{"slug": "eton", "name": "ETON", "meter_label": "8.7.8.7.4.7", "meter": {"pattern": [8, 7, 8, 7, 4, 7], "doubled": false, "refrain": false, "alleluia": false, "irregular": false}, "composer": "joseph-barnby", "canonical_key": {"tonic": "f", "mode": "major"}, "preferred_harmonization": "eton-methodist-1905", "provenance": {"layer": "tune", "source": "methodist-hymnal-1905", "source_kind": "hand_encoded", "pd_basis": "pub_pre_1931", "pd_evidence": "ETON, ascribed on the page to Joseph Barnby, read from The Methodist Hymnal (New York: Eaton & Mains; Cincinnati: Jennings & Graham, 1905), hymn 259, printed page 184 \u2014 archive.org methodi00meth leaf n195, collated against archive.org methodisthymnal01churgoog leaf n202. The header line reads \"259 ETON  8. 7. 8. 7. 4. 7.\" with \"JOSEPH BARNBY\" at the right margin; F major, 4/4, no pickup, 12 bars, 44 soprano noteheads for 41 syllables. THE PAGE PRINTS NO DATE FOR THE TUNE and none is asserted here. US public domain: the printing read is 1905, well before 1931, and no modern arranger stands anywhere in the chain. The same page sets the same text a second time as \"259 GREENVILLE 8s. 7s. 6l. (Second Tune)\" (Jean J. Rousseau); that setting is a different tune and is not encoded. The soprano was fingerprinted against all 291 corpus tunes before engraving (tools/tune_fingerprint.py): best match 0.60 (ANTIOCH), far below the 0.80 duplicate threshold, so this melody is new to the corpus.", "pd_jurisdictions": {"US": true}}, "harmonization": {"slug": "eton-methodist-1905", "tune": "eton", "music_file": "eton-methodist-1905.ly", "canonical_key": {"tonic": "f", "mode": "major"}, "time": "4/4", "pickup": null, "system_breaks": [16, 32], "breaks_manual": true, "final_barline": "|.", "amen": false, "verse_variable_melisma": false, "refrain": null, "alternates": [{"voice": "bass", "voice_note": 41, "pitch": "f,"}], "provenance": {"layer": "harmonization", "source": "methodist-hymnal-1905", "source_kind": "hand_encoded", "pd_basis": "pub_pre_1931", "pd_evidence": "Joseph Barnby's four-part setting of ETON as printed in The Methodist Hymnal (New York: Eaton & Mains; Cincinnati: Jennings & Graham, 1905), hymn 259, printed page 184, headed \"259  ETON  8. 7. 8. 7. 4. 7.\" over \"JOSEPH BARNBY\", in F major and 4/4 with no pickup, twelve bars. US public domain: the printing read is 1905, well before 1931, and no modern arranger stands anywhere in the chain. NO SCAFFOLD WAS USED AND NOTHING WAS CHECKED AGAINST ONE \u2014 no MIDI, no Cyber Hymnal asset and no existing encoding was opened at any point. TWO DATED SCANS OF THAT PLATE WERE READ END TO END AND REQUIRED TO AGREE: archive.org methodi00meth leaf n195 (2813 px across, greyscale, staff space 29.0 px) and archive.org methodisthymnal01churgoog leaf n202 (3270 px across, bitonal, staff space 34.6 px). THE NOTEHEADS WERE MEASURED, NOT EYEBALLED: each scan was binarised against its OWN local paper level (ink darker than 82% of a 41x41 grey-dilation + box filter) and closed 5x5 to seal the pale gaps inside a head; holes under 0.40 staff space^2 and under 0.85 x 0.70 staff spaces were filled so a HOLLOW head reads as a disc while the paper enclosed by staff lines and barlines does not; heads were then isolated by morphological erosion with an ellipse of semi-axes 0.30 x 0.215 staff spaces; cores taller than 1.15 staff spaces are merged upper/lower pairs and were split at the midpoint of their two clean outer edges with the separation snapped to a whole step; and every core was converted to a diatonic step against a FRESH five-line fit made in a +-2.4-staff-space window around that core's own x, so the plate's skew and curvature cannot drag a reading. Scan 1 resolved 150 notehead cores, 145 of them within 0.15 diatonic step of an exact staff position and the other five between 0.16 and 0.19; scan 2, read independently the same way with its own binarisation and its own staff-line fits, resolved 144, 138 of them within 0.15. Over the 78 note columns both passes resolved \u2014 139 noteheads \u2014 there are 0 disagreements. Voices were separated by the plate's own engraving: two heads in a staff are its upper and lower voice; one head carrying both an up and a down stem is a unison (this happens four times \u2014 soprano/alto on \"-y\" in m. 5, on \"pit-\" in m. 7 and on \"power:\" in m. 8, and tenor/bass on \"and\" in m. 4); one head carrying a single stem where another voice is holding is that stem's voice only. HOLLOW HEADS WERE READ A THIRD WAY REGARDLESS: all 19 half and whole heads in the hymn proper (the five columns at soprano notes 16, 32, 42, 43 and 44, across both staves) and the nine in the two engraved Amen chords were read off crops with the FITTED five-line grid drawn over them at 5x to 12x, on scan 1 and again on scan 2, and that by-eye reading is what ships; the same treatment was given to every core that measured outside 0.15 step. FINALLY THE COMMITTED .ly WAS PARSED BACK AND CHECKED AGAINST BOTH RIGS' READINGS: for each of the six staves, the measured columns must form a subsequence of the .ly's columns with every measured head present in the column it lands on. All twelve staff-by-scan runs pass, and the only three measured columns that do not sit inside a .ly column are accounted for: two are the engraved Amen chords, which this file deliberately does not carry, and the third is scan 2's reading of the lower hollow head of the \"no\" chord (m. 11, bass staff) at 3.61 steps \u2014 0.39 off any staff position, because the ring merges with the D3 line. Read on the grid at 9x on scan 1 and again on scan 2, that head sits clear of both neighbouring lines in the C3 space, and C3 is what is encoded. ACCIDENTALS WERE SWEPT SEPARATELY, BY EYE, AND THE AUTOMATED AGREEMENT FIGURE ABOVE COVERS NONE OF THEM \u2014 a notehead probe reads staff positions, not the sharps and naturals in front of them. Both leaves were swept column by column at high magnification and the page carries exactly SEVEN printed accidental glyphs, in six columns: alto B natural before \"and\" in m. 4; alto C sharp before \"-sus\" in m. 5; bass B natural before \"stands\" in m. 6; alto G sharp and soprano B natural before \"and\" in m. 8; alto F sharp on the first chord of m. 9; bass F sharp before \"is\" in m. 9. Two of them are carried into a second beat of the same bar BY THE BAR RULE and are not reprinted, exactly as the plate leaves them: the m. 5 C sharp governs beats 2 and 3 (A7, then a diminished seventh over B flat that resolves by step into the B flat chord on beat 4), and the m. 6 B natural governs beats 1 and 2 (the ascending bass B flat - B natural - B natural - C under G major and the cadential six-four). LilyPond's default accidental style reprints them in exactly the same places the plate does \u2014 that is, nowhere. \u26a0 ONE APPARENT SEVENTH ACCIDENTAL WAS REJECTED: the bitonal scan 2 shows a broken glyph in the lower staff immediately before the first chord of m. 7 (its x 1887-1923); the greyscale scan 1 shows clean paper at the same place, the shape is incoherent, and an accidental there would fall on the bass C, which needs none. It is scan-2 show-through, not ink. DEPARTURES FROM THE PAGE, in full: (1) the page's engraved Amen is not carried, see the `amen` comment in this file; (2) the low f, doubling the bass on the last chord is carried in `alternates` and prints cue-small, where the plate prints it full size; (3) the three two-note SOPRANO syllables (m. 3 a'-g' on \"and\", m. 7 a'-g' on \"of\", m. 10 g'-f' on \"-ing:\") are slurred here because that is how this corpus marks a two-note syllable \u2014 the plate beams them as eighth pairs and draws no slur, a notation change and not a pitch or duration change; (4) the thick rules the page draws at the ends of poetic lines 2 and 4 are ordinary barlines drawn heavy, and are reproduced as ordinary barlines with a system break at the same soprano ordinals. Nothing else departs: every pitch and every duration below is the plate's, the one slur the plate draws in a lower voice (tenor g( a) in m. 11) is kept, the plate's other beamed eighth pairs (tenor m. 7, alto m. 10, tenor m. 10) are left unslurred as it has them, and the casting into three systems is the page's own.", "pd_jurisdictions": {"US": true}, "encoding_origin": "Hand-encoded from the page images of The Methodist Hymnal 1905, hymn 259, p. 184 \u2014 archive.org methodi00meth leaf n195, collated notehead by notehead against archive.org methodisthymnal01churgoog leaf n202. No MIDI, no Cyber Hymnal asset and no existing encoding was used at any point.", "encoding_license": "PD-dedication"}, "music_ly": "%% ETON (8.7.8.7.4.7) \u2014 Joseph Barnby. Four SATB voice streams, canonical key\n%% F major, 4/4, no pickup, 12 measures.\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 Methodist Hymnal (New York: Eaton & Mains;\n%% Cincinnati: Jennings & Graham, 1905), hymn 259, printed page 184, headed\n%% \"259  ETON  8. 7. 8. 7. 4. 7.\" with \"JOSEPH BARNBY\" at the right and\n%% \"JOSEPH HART\" under the stanzas. It is the FIRST of two tunes the page sets\n%% to this text; a second setting, \"259 GREENVILLE 8s. 7s. 6l. (Second Tune)\"\n%% by Jean J. Rousseau, begins at the foot of the same page and is NOT encoded\n%% here. Read off TWO independent scans of that plate: archive.org\n%% methodi00meth leaf n195 (2813 px across, greyscale) and archive.org\n%% methodisthymnal01churgoog leaf n202 (3270 px across, bitonal).\n%%\n%% HOW THE NOTEHEADS WERE READ (not by eye). Per scan, with its OWN\n%% binarisation and its OWN staff-line fits: ink = darker than 82% of the LOCAL\n%% paper level (41x41 grey-dilation + box filter), then a 5x5 binary closing to\n%% seal the pale gaps the plate's ink leaves inside a head; holes smaller than\n%% 0.40 staff space^2 AND under 0.85 x 0.70 staff spaces filled, which seals a\n%% HOLLOW head's interior without flooding the paper enclosed by staff lines and\n%% barlines; noteheads then isolated by morphological erosion with an ellipse of\n%% semi-axes 0.30 x 0.215 staff spaces, which deletes staff lines, stems and\n%% beams. A core taller than 1.15 staff spaces is a MERGED soprano-over-alto (or\n%% tenor-over-bass) pair and was split at the midpoint of its two clean outer\n%% edges, the separation snapping to a whole diatonic step. Every core centroid\n%% was converted to a diatonic step against a FRESH five-line fit made in a\n%% +-2.4-staff-space window around that core's own x, so the plate's skew and\n%% curvature cannot drag a reading. Staff space is 29.0 px on scan 1 and 34.6 px\n%% on scan 2 \u2014 14.5 and 17.3 px per diatonic step.\n%%\n%% THE SECOND WITNESS. Scan 2 was read end to end by the same rig with its own\n%% binarisation and its own staff-line fits, and the two readings were aligned\n%% by a single linear x-map fitted on the barlines. The automated pass resolved\n%% 150 notehead cores on scan 1 (145 of them within 0.15 diatonic step of an\n%% exact staff position, the other five between 0.16 and 0.19) and 144 on scan\n%% 2 (138 within 0.15). Over the 78 note columns BOTH passes resolved, covering\n%% 139 noteheads, there are ZERO pitch disagreements. Where the automated pass\n%% failed \u2014 the pairs an accidental glyph merges into, and the hollow heads \u2014\n%% the note was settled by the third, by-eye reading below, and every core that\n%% measured outside 0.15 step was re-read that way too.\n%%\n%% HOLLOW HEADS READ A THIRD WAY. Half and whole heads are outlines, not blobs.\n%% All 19 of them in the hymn proper \u2014 the five columns at soprano notes 16, 32,\n%% 42, 43 and 44, across both staves \u2014 and the nine in the two engraved Amen\n%% chords were read off crops with the FITTED five-line grid drawn over them at\n%% 5x to 12x, on scan 1 and again on scan 2, and THAT reading is what ships.\n%%\n%% ACCIDENTALS WERE SWEPT SEPARATELY, BY EYE. A notehead probe reads heads, not\n%% the sharps and naturals in front of them, and the automated agreement figure\n%% above covers NO accidental. Both leaves were swept column by column at high\n%% magnification, and the seven printed accidental glyphs (six columns) are, in reading order:\n%%   m. 4  alto  B natural  (before \"and\", soprano note 15)\n%%   m. 5  alto  C sharp    (before \"-sus\",  soprano note 18)\n%%   m. 6  bass  B natural   (before \"stands\", soprano note 21)\n%%   m. 8  alto  G sharp and soprano B natural (before \"and\", soprano note 31)\n%%   m. 9  alto  F sharp    (on the first chord of the system)\n%%   m. 9  bass  F sharp    (before \"is\", soprano note 34)\n%% Two of them are RESTATED HERE ONLY BY THE BAR RULE, exactly as the plate\n%% leaves them: the alto's C sharp in m. 5 is printed once and governs both the\n%% 2nd and the 3rd beat (Barnby's A7 then a diminished seventh over B flat that\n%% resolves into the B flat chord), and the bass's B natural in m. 6 is printed\n%% once and governs both the 1st and the 2nd beat (the ascending bass B flat -\n%% B natural - B natural - C under G major and the cadential six-four).\n%% LilyPond's default accidental style reprints them in exactly the same places\n%% the plate does \u2014 that is, not at all.\n%%\n%% RHYTHM. 4/4 throughout, no pickup, 12 full bars; the hymn is metrically\n%% complete. All four voices restrike together except at five places, all of\n%% which the plate prints: m. 2 the soprano alone has a dotted quarter + eighth\n%% over four moving quarters below; m. 3 and m. 7 the soprano has an eighth pair\n%% (in m. 7 the tenor has one too, on the same two beats); m. 10 the alto has an\n%% eighth pair on beat 3 while the soprano and tenor have theirs on beat 4; and\n%% m. 11 the tenor moves g-a under the upper voices' half notes.\n%%\n%% THE SLURS. Only one slur on this page belongs to a lower voice: the tenor's\n%% g( a) in m. 11, which the plate draws. The three SOPRANO slurs (m. 3 a'-g',\n%% m. 7 a'-g', m. 10 g'-f') are this corpus's mark for a two-note syllable; the\n%% plate BEAMS those pairs and prints no slur over them. That is a departure in\n%% notation, not in pitch or duration. The plate's other beamed eighth pairs \u2014\n%% the tenor in m. 7, the alto in m. 10, the tenor in m. 10 \u2014 carry no lyric\n%% group and are left unslurred, as the plate has them.\n%%\n%% THE PAGE'S ENGRAVED AMEN IS NOT CARRIED HERE. After a thin-thick double bar\n%% the plate prints two whole-note chords under \"A-MEN.\": B flat major\n%% (S f' / A d' / T bes / B bes,) then F major (S f' / A c' / T a / B f, with a\n%% second, lower f, head an octave below). This corpus never authors an amen per\n%% harmonization \u2014 amens are synthesized as an optional book feature \u2014 so\n%% `amen: false` and those two chords are simply absent.\n%%\n%% THE LOW OCTAVE ON THE LAST CHORD. Under \"more.\" the plate's lower staff\n%% prints THREE heads \u2014 a, f and f, \u2014 the lowest a full-size octave doubling of\n%% the bass. Divisi is forbidden here, so f, is carried in the sidecar's\n%% `alternates` (bass note 41) and the engine prints it cue-small; the plate\n%% prints it full size. That size is the only departure; the pitch is the\n%% plate's.\n%%\n%% 44 soprano noteheads carry 41 syllables (8+7+8+7+4+7), the three two-note\n%% syllables named above accounting for the difference.\n\\version \"2.24.0\"\n\nglobal  = { \\key f \\major \\time 4/4 }\n\nsoprano = {\n  f'4 e'4 d'4 c'4 |\n  a'4. g'8 f'4 e'4 |\n  bes'4 a'8( g'8) a'4 g'4 |\n  f'4 d'4 e'2 |\n  f'4 e'4 e'4 d'4 |\n  a'4 g'4 f'4 e'4 |\n  bes'4 a'8( g'8) f'4 a'4 |\n  a'4 b'4 a'2 |\n  a'4 d''4 c''4 bes'4 |\n  c''4 bes'4 a'4 g'8( f'8) |\n  f'2 e'2 |\n  f'1\n}\n\nalto = {\n  c'4 c'4 bes4 bes4 |\n  a4 d'4 c'4 c'4 |\n  bes4 c'4 c'4 c'4 |\n  c'4 b4 c'2 |\n  a4 cis'4 cis'4 d'4 |\n  d'4 d'4 c'4 c'4 |\n  e'4 e'4 f'4 d'4 |\n  e'4 gis'4 a'2 |\n  fis'4 a'4 a'4 g'4 |\n  g'4 e'4 f'8 e'8 d'4 |\n  c'2 c'2 |\n  c'1\n}\n\ntenor = {\n  a4 g4 f4 g4 |\n  a4 bes4 a4 g4 |\n  f4 bes4 a4 c'4 |\n  a4 g4 g2 |\n  f4 g4 g4 f4 |\n  f4 g4 a4 g4 |\n  g4 a8 bes8 a4 a4 |\n  c'4 d'4 c'2 |\n  d'4 d'4 d'4 d'4 |\n  c'4 c'4 c'4 bes8 a8 |\n  g4( a4) bes2 |\n  a1\n}\n\nbass = {\n  f4 c4 d4 e4 |\n  f4 bes,4 c4 c4 |\n  d4 e4 f4 e4 |\n  d4 g4 c2 |\n  d4 a,4 bes,4 bes,4 |\n  b,4 b,4 c4 c4 |\n  c4 c4 d4 f4 |\n  e4 e4 a2 |\n  d4 fis4 g4 g4 |\n  e4 c4 f4 bes,4 |\n  c2 c2 |\n  f1\n}\n"}, "harmonizations": [{"slug": "eton-methodist-1905", "tune": "eton", "music_file": "eton-methodist-1905.ly", "canonical_key": {"tonic": "f", "mode": "major"}, "time": "4/4", "pickup": null, "system_breaks": [16, 32], "breaks_manual": true, "final_barline": "|.", "amen": false, "verse_variable_melisma": false, "refrain": null, "alternates": [{"voice": "bass", "voice_note": 41, "pitch": "f,"}], "provenance": {"layer": "harmonization", "source": "methodist-hymnal-1905", "source_kind": "hand_encoded", "pd_basis": "pub_pre_1931", "pd_evidence": "Joseph Barnby's four-part setting of ETON as printed in The Methodist Hymnal (New York: Eaton & Mains; Cincinnati: Jennings & Graham, 1905), hymn 259, printed page 184, headed \"259  ETON  8. 7. 8. 7. 4. 7.\" over \"JOSEPH BARNBY\", in F major and 4/4 with no pickup, twelve bars. US public domain: the printing read is 1905, well before 1931, and no modern arranger stands anywhere in the chain. NO SCAFFOLD WAS USED AND NOTHING WAS CHECKED AGAINST ONE \u2014 no MIDI, no Cyber Hymnal asset and no existing encoding was opened at any point. TWO DATED SCANS OF THAT PLATE WERE READ END TO END AND REQUIRED TO AGREE: archive.org methodi00meth leaf n195 (2813 px across, greyscale, staff space 29.0 px) and archive.org methodisthymnal01churgoog leaf n202 (3270 px across, bitonal, staff space 34.6 px). THE NOTEHEADS WERE MEASURED, NOT EYEBALLED: each scan was binarised against its OWN local paper level (ink darker than 82% of a 41x41 grey-dilation + box filter) and closed 5x5 to seal the pale gaps inside a head; holes under 0.40 staff space^2 and under 0.85 x 0.70 staff spaces were filled so a HOLLOW head reads as a disc while the paper enclosed by staff lines and barlines does not; heads were then isolated by morphological erosion with an ellipse of semi-axes 0.30 x 0.215 staff spaces; cores taller than 1.15 staff spaces are merged upper/lower pairs and were split at the midpoint of their two clean outer edges with the separation snapped to a whole step; and every core was converted to a diatonic step against a FRESH five-line fit made in a +-2.4-staff-space window around that core's own x, so the plate's skew and curvature cannot drag a reading. Scan 1 resolved 150 notehead cores, 145 of them within 0.15 diatonic step of an exact staff position and the other five between 0.16 and 0.19; scan 2, read independently the same way with its own binarisation and its own staff-line fits, resolved 144, 138 of them within 0.15. Over the 78 note columns both passes resolved \u2014 139 noteheads \u2014 there are 0 disagreements. Voices were separated by the plate's own engraving: two heads in a staff are its upper and lower voice; one head carrying both an up and a down stem is a unison (this happens four times \u2014 soprano/alto on \"-y\" in m. 5, on \"pit-\" in m. 7 and on \"power:\" in m. 8, and tenor/bass on \"and\" in m. 4); one head carrying a single stem where another voice is holding is that stem's voice only. HOLLOW HEADS WERE READ A THIRD WAY REGARDLESS: all 19 half and whole heads in the hymn proper (the five columns at soprano notes 16, 32, 42, 43 and 44, across both staves) and the nine in the two engraved Amen chords were read off crops with the FITTED five-line grid drawn over them at 5x to 12x, on scan 1 and again on scan 2, and that by-eye reading is what ships; the same treatment was given to every core that measured outside 0.15 step. FINALLY THE COMMITTED .ly WAS PARSED BACK AND CHECKED AGAINST BOTH RIGS' READINGS: for each of the six staves, the measured columns must form a subsequence of the .ly's columns with every measured head present in the column it lands on. All twelve staff-by-scan runs pass, and the only three measured columns that do not sit inside a .ly column are accounted for: two are the engraved Amen chords, which this file deliberately does not carry, and the third is scan 2's reading of the lower hollow head of the \"no\" chord (m. 11, bass staff) at 3.61 steps \u2014 0.39 off any staff position, because the ring merges with the D3 line. Read on the grid at 9x on scan 1 and again on scan 2, that head sits clear of both neighbouring lines in the C3 space, and C3 is what is encoded. ACCIDENTALS WERE SWEPT SEPARATELY, BY EYE, AND THE AUTOMATED AGREEMENT FIGURE ABOVE COVERS NONE OF THEM \u2014 a notehead probe reads staff positions, not the sharps and naturals in front of them. Both leaves were swept column by column at high magnification and the page carries exactly SEVEN printed accidental glyphs, in six columns: alto B natural before \"and\" in m. 4; alto C sharp before \"-sus\" in m. 5; bass B natural before \"stands\" in m. 6; alto G sharp and soprano B natural before \"and\" in m. 8; alto F sharp on the first chord of m. 9; bass F sharp before \"is\" in m. 9. Two of them are carried into a second beat of the same bar BY THE BAR RULE and are not reprinted, exactly as the plate leaves them: the m. 5 C sharp governs beats 2 and 3 (A7, then a diminished seventh over B flat that resolves by step into the B flat chord on beat 4), and the m. 6 B natural governs beats 1 and 2 (the ascending bass B flat - B natural - B natural - C under G major and the cadential six-four). LilyPond's default accidental style reprints them in exactly the same places the plate does \u2014 that is, nowhere. \u26a0 ONE APPARENT SEVENTH ACCIDENTAL WAS REJECTED: the bitonal scan 2 shows a broken glyph in the lower staff immediately before the first chord of m. 7 (its x 1887-1923); the greyscale scan 1 shows clean paper at the same place, the shape is incoherent, and an accidental there would fall on the bass C, which needs none. It is scan-2 show-through, not ink. DEPARTURES FROM THE PAGE, in full: (1) the page's engraved Amen is not carried, see the `amen` comment in this file; (2) the low f, doubling the bass on the last chord is carried in `alternates` and prints cue-small, where the plate prints it full size; (3) the three two-note SOPRANO syllables (m. 3 a'-g' on \"and\", m. 7 a'-g' on \"of\", m. 10 g'-f' on \"-ing:\") are slurred here because that is how this corpus marks a two-note syllable \u2014 the plate beams them as eighth pairs and draws no slur, a notation change and not a pitch or duration change; (4) the thick rules the page draws at the ends of poetic lines 2 and 4 are ordinary barlines drawn heavy, and are reproduced as ordinary barlines with a system break at the same soprano ordinals. Nothing else departs: every pitch and every duration below is the plate's, the one slur the plate draws in a lower voice (tenor g( a) in m. 11) is kept, the plate's other beamed eighth pairs (tenor m. 7, alto m. 10, tenor m. 10) are left unslurred as it has them, and the casting into three systems is the page's own.", "pd_jurisdictions": {"US": true}, "encoding_origin": "Hand-encoded from the page images of The Methodist Hymnal 1905, hymn 259, p. 184 \u2014 archive.org methodi00meth leaf n195, collated notehead by notehead against archive.org methodisthymnal01churgoog leaf n202. No MIDI, no Cyber Hymnal asset and no existing encoding was used at any point.", "encoding_license": "PD-dedication"}, "music_ly": "%% ETON (8.7.8.7.4.7) \u2014 Joseph Barnby. Four SATB voice streams, canonical key\n%% F major, 4/4, no pickup, 12 measures.\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 Methodist Hymnal (New York: Eaton & Mains;\n%% Cincinnati: Jennings & Graham, 1905), hymn 259, printed page 184, headed\n%% \"259  ETON  8. 7. 8. 7. 4. 7.\" with \"JOSEPH BARNBY\" at the right and\n%% \"JOSEPH HART\" under the stanzas. It is the FIRST of two tunes the page sets\n%% to this text; a second setting, \"259 GREENVILLE 8s. 7s. 6l. (Second Tune)\"\n%% by Jean J. Rousseau, begins at the foot of the same page and is NOT encoded\n%% here. Read off TWO independent scans of that plate: archive.org\n%% methodi00meth leaf n195 (2813 px across, greyscale) and archive.org\n%% methodisthymnal01churgoog leaf n202 (3270 px across, bitonal).\n%%\n%% HOW THE NOTEHEADS WERE READ (not by eye). Per scan, with its OWN\n%% binarisation and its OWN staff-line fits: ink = darker than 82% of the LOCAL\n%% paper level (41x41 grey-dilation + box filter), then a 5x5 binary closing to\n%% seal the pale gaps the plate's ink leaves inside a head; holes smaller than\n%% 0.40 staff space^2 AND under 0.85 x 0.70 staff spaces filled, which seals a\n%% HOLLOW head's interior without flooding the paper enclosed by staff lines and\n%% barlines; noteheads then isolated by morphological erosion with an ellipse of\n%% semi-axes 0.30 x 0.215 staff spaces, which deletes staff lines, stems and\n%% beams. A core taller than 1.15 staff spaces is a MERGED soprano-over-alto (or\n%% tenor-over-bass) pair and was split at the midpoint of its two clean outer\n%% edges, the separation snapping to a whole diatonic step. Every core centroid\n%% was converted to a diatonic step against a FRESH five-line fit made in a\n%% +-2.4-staff-space window around that core's own x, so the plate's skew and\n%% curvature cannot drag a reading. Staff space is 29.0 px on scan 1 and 34.6 px\n%% on scan 2 \u2014 14.5 and 17.3 px per diatonic step.\n%%\n%% THE SECOND WITNESS. Scan 2 was read end to end by the same rig with its own\n%% binarisation and its own staff-line fits, and the two readings were aligned\n%% by a single linear x-map fitted on the barlines. The automated pass resolved\n%% 150 notehead cores on scan 1 (145 of them within 0.15 diatonic step of an\n%% exact staff position, the other five between 0.16 and 0.19) and 144 on scan\n%% 2 (138 within 0.15). Over the 78 note columns BOTH passes resolved, covering\n%% 139 noteheads, there are ZERO pitch disagreements. Where the automated pass\n%% failed \u2014 the pairs an accidental glyph merges into, and the hollow heads \u2014\n%% the note was settled by the third, by-eye reading below, and every core that\n%% measured outside 0.15 step was re-read that way too.\n%%\n%% HOLLOW HEADS READ A THIRD WAY. Half and whole heads are outlines, not blobs.\n%% All 19 of them in the hymn proper \u2014 the five columns at soprano notes 16, 32,\n%% 42, 43 and 44, across both staves \u2014 and the nine in the two engraved Amen\n%% chords were read off crops with the FITTED five-line grid drawn over them at\n%% 5x to 12x, on scan 1 and again on scan 2, and THAT reading is what ships.\n%%\n%% ACCIDENTALS WERE SWEPT SEPARATELY, BY EYE. A notehead probe reads heads, not\n%% the sharps and naturals in front of them, and the automated agreement figure\n%% above covers NO accidental. Both leaves were swept column by column at high\n%% magnification, and the seven printed accidental glyphs (six columns) are, in reading order:\n%%   m. 4  alto  B natural  (before \"and\", soprano note 15)\n%%   m. 5  alto  C sharp    (before \"-sus\",  soprano note 18)\n%%   m. 6  bass  B natural   (before \"stands\", soprano note 21)\n%%   m. 8  alto  G sharp and soprano B natural (before \"and\", soprano note 31)\n%%   m. 9  alto  F sharp    (on the first chord of the system)\n%%   m. 9  bass  F sharp    (before \"is\", soprano note 34)\n%% Two of them are RESTATED HERE ONLY BY THE BAR RULE, exactly as the plate\n%% leaves them: the alto's C sharp in m. 5 is printed once and governs both the\n%% 2nd and the 3rd beat (Barnby's A7 then a diminished seventh over B flat that\n%% resolves into the B flat chord), and the bass's B natural in m. 6 is printed\n%% once and governs both the 1st and the 2nd beat (the ascending bass B flat -\n%% B natural - B natural - C under G major and the cadential six-four).\n%% LilyPond's default accidental style reprints them in exactly the same places\n%% the plate does \u2014 that is, not at all.\n%%\n%% RHYTHM. 4/4 throughout, no pickup, 12 full bars; the hymn is metrically\n%% complete. All four voices restrike together except at five places, all of\n%% which the plate prints: m. 2 the soprano alone has a dotted quarter + eighth\n%% over four moving quarters below; m. 3 and m. 7 the soprano has an eighth pair\n%% (in m. 7 the tenor has one too, on the same two beats); m. 10 the alto has an\n%% eighth pair on beat 3 while the soprano and tenor have theirs on beat 4; and\n%% m. 11 the tenor moves g-a under the upper voices' half notes.\n%%\n%% THE SLURS. Only one slur on this page belongs to a lower voice: the tenor's\n%% g( a) in m. 11, which the plate draws. The three SOPRANO slurs (m. 3 a'-g',\n%% m. 7 a'-g', m. 10 g'-f') are this corpus's mark for a two-note syllable; the\n%% plate BEAMS those pairs and prints no slur over them. That is a departure in\n%% notation, not in pitch or duration. The plate's other beamed eighth pairs \u2014\n%% the tenor in m. 7, the alto in m. 10, the tenor in m. 10 \u2014 carry no lyric\n%% group and are left unslurred, as the plate has them.\n%%\n%% THE PAGE'S ENGRAVED AMEN IS NOT CARRIED HERE. After a thin-thick double bar\n%% the plate prints two whole-note chords under \"A-MEN.\": B flat major\n%% (S f' / A d' / T bes / B bes,) then F major (S f' / A c' / T a / B f, with a\n%% second, lower f, head an octave below). This corpus never authors an amen per\n%% harmonization \u2014 amens are synthesized as an optional book feature \u2014 so\n%% `amen: false` and those two chords are simply absent.\n%%\n%% THE LOW OCTAVE ON THE LAST CHORD. Under \"more.\" the plate's lower staff\n%% prints THREE heads \u2014 a, f and f, \u2014 the lowest a full-size octave doubling of\n%% the bass. Divisi is forbidden here, so f, is carried in the sidecar's\n%% `alternates` (bass note 41) and the engine prints it cue-small; the plate\n%% prints it full size. That size is the only departure; the pitch is the\n%% plate's.\n%%\n%% 44 soprano noteheads carry 41 syllables (8+7+8+7+4+7), the three two-note\n%% syllables named above accounting for the difference.\n\\version \"2.24.0\"\n\nglobal  = { \\key f \\major \\time 4/4 }\n\nsoprano = {\n  f'4 e'4 d'4 c'4 |\n  a'4. g'8 f'4 e'4 |\n  bes'4 a'8( g'8) a'4 g'4 |\n  f'4 d'4 e'2 |\n  f'4 e'4 e'4 d'4 |\n  a'4 g'4 f'4 e'4 |\n  bes'4 a'8( g'8) f'4 a'4 |\n  a'4 b'4 a'2 |\n  a'4 d''4 c''4 bes'4 |\n  c''4 bes'4 a'4 g'8( f'8) |\n  f'2 e'2 |\n  f'1\n}\n\nalto = {\n  c'4 c'4 bes4 bes4 |\n  a4 d'4 c'4 c'4 |\n  bes4 c'4 c'4 c'4 |\n  c'4 b4 c'2 |\n  a4 cis'4 cis'4 d'4 |\n  d'4 d'4 c'4 c'4 |\n  e'4 e'4 f'4 d'4 |\n  e'4 gis'4 a'2 |\n  fis'4 a'4 a'4 g'4 |\n  g'4 e'4 f'8 e'8 d'4 |\n  c'2 c'2 |\n  c'1\n}\n\ntenor = {\n  a4 g4 f4 g4 |\n  a4 bes4 a4 g4 |\n  f4 bes4 a4 c'4 |\n  a4 g4 g2 |\n  f4 g4 g4 f4 |\n  f4 g4 a4 g4 |\n  g4 a8 bes8 a4 a4 |\n  c'4 d'4 c'2 |\n  d'4 d'4 d'4 d'4 |\n  c'4 c'4 c'4 bes8 a8 |\n  g4( a4) bes2 |\n  a1\n}\n\nbass = {\n  f4 c4 d4 e4 |\n  f4 bes,4 c4 c4 |\n  d4 e4 f4 e4 |\n  d4 g4 c2 |\n  d4 a,4 bes,4 bes,4 |\n  b,4 b,4 c4 c4 |\n  c4 c4 d4 f4 |\n  e4 e4 a2 |\n  d4 fis4 g4 g4 |\n  e4 c4 f4 bes,4 |\n  c2 c2 |\n  f1\n}\n", "preferred": true}], "license": "The hymn texts and tunes are in the public domain; per-layer provenance is recorded on each record."}