 step of an exact staff\n%% position. TWO PLACES WERE NOT SETTLED BY THAT PROBE and were read by eye at\n%% 3.2-4.5x on BOTH leaves instead: the four hollow half-note chords (\"more,\",\n%% \"shore.\", \"wave!\", \"save.\") which fell below the ellipse threshold, and the\n%% alto's first quarter under the slur at \"more,\" and \"wave!\", whose centroid the\n%% slur pulls down to 0.58-1.08 step. Everything else read inside tolerance on\n%% both leaves. 242 note positions are encoded here (60 soprano, 62 alto, 60\n%% tenor, 60 bass) and not one of them differs between the two dated leaves.\n%%\n%% ACCIDENTALS WERE NOT COVERED BY THAT MEASUREMENT. A notehead probe reads\n%% heads, not the sharps, flats and naturals in front of them, so every\n%% accidental on both leaves was checked SEPARATELY, by eye, at 3-5x\n%% magnification. There are exactly four accidental glyphs on the page and all\n%% four sit in the ALTO:\n%%   * a NATURAL before the alto's e' at the start of verse bar 3 (\"light-house\n%%     ev-er\"), holding for that whole bar \u2014 the bar is a C7 (c/g/e'/bes'-c'')\n%%     leading to the half cadence on F;\n%%   * a courtesy FLAT before the alto's ees' on the second half of verse bar 4\n%%     (\"more,\"), cancelling that natural;\n%%   * the same NATURAL before the alto's e' at the start of refrain bar 3\n%%     (\"gleam a-cross the\"), again a C7;\n%%   * the same courtesy FLAT before the alto's ees' in refrain bar 4 (\"wave!\").\n%% No accidental appears anywhere in the soprano, tenor or bass on either leaf.\n%%\n%% DEPARTURES FROM THE PAGE (declared, and repeated in pd_evidence):\n%%  1. The two courtesy FLATS above are NOT reprinted here. LilyPond's default\n%%     accidental style does not restate a key-signature flat in the bar after a\n%%     cancelled natural, and no accidental was forced. The PITCHES are the\n%%     page's (e-flat); only the cautionary glyph is absent.\n%%  2. The page's SLUR over the alto's two-quarter \"more,\" and \"wave!\" figure is\n%%     kept (see below), but the page draws no other slur or tie anywhere and\n%%     none is added.\n%%  3. Nothing else. No note is restruck where the page holds and none is held\n%%     where the page restrikes; no rest is added or removed (the page prints no\n%%     rest at all); no voice is transposed; no amen is added.\n%%\n%% CLOSE SCORE. The page is a two-staff close score: soprano over alto on the\n%% treble staff and tenor over bass on the bass staff, normally sharing one\n%% stem, so the two voices of a staff share a rhythm. The one exception is the\n%% phrase-end \"more,\" (verse bar 4) and \"wave!\" (refrain bar 4): there the\n%% soprano, tenor and bass hold a HALF note while the ALTO moves f' -> ees' in\n%% two quarters under a printed slur, its stem turned down against the soprano's\n%% up-stem. That is engraved identically on both dated leaves and is reproduced\n%% here exactly.\n%%\n%% UNISONS ARE THE PAGE'S. Four places print ONE notehead in the lower staff\n%% where two voices sound: the refrain pickup \"Let the\" and the \"Send a\" figure\n%% of refrain bar 2 (tenor and bass both bes), and the second note of verse bar\n%% 7 / refrain bar 7 (\"a\" of \"a-long\", \"cue,\" of \"res-cue,\") where tenor and\n%% bass are both f. The treble staff prints a unison c' for alto and tenor on\n%% \"the\" of \"the shore\" / \"may\" of \"may save\" (verse bar 7, refrain bar 7 last\n%% quarter) \u2014 the alto's c' is a ledger note under the treble staff and the\n%% tenor's c' a ledger note over the bass staff. Both voices carry the pitch here\n%% and LilyPond's default collision resolver is left to kiss the noteheads.\n%%\n%% RHYTHM, bar by bar (3/4 throughout; the page's gospel snap is real and was\n%% read off flags and dots at high magnification, then cross-checked against the\n%% horizontal spacing of the chords, which is proportional on both plates):\n%%            VERSE                             REFRAIN\n%%   pickup   8. 16   \"Bright-ly\"               8. 16   \"Let the\"\n%%   bar 1    4. 8 8 8   \"beams our Fa-ther's\"  4. 8 8. 16  \"low-er lights be\"\n%%   bar 2    8 4. 8. 16 \"mer-cy From His\"      8 4. 8. 16  \"burn-ing! Send a\"\n%%   bar 3    4. 8 8 8   \"light-house ev-er\"    4. 8 8. 16  \"gleam a-cross the\"\n%%   bar 4    2 + 8. 16  \"more, But to\"         2 + 8. 16   \"wave! Some poor\"\n%%   bar 5    4. 8 8 8   \"us He gives the\"      4. 8 8 8    \"faint-ing, strug-gling\"\n%%   bar 6    8 4. 8. 16 \"keep-ing Of the\"      8 4. 8. 16  \"sea-man You may\"\n%%   bar 7    8 8 4. 8   \"lights a-long the\"    8 8 4. 8    \"res-cue, you may\"\n%%   bar 8    2          \"shore.\"               2           \"save.\"\n%% Note the two places where the refrain does NOT copy the verse's figure:\n%% refrain bars 1 and 3 print a dotted eighth and a sixteenth on their third and\n%% fourth notes (\"lights be\", \"cross the\") where verse bars 1 and 3 print two\n%% plain eighths (\"Fa-ther's\", \"ev-er\"). The third note of refrain bars 1 and 3\n%% carries an augmentation dot AND a single flag and the fourth carries two\n%% flags, while the third and fourth notes of verse bars 1 and 3 carry a single\n%% flag and no dot. EVERY distinct figure above \u2014 the pickup and bars 1, 2, 3 and\n%% 7 of both halves \u2014 was read off its flags and dots at 2.2-4.5x on the 1875\n%% leaf AND again at 3.6-4.5x on the 1883 leaf (bars 5 and 6 repeat the figures\n%% of bars 1 and 2 note value for note value). The proportional chord spacing\n%% agrees on both plates as well (1875: 146/84/114 px in refrain bar 1 against\n%% 145/114/116 px in verse bar 1, and 86/87/143 px in bar 7).\n\\version \"2.24.0\"\n\nglobal  = { \\key bes \\major \\time 3/4 \\partial 4 }\n\nsoprano = {\n  f'8. g'16 |\n  f'4. d''8 c''8 bes'8 |\n  a'8 g'4. g'8. g'16 |\n  c''4. bes'8 a'8 bes'8 |\n  c''2 f'8. g'16 |\n  f'4. bes'8 c''8 d''8 |\n  ees''8 g'4. bes'8. c''16 |\n  d''8 bes'8 c''4. a'8 |\n  bes'2 bes'8. bes'16 |\n  ees''4. ees''8 ees''8. ees''16 |\n  ees''8 d''4. d''8. d''16 |\n  d''4. c''8 g'8. c''16 |\n  c''2 f'8. g'16 |\n  f'4. bes'8 c''8 d''8 |\n  ees''8 g'4. bes'8. c''16 |\n  d''8 bes'8 c''4. a'8 |\n  bes'2\n}\n\nalto = {\n  d'8. ees'16 |\n  d'4. f'8 ees'8 d'8 |\n  f'8 ees'4. ees'8. ees'16 |\n  e'4. e'8 e'8 e'8 |\n  f'4( ees'4) d'8. ees'16 |\n  d'4. d'8 ees'8 f'8 |\n  g'8 ees'4. g'8. g'16 |\n  f'8 d'8 ees'4. c'8 |\n  d'2 d'8. d'16 |\n  g'4. g'8 g'8. g'16 |\n  g'8 f'4. f'8. f'16 |\n  e'4. e'8 e'8. e'16 |\n  f'4( ees'4) d'8. ees'16 |\n  d'4. d'8 ees'8 f'8 |\n  g'8 ees'4. g'8. g'16 |\n  f'8 d'8 ees'4. c'8 |\n  d'2\n}\n\ntenor = {\n  bes8. bes16 |\n  bes4. bes8 bes8 bes8 |\n  bes8 bes4. bes8. g16 |\n  g4. c'8 c'8 bes8 |\n  a2 bes8. bes16 |\n  bes4. bes8 bes8 bes8 |\n  bes8 bes4. bes8. bes16 |\n  bes8 f8 a4. c'8 |\n  bes2 bes8. bes16 |\n  bes4. bes8 bes8. bes16 |\n  bes8 bes4. bes8. bes16 |\n  c'4. c'8 c'8. bes16 |\n  a2 bes8. bes16 |\n  bes4. bes8 bes8 bes8 |\n  bes8 bes4. bes8. bes16 |\n  bes8 f8 a4. c'8 |\n  bes2\n}\n\nbass = {\n  bes,8. bes,16 |\n  bes,4. bes,8 bes,8 bes,8 |\n  ees8 ees4. ees8. ees16 |\n  c4. c8 c8 c8 |\n  f2 bes,8. bes,16 |\n  bes,4. bes,8 bes,8 bes,8 |\n  ees8 ees4. ees8. ees16 |\n  f8 f8 f4. f8 |\n  bes,2 bes8. bes16 |\n  ees4. ees8 ees8. ees16 |\n  bes,8 bes,4. bes8. bes16 |\n  g4. g8 g8. g16 |\n  f2 bes,8. bes,16 |\n  bes,4. bes,8 bes,8 bes,8 |\n  ees8 ees4. ees8. ees16 |\n  f8 f8 f4. f8 |\n  bes,2\n}\n", "preferred": true}], "license": "The hymn texts and tunes are in the public domain; per-layer provenance is recorded on each record."}