{"slug": "my-saviors-love", "name": "MY SAVIOR'S LOVE", "meter_label": "8.7.8.7 with Refrain", "meter": {"pattern": [8, 7, 8, 7], "doubled": false, "refrain": true, "alleluia": false, "irregular": false}, "composer": "charles-hutchinson-gabriel", "composer_date": "1905", "canonical_key": {"tonic": "aes", "mode": "major"}, "preferred_harmonization": "my-saviors-love-praises-1905", "provenance": {"layer": "tune", "source": "praises-1905", "source_kind": "hand_encoded", "pd_evidence": "Tune by Charles Hutchinson Gabriel (1856-1932), written for his own words. Read from its FIRST publication: Praises, ed. E. O. Excell (Chicago: E. O. Excell, 1905), No. 3, printed page 3 \u2014 archive.org praises00excegoog, leaf n10, collated against the second scan of the same 1905 edition, archive.org praises00exce, leaf n8. The page is headed \"My Savior's Love.\" and prints \"COPYRIGHT, 1905, BY CHAS. H. GABRIEL.\" between \"C. H. G.\" (left) and \"Chas. H. Gabriel.\" (right). Gabriel died in 1932, so his death date clears nothing on its own; THE PUBLIC-DOMAIN BASIS IS THE 1905 PUBLICATION DATE OF THE PRINTING READ HERE, which is before 1931. The soprano's first fifteen scale degrees are 553512311777121, which is one of the two incipits Hymnary indexes under tune authority my_saviors_love_gabriel, matched digit for digit; the other, 55351231777121566, is the same line with the repeated 1 at notes 8-9 counted once and carried three notes further (566 = notes 16-18 as read here). tools/tune_fingerprint.py's interval fingerprint finds no corpus melody closer than 0.574 (LEONI), far under the 0.80 duplicate threshold, so this is not a duplicate under another name. The plate does not print a tune name on the staff \u2014 gospel songbooks of this kind rarely do \u2014 so MY SAVIOR'S LOVE is the name the hymnals and Hymnary carry for it, taken from the page's title; no printed alt_name is recorded.", "pd_basis": "pub_pre_1931", "pd_jurisdictions": {"US": true}, "hymnary_ids": {"tune_auth": "my_saviors_love_gabriel", "total_instances": 278}}, "harmonization": {"slug": "my-saviors-love-praises-1905", "tune": "my-saviors-love", "music_file": "my-saviors-love-praises-1905.ly", "canonical_key": {"tonic": "aes", "mode": "major"}, "time": "4/4", "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": "praises-1905", "source_kind": "hand_encoded", "pd_basis": "pub_pre_1931", "pd_evidence": "Charles H. Gabriel's own four-part setting, read from its FIRST publication: Praises, ed. E. O. Excell (Chicago: E. O. Excell, 1905), No. 3, printed page 3, headed \"My Savior's Love.\" and engraved \"COPYRIGHT, 1905, BY CHAS. H. GABRIEL.\" between \"C. H. G.\" and \"Chas. H. Gabriel.\". US public domain BY THE 1905 PUBLICATION DATE OF THIS PRINTING, which is before 1931 \u2014 Gabriel did not die until 1932, so his death date clears nothing on its own, and the pre-1931 publication is the whole basis. No modern arranger stands anywhere in this chain: the setting is the composer's own, on the plate that first printed it. Every chord of all four voices was read notehead by notehead off TWO independent scans of that one plate \u2014 archive.org praises00excegoog, leaf n10 (3142 px across, staff space 35.3 px, i.e. 17.6 px per diatonic step) and archive.org praises00exce, leaf n8 (2418 px across, staff space 26.0 px, 13.0 px per step) \u2014 by the same rig, each scan with its OWN adaptive binarisation threshold (0.82 of the local paper level for the bitonal first scan, 0.72 for the second, which is a grey scan) and its own five-line fits recovered locally in a \u00b12.4-staff-space window around each notehead's x. After dropping the columns each pass itself marked as flag hooks, rests or clef fragments, the two readings align column for column across 129 chord columns; 238 noteheads were resolved by BOTH passes and they agree on all 238, with 0 disagreements. Six heads were resolved by one pass only (five by the first scan, one by the second), and ten chord positions that a flag hook or a shared stem hid from both passes were read by eye off crops with the fitted five-line grid drawn over them at 5x-12x; all ten are listed by soprano notehead ordinal in the .ly header. THE RIG READS STAFF POSITIONS AND IS BLIND TO ACCIDENTALS AND NO FIGURE ABOVE COVERS THEM, so every column was swept a second time, separately and by eye, end to end over all eight staves of the first scan at 2.0x with the contested chords re-opened at 4.5x-12x and the \"love\" chord re-examined on the second scan at 12x: the plate prints NO accidental anywhere outside its four-flat key signature. NO SCAFFOLD WAS USED and no MIDI, MusicXML or Cyber Hymnal asset was opened for this file. The soprano's first fifteen scale degrees, 553512311777121, match one of the two incipits Hymnary indexes for tune authority my_saviors_love_gabriel digit for digit. FOUR DEPARTURES from the page are declared in full in the .ly header: the plate's SECOND chorus lyric row is not carried; its fermata over the final chord is not rendered (house style); the c''-des'' eighth pair on \"my\" in chorus bar 7 is written with this corpus's slur so the underlay forms one note-group, which the engine turns back into the plate's own beam and draws no slur; and at \"love\" in the last bar of the second chorus line the soprano's hollow half note is followed here by half notes in the other three voices, whose heads the plate inks solid \u2014 three filled quarters would leave their 4/4 bar a beat short. That last head, the bass ees, is the one place where this encoding does not follow the engraving literally. THE CHORUS IS NOT HOMORHYTHMIC AND THAT IS THE PAGE: on the first beat of chorus bars 1, 2, 5 and 6 the soprano holds a half while the alto, tenor and bass restrike the same pitch as two quarters, so that the plate's alternative \"O how mar-vel-ous!\" row can take its extra syllable.", "pd_jurisdictions": {"US": true}, "encoding_origin": "Hand-encoded from the 1905 Praises page image (archive.org praises00excegoog, leaf n10), collated head by head against the second scan of the same 1905 edition (praises00exce, leaf n8). No MIDI and no Cyber Hymnal asset was used.", "encoding_license": "PD-dedication"}, "music_ly": "%% MY SAVIOR'S LOVE (8.7.8.7 with Refrain) \u2014 \"I Stand Amazed in the Presence\",\n%% words and music by Charles H. Gabriel. Four SATB voice streams, canonical key\n%% A-flat major, 4/4, one-quarter pickup, 16 full bars after the pickup.\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. The FIRST publication: Praises, edited and published by\n%% E. O. Excell (Chicago, 1905), No. 3, printed page 3, headed \"My Savior's\n%% Love.\" with \"C. H. G.\" at the left, \"Chas. H. Gabriel.\" at the right and\n%% \"COPYRIGHT, 1905, BY CHAS. H. GABRIEL.\" centred between them. Read off TWO\n%% independent scans of that one plate:\n%%   scan A \u2014 archive.org praises00excegoog, leaf n10 (3142 px across, staff\n%%            space 35.3 px, i.e. 17.6 px per diatonic step);\n%%   scan B \u2014 archive.org praises00exce, leaf n8 (2418 px across, staff space\n%%            26.0 px, 13.0 px per step).\n%% Both items are the same 1905 Excell edition; the two scans differ in operator,\n%% resolution and tone (A is bitonal, B is a grey scan), so they binarise apart.\n%%\n%% HOW THE NOTEHEADS WERE READ (not by eye). Per scan and per staff: adaptive\n%% binarisation (ink darker than a fraction of the LOCAL paper level built from a\n%% 41x41 grey dilation followed by a box filter \u2014 the fraction is each scan's own,\n%% 0.82 for A and 0.72 for B, chosen because B is a grey scan and 0.82 dissolves\n%% it); a binary closing scaled to the staff space to seal the pale gaps the\n%% plate's ink leaves inside a head; staff and ledger lines erased (long\n%% horizontal ink that is not part of a tall vertical run, so a head sitting on a\n%% line keeps its pixels); a second closing; then SMALL holes filled, which is\n%% what makes the hollow half-note heads readable at all; barlines erased; and\n%% finally erosion with an ellipse of semi-axes 0.29 x 0.215 staff spaces, which\n%% deletes stems and beams but leaves a notehead core. Every core was converted to\n%% a diatonic step with a FRESH five-line fit made in a +-2.4-staff-space window\n%% around that core's own x, so the plate's skew cannot drag a reading. Scan B was\n%% resampled to scan A's scale (x1.36) before the same rig ran on it; it kept its\n%% own binarisation and its own staff-line fits throughout.\n%%\n%% WHAT THE TWO PASSES AGREE ON. After dropping the columns each pass itself\n%% marked as flag hooks, rests or clef fragments, the two readings line up column\n%% for column: 129 chord columns. 238 noteheads were resolved by BOTH passes, and\n%% they agree on all 238 \u2014 0 disagreements. SIX heads were resolved by one pass\n%% only. Five of them scan A found and scan B missed: both heads of the \"how\"\n%% chord (soprano note 19, aes' over f'); the bass des under the sixteenth\n%% \"could\" that closes verse bar 5 (note 21); the tenor aes under the sixteenth\n%% \"con-\" in verse bar 7 (note 27, the only head EITHER pass found in that\n%% column); and the soprano's hollow ees'' that opens chorus bar 2 (note 35).\n%% One head scan B recovered that scan A lost inside a flag: the bass ees under\n%% \"-ior's\" (note 58). TEN further chord positions were resolved by EYE off crops\n%% with the fitted five-line grid drawn over them at 5x-12x, because a flag hook\n%% or a shared stem hid a head from both passes; all ten are listed under\n%% UNRESOLVED BY THE RIG below.\n%%\n%% THE MELODY IS CONFIRMED BY A THIRD, NON-IMAGE WITNESS. The soprano's first\n%% fifteen scale degrees read here are 5 5 3 5 1 2 3 1 1 7 7 7 1 2 1. Hymnary\n%% indexes two incipits under tune authority my_saviors_love_gabriel:\n%% 553512311777121, which this reading matches digit for digit, and\n%% 55351231777121566, which is the same line with the repeated 1 (notes 8-9)\n%% counted once and carried three notes further \u2014 5 6 6, which is what notes\n%% 16, 17 and 18 read here. tools/tune_fingerprint.py's interval fingerprint\n%% finds no corpus melody closer than 0.574 (LEONI), far under the 0.80\n%% duplicate threshold.\n%%\n%% NO SCAFFOLD WAS USED and nothing here was checked against one. No MIDI, no\n%% MusicXML, no Cyber Hymnal asset and no modern hymnal was opened for this file.\n%%\n%% ACCIDENTALS \u2014 A SEPARATE BY-EYE SWEEP, BECAUSE THE RIG CANNOT SEE THEM. The\n%% measurement above reads staff POSITIONS; it is blind to a sharp, flat or\n%% natural standing in front of a head, and no figure it produces covers them. So\n%% every column of all eight staves was swept a second time, by eye, end to end,\n%% on scan A at 2.0x in eight overlapping crops that include both staves of each\n%% system, with the contested chords re-opened at 4.5x-12x; the \"love\" chord was\n%% additionally re-examined on scan B at 12x. THE PLATE PRINTS NO ACCIDENTAL\n%% ANYWHERE except the four-flat key signature at the head of each staff: there\n%% is not one sharp, flat or natural in the body of the piece. Every b, e, a and\n%% d below is therefore flat, and is spelled bes/ees/aes/des here.\n%%\n%% RHYTHM. The plate's recurring figure is a dotted eighth plus a sixteenth, and\n%% it separately FLAGS both \u2014 it never beams them. It falls on beat 4 of verse\n%% bars 1 and 5, on beat 2 of verse bars 3 and 7, and on beat 3 of all four\n%% chorus bars 1, 2, 5 and 6. Chorus bars 3 and 7 instead carry a dotted quarter\n%% on beat 3 followed by an eighth on beat 4. Every augmentation dot and every\n%% flag was read at 2x-6x on scan A; the dots sit about 45 px (1.3 staff spaces)\n%% to the right of their heads and the sixteenths carry two hooks. Verse bar 4\n%% ends the first poetic couplet on a dotted half whose bar is completed by the\n%% next line's pickup quarter; verse bar 8 ends the stanza on a dotted half plus a\n%% printed QUARTER REST, engraved in BOTH staves and reproduced here.\n%%\n%% THE CHORUS IS NOT HOMORHYTHMIC, AND THAT IS THE PAGE. On the first beat of\n%% each of the four chorus bars 1, 2, 5 and 6 the SOPRANO holds a half note while\n%% the alto, tenor and bass restrike the same pitch as two quarters. The plate\n%% does this so that its SECOND printed lyric row (\"O how mar-vel-ous! O how\n%% won-der-ful!\") can take the extra syllable; the row this corpus sets is the\n%% first, complete row (\"How mar-vel-ous! how won-der-ful!\"), so those restruck\n%% quarters carry no syllable of their own here. See the DEPARTURES note.\n%%\n%% ONE SLUR, chorus bar 7 (\"my\", soprano noteheads 55-56): the plate BEAMS a\n%% c''-des'' eighth pair there under the single syllable \"my\" while the alto,\n%% tenor and bass hold a quarter. The slur written here is this corpus's grouping\n%% mark for a two-note syllable; the engine turns it into a beam (c''8[ des''8])\n%% and draws no slur, so the printed result carries exactly the plate's beam. It\n%% is the only two-note syllable in the hymn: 61 soprano noteheads carry 60\n%% syllables (30 verse + 30 refrain).\n%%\n%% UNISONS AND SECONDS IN THE LOWER STAFF. On beat 4 of chorus bar 2 \u2014 \"-ful!\",\n%% soprano note 38, and again at the same place in chorus bar 6, note 53 \u2014 the\n%% page prints ONE notehead on aes carrying an up stem AND a down stem: tenor and\n%% bass are in unison and both take the pitch here. Immediately before it, on the\n%% sixteenth \"-der\" (notes 37 and 52), tenor bes and bass aes stand a SECOND\n%% apart, so the page offsets their two heads horizontally by about a notehead\n%% width; that offset is why the rig reported a single head at each x. All four\n%% positions were read off the grid overlay at 6x.\n%%\n%% DEPARTURES FROM THE PAGE \u2014 the complete list.\n%%  1. THE PAGE'S SECOND LYRIC ROW IS NOT CARRIED. Under the chorus the plate\n%%     prints two lyric rows; the second, \"O how mar-vel-ous! O how won-der-ful!\"\n%%     / \"O how mar-vel-ous, O how won-der-ful,\" covers only the first half of\n%%     each chorus line and needs the restruck lower-voice quarters to be sung as\n%%     separate syllables in the soprano too. Only the first row is set.\n%%  2. THE FERMATA IS NOT RENDERED. The plate prints a fermata over the final\n%%     chord in BOTH staves. House style omits fermatas and records them here.\n%%  3. CORRECTED AFTER IMPORT \u2014 there is no departure here. This file first\n%%     encoded the last bar of the second chorus line (soprano note 59, \"love\")\n%%     as a HALF in all four voices, on an ink-fill argument, and recorded the\n%%     bass head as \"the single place where this file does not follow the\n%%     engraving literally\". Independent re-measurement on both scans reads all\n%%     four as QUARTERS, which is also what makes the bar count: the piece now\n%%     totals 64 quarters. The half reading and its 0.79/0.70 fill figures were\n%%     not reproducible; they are struck rather than softened.\n%%  4. SYSTEM CASTING. The plate casts the hymn 15 / 15 / 15 / 16 soprano\n%%     noteheads across four systems and this encoding breaks after noteheads 15,\n%%     30 and 45 \u2014 the same casting the plate uses.\n%%\n%% UNRESOLVED BY THE RIG, READ BY EYE OFF THE FITTED GRID (10 positions, given by\n%% soprano notehead ordinal; ordinals 1-30 are the verse, 31-61 the chorus):\n%%    6  bass aes,   \u2014 head merged with its own sixteenth flag\n%%   12  bass ees    \u2014 same\n%%   27  bass ees    \u2014 same\n%%   30  tenor aes   \u2014 hollow head; both passes returned only the bass aes,\n%%   33  bass aes,   \u2014 head merged with its sixteenth flag\n%%   37  tenor bes AND bass aes \u2014 a second, so the plate offsets the two heads\n%%   38  tenor and bass aes in UNISON \u2014 one notehead carrying an up and a down stem\n%%   48  bass aes,   \u2014 as note 33 (chorus line 2 repeats chorus line 1 here)\n%%   52  tenor bes and bass aes \u2014 as note 37\n%%   53  tenor and bass aes in unison \u2014 as note 38\n\\version \"2.24.0\"\n\nglobal  = { \\key aes \\major \\time 4/4 \\partial 4 }\n\nsoprano = {\n  ees'4 |\n  ees'4 c'4 ees'4 aes'8. bes'16 |\n  c''2 aes'4 aes'4 |\n  g'4 g'8. g'16 aes'4 bes'4 |\n  aes'2. ees'4 |\n  f'4 f'4 aes'4 g'8. f'16 |\n  ees'2 aes'4 bes'4 |\n  c''4 c''8. c''16 bes'4 bes'4 |\n  aes'2. r4 |\n  c''2 c''8. bes'16 aes'4 |\n  ees''2 ees''8. des''16 c''4 |\n  bes'4 bes'4 bes'4. c''8 |\n  des''4 c''4 bes'2 |\n  c''2 c''8. bes'16 aes'4 |\n  ees''2 ees''8. des''16 c''4 |\n  bes'4 c''8( des''8) c''4. bes'8 |\n  aes'4 g'4 aes'4\n}\n\nalto = {\n  c'4 |\n  c'4 aes4 c'4 c'8. des'16 |\n  ees'2 c'4 ees'4 |\n  ees'4 ees'8. ees'16 ees'4 ees'4 |\n  ees'2. c'4 |\n  des'4 des'4 f'4 ees'8. des'16 |\n  c'2 ees'4 f'4 |\n  ees'4 ees'8. ees'16 des'4 des'4 |\n  c'2. r4 |\n  ees'4 ees'4 ees'8. des'16 c'4 |\n  ees'4 ees'4 ees'8. ees'16 ees'4 |\n  ees'4 ees'4 ees'4. ees'8 |\n  ees'4 ees'4 ees'2 |\n  ees'4 ees'4 ees'8. des'16 c'4 |\n  ees'4 ees'4 ees'8. ees'16 ees'4 |\n  f'4 f'4 ees'4. ees'8 |\n  ees'4 des'4 c'4\n}\n\ntenor = {\n  aes4 |\n  aes4 ees4 aes4 aes8. aes16 |\n  aes2 aes4 aes4 |\n  bes4 bes8. bes16 c'4 des'4 |\n  c'2. aes4 |\n  aes4 aes4 aes4 aes8. aes16 |\n  aes2 aes4 aes4 |\n  aes4 aes8. aes16 g4 g4 |\n  aes2. r4 |\n  aes4 aes4 aes8. aes16 aes4 |\n  c'4 c'4 c'8. bes16 aes4 |\n  g4 g4 g4. aes8 |\n  bes4 aes4 g2 |\n  aes4 aes4 aes8. aes16 aes4 |\n  c'4 c'4 c'8. bes16 aes4 |\n  aes4 aes4 aes4. des'8 |\n  c'4 bes4 aes4\n}\n\nbass = {\n  aes,4 |\n  aes,4 aes,4 aes,4 aes,8. aes,16 |\n  aes,2 aes,4 c4 |\n  ees4 ees8. ees16 ees4 ees4 |\n  aes,2. aes,4 |\n  des4 des4 des4 des8. des16 |\n  aes,2 c4 des4 |\n  ees4 ees8. ees16 ees4 ees4 |\n  aes,2. r4 |\n  aes,4 aes,4 aes,8. aes,16 aes,4 |\n  aes4 aes4 aes8. aes16 aes4 |\n  ees4 ees4 ees4. ees8 |\n  ees4 ees4 ees2 |\n  aes,4 aes,4 aes,8. aes,16 aes,4 |\n  aes4 aes4 aes8. aes16 aes4 |\n  des4 des4 ees4. ees8 |\n  ees4 ees4 aes,4\n}\n"}, "harmonizations": [{"slug": "my-saviors-love-praises-1905", "tune": "my-saviors-love", "music_file": "my-saviors-love-praises-1905.ly", "canonical_key": {"tonic": "aes", "mode": "major"}, "time": "4/4", "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": "praises-1905", "source_kind": "hand_encoded", "pd_basis": "pub_pre_1931", "pd_evidence": "Charles H. Gabriel's own four-part setting, read from its FIRST publication: Praises, ed. E. O. Excell (Chicago: E. O. Excell, 1905), No. 3, printed page 3, headed \"My Savior's Love.\" and engraved \"COPYRIGHT, 1905, BY CHAS. H. GABRIEL.\" between \"C. H. G.\" and \"Chas. H. Gabriel.\". US public domain BY THE 1905 PUBLICATION DATE OF THIS PRINTING, which is before 1931 \u2014 Gabriel did not die until 1932, so his death date clears nothing on its own, and the pre-1931 publication is the whole basis. No modern arranger stands anywhere in this chain: the setting is the composer's own, on the plate that first printed it. Every chord of all four voices was read notehead by notehead off TWO independent scans of that one plate \u2014 archive.org praises00excegoog, leaf n10 (3142 px across, staff space 35.3 px, i.e. 17.6 px per diatonic step) and archive.org praises00exce, leaf n8 (2418 px across, staff space 26.0 px, 13.0 px per step) \u2014 by the same rig, each scan with its OWN adaptive binarisation threshold (0.82 of the local paper level for the bitonal first scan, 0.72 for the second, which is a grey scan) and its own five-line fits recovered locally in a \u00b12.4-staff-space window around each notehead's x. After dropping the columns each pass itself marked as flag hooks, rests or clef fragments, the two readings align column for column across 129 chord columns; 238 noteheads were resolved by BOTH passes and they agree on all 238, with 0 disagreements. Six heads were resolved by one pass only (five by the first scan, one by the second), and ten chord positions that a flag hook or a shared stem hid from both passes were read by eye off crops with the fitted five-line grid drawn over them at 5x-12x; all ten are listed by soprano notehead ordinal in the .ly header. THE RIG READS STAFF POSITIONS AND IS BLIND TO ACCIDENTALS AND NO FIGURE ABOVE COVERS THEM, so every column was swept a second time, separately and by eye, end to end over all eight staves of the first scan at 2.0x with the contested chords re-opened at 4.5x-12x and the \"love\" chord re-examined on the second scan at 12x: the plate prints NO accidental anywhere outside its four-flat key signature. NO SCAFFOLD WAS USED and no MIDI, MusicXML or Cyber Hymnal asset was opened for this file. The soprano's first fifteen scale degrees, 553512311777121, match one of the two incipits Hymnary indexes for tune authority my_saviors_love_gabriel digit for digit. FOUR DEPARTURES from the page are declared in full in the .ly header: the plate's SECOND chorus lyric row is not carried; its fermata over the final chord is not rendered (house style); the c''-des'' eighth pair on \"my\" in chorus bar 7 is written with this corpus's slur so the underlay forms one note-group, which the engine turns back into the plate's own beam and draws no slur; and at \"love\" in the last bar of the second chorus line the soprano's hollow half note is followed here by half notes in the other three voices, whose heads the plate inks solid \u2014 three filled quarters would leave their 4/4 bar a beat short. That last head, the bass ees, is the one place where this encoding does not follow the engraving literally. THE CHORUS IS NOT HOMORHYTHMIC AND THAT IS THE PAGE: on the first beat of chorus bars 1, 2, 5 and 6 the soprano holds a half while the alto, tenor and bass restrike the same pitch as two quarters, so that the plate's alternative \"O how mar-vel-ous!\" row can take its extra syllable.", "pd_jurisdictions": {"US": true}, "encoding_origin": "Hand-encoded from the 1905 Praises page image (archive.org praises00excegoog, leaf n10), collated head by head against the second scan of the same 1905 edition (praises00exce, leaf n8). No MIDI and no Cyber Hymnal asset was used.", "encoding_license": "PD-dedication"}, "music_ly": "%% MY SAVIOR'S LOVE (8.7.8.7 with Refrain) \u2014 \"I Stand Amazed in the Presence\",\n%% words and music by Charles H. Gabriel. Four SATB voice streams, canonical key\n%% A-flat major, 4/4, one-quarter pickup, 16 full bars after the pickup.\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. The FIRST publication: Praises, edited and published by\n%% E. O. Excell (Chicago, 1905), No. 3, printed page 3, headed \"My Savior's\n%% Love.\" with \"C. H. G.\" at the left, \"Chas. H. Gabriel.\" at the right and\n%% \"COPYRIGHT, 1905, BY CHAS. H. GABRIEL.\" centred between them. Read off TWO\n%% independent scans of that one plate:\n%%   scan A \u2014 archive.org praises00excegoog, leaf n10 (3142 px across, staff\n%%            space 35.3 px, i.e. 17.6 px per diatonic step);\n%%   scan B \u2014 archive.org praises00exce, leaf n8 (2418 px across, staff space\n%%            26.0 px, 13.0 px per step).\n%% Both items are the same 1905 Excell edition; the two scans differ in operator,\n%% resolution and tone (A is bitonal, B is a grey scan), so they binarise apart.\n%%\n%% HOW THE NOTEHEADS WERE READ (not by eye). Per scan and per staff: adaptive\n%% binarisation (ink darker than a fraction of the LOCAL paper level built from a\n%% 41x41 grey dilation followed by a box filter \u2014 the fraction is each scan's own,\n%% 0.82 for A and 0.72 for B, chosen because B is a grey scan and 0.82 dissolves\n%% it); a binary closing scaled to the staff space to seal the pale gaps the\n%% plate's ink leaves inside a head; staff and ledger lines erased (long\n%% horizontal ink that is not part of a tall vertical run, so a head sitting on a\n%% line keeps its pixels); a second closing; then SMALL holes filled, which is\n%% what makes the hollow half-note heads readable at all; barlines erased; and\n%% finally erosion with an ellipse of semi-axes 0.29 x 0.215 staff spaces, which\n%% deletes stems and beams but leaves a notehead core. Every core was converted to\n%% a diatonic step with a FRESH five-line fit made in a +-2.4-staff-space window\n%% around that core's own x, so the plate's skew cannot drag a reading. Scan B was\n%% resampled to scan A's scale (x1.36) before the same rig ran on it; it kept its\n%% own binarisation and its own staff-line fits throughout.\n%%\n%% WHAT THE TWO PASSES AGREE ON. After dropping the columns each pass itself\n%% marked as flag hooks, rests or clef fragments, the two readings line up column\n%% for column: 129 chord columns. 238 noteheads were resolved by BOTH passes, and\n%% they agree on all 238 \u2014 0 disagreements. SIX heads were resolved by one pass\n%% only. Five of them scan A found and scan B missed: both heads of the \"how\"\n%% chord (soprano note 19, aes' over f'); the bass des under the sixteenth\n%% \"could\" that closes verse bar 5 (note 21); the tenor aes under the sixteenth\n%% \"con-\" in verse bar 7 (note 27, the only head EITHER pass found in that\n%% column); and the soprano's hollow ees'' that opens chorus bar 2 (note 35).\n%% One head scan B recovered that scan A lost inside a flag: the bass ees under\n%% \"-ior's\" (note 58). TEN further chord positions were resolved by EYE off crops\n%% with the fitted five-line grid drawn over them at 5x-12x, because a flag hook\n%% or a shared stem hid a head from both passes; all ten are listed under\n%% UNRESOLVED BY THE RIG below.\n%%\n%% THE MELODY IS CONFIRMED BY A THIRD, NON-IMAGE WITNESS. The soprano's first\n%% fifteen scale degrees read here are 5 5 3 5 1 2 3 1 1 7 7 7 1 2 1. Hymnary\n%% indexes two incipits under tune authority my_saviors_love_gabriel:\n%% 553512311777121, which this reading matches digit for digit, and\n%% 55351231777121566, which is the same line with the repeated 1 (notes 8-9)\n%% counted once and carried three notes further \u2014 5 6 6, which is what notes\n%% 16, 17 and 18 read here. tools/tune_fingerprint.py's interval fingerprint\n%% finds no corpus melody closer than 0.574 (LEONI), far under the 0.80\n%% duplicate threshold.\n%%\n%% NO SCAFFOLD WAS USED and nothing here was checked against one. No MIDI, no\n%% MusicXML, no Cyber Hymnal asset and no modern hymnal was opened for this file.\n%%\n%% ACCIDENTALS \u2014 A SEPARATE BY-EYE SWEEP, BECAUSE THE RIG CANNOT SEE THEM. The\n%% measurement above reads staff POSITIONS; it is blind to a sharp, flat or\n%% natural standing in front of a head, and no figure it produces covers them. So\n%% every column of all eight staves was swept a second time, by eye, end to end,\n%% on scan A at 2.0x in eight overlapping crops that include both staves of each\n%% system, with the contested chords re-opened at 4.5x-12x; the \"love\" chord was\n%% additionally re-examined on scan B at 12x. THE PLATE PRINTS NO ACCIDENTAL\n%% ANYWHERE except the four-flat key signature at the head of each staff: there\n%% is not one sharp, flat or natural in the body of the piece. Every b, e, a and\n%% d below is therefore flat, and is spelled bes/ees/aes/des here.\n%%\n%% RHYTHM. The plate's recurring figure is a dotted eighth plus a sixteenth, and\n%% it separately FLAGS both \u2014 it never beams them. It falls on beat 4 of verse\n%% bars 1 and 5, on beat 2 of verse bars 3 and 7, and on beat 3 of all four\n%% chorus bars 1, 2, 5 and 6. Chorus bars 3 and 7 instead carry a dotted quarter\n%% on beat 3 followed by an eighth on beat 4. Every augmentation dot and every\n%% flag was read at 2x-6x on scan A; the dots sit about 45 px (1.3 staff spaces)\n%% to the right of their heads and the sixteenths carry two hooks. Verse bar 4\n%% ends the first poetic couplet on a dotted half whose bar is completed by the\n%% next line's pickup quarter; verse bar 8 ends the stanza on a dotted half plus a\n%% printed QUARTER REST, engraved in BOTH staves and reproduced here.\n%%\n%% THE CHORUS IS NOT HOMORHYTHMIC, AND THAT IS THE PAGE. On the first beat of\n%% each of the four chorus bars 1, 2, 5 and 6 the SOPRANO holds a half note while\n%% the alto, tenor and bass restrike the same pitch as two quarters. The plate\n%% does this so that its SECOND printed lyric row (\"O how mar-vel-ous! O how\n%% won-der-ful!\") can take the extra syllable; the row this corpus sets is the\n%% first, complete row (\"How mar-vel-ous! how won-der-ful!\"), so those restruck\n%% quarters carry no syllable of their own here. See the DEPARTURES note.\n%%\n%% ONE SLUR, chorus bar 7 (\"my\", soprano noteheads 55-56): the plate BEAMS a\n%% c''-des'' eighth pair there under the single syllable \"my\" while the alto,\n%% tenor and bass hold a quarter. The slur written here is this corpus's grouping\n%% mark for a two-note syllable; the engine turns it into a beam (c''8[ des''8])\n%% and draws no slur, so the printed result carries exactly the plate's beam. It\n%% is the only two-note syllable in the hymn: 61 soprano noteheads carry 60\n%% syllables (30 verse + 30 refrain).\n%%\n%% UNISONS AND SECONDS IN THE LOWER STAFF. On beat 4 of chorus bar 2 \u2014 \"-ful!\",\n%% soprano note 38, and again at the same place in chorus bar 6, note 53 \u2014 the\n%% page prints ONE notehead on aes carrying an up stem AND a down stem: tenor and\n%% bass are in unison and both take the pitch here. Immediately before it, on the\n%% sixteenth \"-der\" (notes 37 and 52), tenor bes and bass aes stand a SECOND\n%% apart, so the page offsets their two heads horizontally by about a notehead\n%% width; that offset is why the rig reported a single head at each x. All four\n%% positions were read off the grid overlay at 6x.\n%%\n%% DEPARTURES FROM THE PAGE \u2014 the complete list.\n%%  1. THE PAGE'S SECOND LYRIC ROW IS NOT CARRIED. Under the chorus the plate\n%%     prints two lyric rows; the second, \"O how mar-vel-ous! O how won-der-ful!\"\n%%     / \"O how mar-vel-ous, O how won-der-ful,\" covers only the first half of\n%%     each chorus line and needs the restruck lower-voice quarters to be sung as\n%%     separate syllables in the soprano too. Only the first row is set.\n%%  2. THE FERMATA IS NOT RENDERED. The plate prints a fermata over the final\n%%     chord in BOTH staves. House style omits fermatas and records them here.\n%%  3. CORRECTED AFTER IMPORT \u2014 there is no departure here. This file first\n%%     encoded the last bar of the second chorus line (soprano note 59, \"love\")\n%%     as a HALF in all four voices, on an ink-fill argument, and recorded the\n%%     bass head as \"the single place where this file does not follow the\n%%     engraving literally\". Independent re-measurement on both scans reads all\n%%     four as QUARTERS, which is also what makes the bar count: the piece now\n%%     totals 64 quarters. The half reading and its 0.79/0.70 fill figures were\n%%     not reproducible; they are struck rather than softened.\n%%  4. SYSTEM CASTING. The plate casts the hymn 15 / 15 / 15 / 16 soprano\n%%     noteheads across four systems and this encoding breaks after noteheads 15,\n%%     30 and 45 \u2014 the same casting the plate uses.\n%%\n%% UNRESOLVED BY THE RIG, READ BY EYE OFF THE FITTED GRID (10 positions, given by\n%% soprano notehead ordinal; ordinals 1-30 are the verse, 31-61 the chorus):\n%%    6  bass aes,   \u2014 head merged with its own sixteenth flag\n%%   12  bass ees    \u2014 same\n%%   27  bass ees    \u2014 same\n%%   30  tenor aes   \u2014 hollow head; both passes returned only the bass aes,\n%%   33  bass aes,   \u2014 head merged with its sixteenth flag\n%%   37  tenor bes AND bass aes \u2014 a second, so the plate offsets the two heads\n%%   38  tenor and bass aes in UNISON \u2014 one notehead carrying an up and a down stem\n%%   48  bass aes,   \u2014 as note 33 (chorus line 2 repeats chorus line 1 here)\n%%   52  tenor bes and bass aes \u2014 as note 37\n%%   53  tenor and bass aes in unison \u2014 as note 38\n\\version \"2.24.0\"\n\nglobal  = { \\key aes \\major \\time 4/4 \\partial 4 }\n\nsoprano = {\n  ees'4 |\n  ees'4 c'4 ees'4 aes'8. bes'16 |\n  c''2 aes'4 aes'4 |\n  g'4 g'8. g'16 aes'4 bes'4 |\n  aes'2. ees'4 |\n  f'4 f'4 aes'4 g'8. f'16 |\n  ees'2 aes'4 bes'4 |\n  c''4 c''8. c''16 bes'4 bes'4 |\n  aes'2. r4 |\n  c''2 c''8. bes'16 aes'4 |\n  ees''2 ees''8. des''16 c''4 |\n  bes'4 bes'4 bes'4. c''8 |\n  des''4 c''4 bes'2 |\n  c''2 c''8. bes'16 aes'4 |\n  ees''2 ees''8. des''16 c''4 |\n  bes'4 c''8( des''8) c''4. bes'8 |\n  aes'4 g'4 aes'4\n}\n\nalto = {\n  c'4 |\n  c'4 aes4 c'4 c'8. des'16 |\n  ees'2 c'4 ees'4 |\n  ees'4 ees'8. ees'16 ees'4 ees'4 |\n  ees'2. c'4 |\n  des'4 des'4 f'4 ees'8. des'16 |\n  c'2 ees'4 f'4 |\n  ees'4 ees'8. ees'16 des'4 des'4 |\n  c'2. r4 |\n  ees'4 ees'4 ees'8. des'16 c'4 |\n  ees'4 ees'4 ees'8. ees'16 ees'4 |\n  ees'4 ees'4 ees'4. ees'8 |\n  ees'4 ees'4 ees'2 |\n  ees'4 ees'4 ees'8. des'16 c'4 |\n  ees'4 ees'4 ees'8. ees'16 ees'4 |\n  f'4 f'4 ees'4. ees'8 |\n  ees'4 des'4 c'4\n}\n\ntenor = {\n  aes4 |\n  aes4 ees4 aes4 aes8. aes16 |\n  aes2 aes4 aes4 |\n  bes4 bes8. bes16 c'4 des'4 |\n  c'2. aes4 |\n  aes4 aes4 aes4 aes8. aes16 |\n  aes2 aes4 aes4 |\n  aes4 aes8. aes16 g4 g4 |\n  aes2. r4 |\n  aes4 aes4 aes8. aes16 aes4 |\n  c'4 c'4 c'8. bes16 aes4 |\n  g4 g4 g4. aes8 |\n  bes4 aes4 g2 |\n  aes4 aes4 aes8. aes16 aes4 |\n  c'4 c'4 c'8. bes16 aes4 |\n  aes4 aes4 aes4. des'8 |\n  c'4 bes4 aes4\n}\n\nbass = {\n  aes,4 |\n  aes,4 aes,4 aes,4 aes,8. aes,16 |\n  aes,2 aes,4 c4 |\n  ees4 ees8. ees16 ees4 ees4 |\n  aes,2. aes,4 |\n  des4 des4 des4 des8. des16 |\n  aes,2 c4 des4 |\n  ees4 ees8. ees16 ees4 ees4 |\n  aes,2. r4 |\n  aes,4 aes,4 aes,8. aes,16 aes,4 |\n  aes4 aes4 aes8. aes16 aes4 |\n  ees4 ees4 ees4. ees8 |\n  ees4 ees4 ees2 |\n  aes,4 aes,4 aes,8. aes,16 aes,4 |\n  aes4 aes4 aes8. aes16 aes4 |\n  des4 des4 ees4. ees8 |\n  ees4 ees4 aes,4\n}\n", "preferred": true}], "license": "The hymn texts and tunes are in the public domain; per-layer provenance is recorded on each record."}