Reference

Data and method

Every number on this site is read from a file in site/data. The files are written by glyphstack.export_site.

107printed e's, from 7 cards
29glyphs with at least 12 instances
7.5 µmsingle-scan edge noise, 5.3 µm for the mean of two views
47.32scanner pixels per millimetre
0.40–0.56edge level that matched ink area across 14 scans

These come from one scanner and 7 pilot cards. The scanner blurs edges over 100 to 125 µm, so stacking reduces noise but does not recover finer detail.

Noise

Two views of the same printed glyph, scanned at different angles, differ by an edge RMS of 10.6 µm over 1018 pairs when each scan’s affine is fitted on its own. That gives 7.5 µm single-scan noise and 5.3 µm for a pair mean.

With scale and shear tied across a card’s two scans the same comparison gives 7.2 µm and 5.1 µm. The untied fit absorbs per-scan placement distortion, which is why the two differ.

1018 printed glyphs, each scanned twice. Dashed line: the middle pair, 9.3 µm. The RMS over all pairs is 10.6 µm.

Calibration

Ruler ticks on the calibration target give 47.32 px/mm, with x and y scales equal to 0.03 %. Edge widths measured on slanted printed edges range 93 to 131 µm over 24 edges, median 116 µm. The scanner pixel is 21 µm, so the blur is several pixels wide.

Scale from each ruler on the calibration target. Scan A has the sheet upright, scan B has it turned 90°.
Scan and paper axis	Ticks	px per mm	Tilt °	Fit residual px
Scan A, paper x	61	46.991	0.18	0.85
Scan A, paper y	61	47.325	-0.01	3.18
Scan B, paper x	61	47.006	0.54	2.83
Scan B, paper y	61	47.320	0.68	1.13

Grey level read for each printed tone step, 0 is black and 255 is white
Percent black printed	0	10	20	30	40	50	60	70	80	90	100
Scan A	255	219	150	108	81	57	43	34	30	28	24
Scan B	255	232	160	115	84	61	44	36	33	31	27

Edge levels

To reproduce the same pooled ink area, each of the 14 scans needed its own coverage level, from 0.40 to 0.56.

Files

Contents of site/data, 29 glyphs, 29 glyph files
File	Content	Size
glyphs.json	Glyph index: name, instances, cards, nodes, edge RMS, IoU	6838 B
glyphs/{name}.json	Per glyph: trace, mean contours 0.1 to 0.9, spread grid, up to 64 instance paths, cutaways
sweep/{name}.json	Edge RMS and noise against the number of stacked instances
kerning.json	Observed spacing for each adjacent pair, with the designed value	101 KB
noise.json	Comparison of views of the same printed glyph scanned at different angles	1405 B
calibration.json	Scale, isotropy, edge widths from the calibration scans	1726 B
cards.json	Cards, their scans, edge levels and print scale	2325 B

Glyphs

Statistics for each glyph, from glyphs.json. Edge RMS and IoU are against the measuring stick.
Glyph	Instances	Cards	Nodes	Edge RMS µm	IoU	Recommended N
d	20	7	73	6.14	0.973	16
l	34	7	18	5.62	0.972	16
t	74	7	25	6.89	0.964	16
a	66	6	45	6.24	0.968	16
c	29	5	42	4.86	0.971	16
s	33	7	40	5.46	0.966	16
r	64	7	29	8.73	0.964	16
n	57	7	39	6.84	0.969	16
e	107	7	35	5.20	0.973	16
p (low confidence)	13	4	79	7.10	0.963	13
o	46	7	53	2.21	0.986	16
i	33	7	27	8.13	0.952	16
u	31	7	40	6.69	0.965	16
y (low confidence)	12	3	93	8.46	0.950	12
h (low confidence)	14	5	55	6.84	0.968	14
full stop (low confidence)	15	7	6	8.06	0.930	15
h, italic (low confidence)	19	5	112	14.49	0.938	16
l, italic (low confidence)	17	5	68	9.86	0.933	16
t, italic	41	5	41	9.39	0.940	16
o, italic	33	5	44	3.21	0.976	16
u, italic (low confidence)	12	4	96	14.41	0.953	12
a, italic	31	5	60	9.78	0.954	16
c, italic (low confidence)	14	5	61	6.54	0.955	14
s, italic	31	5	28	6.24	0.960	16
e, italic	55	5	45	6.55	0.952	16
f, italic (low confidence)	12	4	87	14.28	0.909	12
d, italic (low confidence)	18	5	95	12.68	0.933	16
r, italic	24	5	49	17.64	0.946	16
n, italic	27	5	92	14.98	0.942	16

Stacking, glyph by glyph

For each glyph, the trace from the first N instances of one fixed random order, and the N-sweep over random subsets.

d, 20 instances

Trace from the first N instances of d
N	Cards	Nodes	Edge RMS µm
1	1	144	11.61
2	2	142	10.68
3	3	120	8.73
4	4	126	7.88
5	4	114	7.43
6	4	98	7.03
8	4	104	6.68
10	5	90	6.16
12	6	87	6.22
16	7	72	6.19
20	7	73	6.14

N-sweep for d over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	10.40	8.59	12.76	8.67
2	8.57	7.43	9.92	6.13
3	7.59	6.98	8.69	5.01
4	7.25	6.48	7.89	4.34
5	7.07	6.46	7.76	3.88
6	6.85	6.40	7.35	3.54
8	6.50	6.20	6.84	3.07
10	6.39	6.11	6.65	2.74
12	6.35	6.12	6.62	2.50
14	6.28	6.06	6.50	2.32
16	6.23	6.09	6.37	2.17
20	6.14	6.14	6.14	1.94

l, 34 instances

Trace from the first N instances of l
N	Cards	Nodes	Edge RMS µm
1	1	86	7.19
2	2	56	5.71
3	3	40	5.74
4	4	37	6.08
5	4	34	6.10
6	4	28	6.27
8	5	28	6.10
10	6	29	5.66
12	7	27	5.75
16	7	20	5.72
20	7	22	5.67
24	7	18	5.72
32	7	19	5.59
34	7	18	5.62

N-sweep for l over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	8.62	7.42	10.37	6.56
2	7.30	6.08	8.52	4.64
3	6.45	5.35	7.23	3.79
4	6.32	5.49	7.13	3.28
5	6.27	5.53	6.86	2.94
6	6.21	5.62	6.84	2.68
8	6.04	5.55	6.62	2.32
10	5.84	5.28	6.46	2.08
12	5.86	5.44	6.39	1.89
14	5.79	5.43	6.13	1.75
16	5.74	5.41	6.09	1.64
20	5.66	5.36	5.86	1.47
24	5.63	5.46	5.80	1.34
28	5.64	5.51	5.78	1.24
34	5.62	5.62	5.62	1.13

t, 74 instances

Trace from the first N instances of t
N	Cards	Nodes	Edge RMS µm
1	1	68	12.79
2	2	58	11.01
3	2	49	9.49
4	3	45	7.79
5	4	38	7.13
6	4	40	7.11
8	4	40	7.62
10	4	30	7.89
12	5	38	7.95
16	5	29	7.47
20	6	25	7.50
24	6	30	7.19
32	6	28	7.02
48	7	24	7.07
64	7	23	6.98
74	7	25	6.89

N-sweep for t over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	9.71	7.82	12.31	6.91
2	8.62	7.21	10.33	4.89
3	7.70	6.58	8.68	3.99
4	7.69	6.83	8.88	3.46
5	7.56	6.78	8.21	3.09
6	7.57	6.74	8.20	2.82
8	7.22	6.64	8.01	2.44
10	7.25	6.50	7.99	2.19
12	7.19	6.61	7.62	2.00
14	7.23	6.76	7.68	1.85
16	7.12	6.62	7.75	1.73
20	7.08	6.73	7.54	1.55
24	6.89	6.47	7.35	1.41
28	6.99	6.74	7.36	1.31
36	6.96	6.58	7.21	1.15
48	6.90	6.74	7.04	1.00
64	6.93	6.85	7.01	0.86
74	6.89	6.89	6.89	0.80

a, 66 instances

Trace from the first N instances of a
N	Cards	Nodes	Edge RMS µm
1	1	143	9.59
2	2	104	9.35
3	3	92	7.47
4	4	84	7.56
5	5	81	6.42
6	5	77	6.80
8	5	67	7.18
10	5	62	7.15
12	5	67	6.89
16	6	59	6.65
20	6	55	6.62
24	6	51	6.64
32	6	49	6.54
48	6	48	6.55
64	6	40	6.26
66	6	45	6.24

N-sweep for a over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	10.17	8.04	12.97	7.69
2	7.95	6.73	9.36	5.44
3	7.77	6.37	9.01	4.44
4	7.17	6.30	8.72	3.84
5	6.89	6.14	7.90	3.44
6	6.86	5.83	7.77	3.14
8	6.69	5.88	7.48	2.72
10	6.71	5.99	7.45	2.43
12	6.47	5.78	7.20	2.22
14	6.49	5.99	6.95	2.06
16	6.45	5.91	7.00	1.92
20	6.43	5.98	6.93	1.72
24	6.43	6.02	6.79	1.57
28	6.32	6.03	6.66	1.45
36	6.29	5.99	6.59	1.28
48	6.28	6.08	6.47	1.11
64	6.26	6.21	6.31	0.96
66	6.24	6.24	6.24	0.95

c, 29 instances

Trace from the first N instances of c
N	Cards	Nodes	Edge RMS µm
1	1	111	9.87
2	2	89	4.96
3	2	79	4.82
4	2	83	4.85
5	3	73	5.25
6	3	72	5.20
8	3	68	4.74
10	3	65	4.53
12	5	48	4.67
16	5	51	4.70
20	5	48	4.41
24	5	43	4.50
29	5	42	4.86

N-sweep for c over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	9.02	7.72	10.09	7.62
2	6.99	5.98	8.61	5.39
3	6.57	5.56	8.02	4.40
4	6.01	5.16	7.14	3.81
5	5.67	4.91	6.43	3.41
6	5.45	4.90	5.95	3.11
8	5.32	4.67	6.08	2.70
10	5.30	4.71	5.91	2.41
12	5.26	4.85	5.70	2.20
14	5.10	4.68	5.62	2.04
16	4.99	4.65	5.23	1.91
20	4.92	4.66	5.22	1.70
24	4.91	4.69	5.12	1.56
28	4.84	4.77	4.94	1.44
29	4.85	4.85	4.85	1.42

s, 33 instances

Trace from the first N instances of s
N	Cards	Nodes	Edge RMS µm
1	1	124	10.67
2	2	94	7.99
3	3	83	7.08
4	4	80	6.27
5	5	76	5.78
6	5	74	5.55
8	6	67	5.71
10	6	64	5.83
12	7	67	6.04
16	7	50	5.87
20	7	51	5.61
24	7	49	5.35
32	7	36	5.37
33	7	40	5.46

N-sweep for s over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	9.87	7.55	12.51	7.86
2	7.59	6.09	9.60	5.56
3	6.67	5.14	8.08	4.54
4	6.46	5.15	7.90	3.93
5	6.53	5.49	7.59	3.52
6	6.15	5.40	6.97	3.21
8	5.88	5.00	6.93	2.78
10	5.84	5.35	6.42	2.49
12	5.70	5.20	6.27	2.27
14	5.66	5.19	6.13	2.10
16	5.73	5.25	6.12	1.97
20	5.56	5.23	5.87	1.76
24	5.57	5.33	5.82	1.61
28	5.51	5.30	5.70	1.49
33	5.46	5.46	5.46	1.37

r, 64 instances

Trace from the first N instances of r
N	Cards	Nodes	Edge RMS µm
1	1	75	12.63
2	2	67	10.94
3	2	49	10.25
4	3	52	11.07
5	4	48	10.22
6	4	53	10.23
8	4	45	10.27
10	5	47	9.99
12	5	41	9.72
16	6	38	9.13
20	6	33	9.27
24	6	34	9.25
32	6	31	9.01
48	6	32	8.84
64	7	29	8.73

N-sweep for r over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	11.55	9.24	13.46	7.27
2	9.86	8.21	11.54	5.14
3	9.71	8.33	11.10	4.20
4	9.35	8.37	10.51	3.64
5	9.06	8.17	10.34	3.25
6	9.29	8.27	10.68	2.97
8	9.13	8.21	9.95	2.57
10	9.04	8.36	10.09	2.30
12	8.96	8.26	9.55	2.10
14	8.93	8.31	9.64	1.94
16	8.91	8.26	9.56	1.82
20	8.87	8.59	9.20	1.63
24	8.84	8.36	9.24	1.48
28	8.80	8.38	9.19	1.37
36	8.78	8.49	9.07	1.21
48	8.75	8.56	8.93	1.05
64	8.73	8.73	8.73	0.91

n, 57 instances

Trace from the first N instances of n
N	Cards	Nodes	Edge RMS µm
1	1	117	9.69
2	2	100	8.63
3	3	81	7.41
4	3	71	6.91
5	4	79	7.24
6	4	74	6.87
8	5	65	6.97
10	5	53	7.00
12	5	54	6.84
16	5	56	7.19
20	6	48	7.23
24	7	46	7.12
32	7	47	7.10
48	7	39	6.91
57	7	39	6.84

N-sweep for n over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	11.08	8.96	13.85	8.93
2	9.43	7.67	11.89	6.31
3	8.33	7.40	9.51	5.15
4	8.09	7.18	9.28	4.46
5	7.88	6.88	8.87	3.99
6	7.62	6.92	8.57	3.64
8	7.54	6.78	8.38	3.16
10	7.26	6.82	7.84	2.82
12	7.18	6.68	7.77	2.58
14	7.18	6.71	7.60	2.39
16	7.05	6.68	7.41	2.23
20	7.02	6.65	7.38	2.00
24	7.00	6.73	7.24	1.82
28	6.93	6.69	7.19	1.69
36	6.81	6.64	7.01	1.49
48	6.86	6.74	6.97	1.29
57	6.84	6.84	6.84	1.18

e, 107 instances

Trace from the first N instances of e
N	Cards	Nodes	Edge RMS µm
1	1	120	7.30
2	2	103	6.37
3	3	80	6.35
4	4	81	6.02
5	4	70	6.39
6	4	69	6.35
8	4	73	6.24
10	5	67	6.08
12	5	71	5.81
16	5	55	5.83
20	6	58	5.74
24	6	44	5.60
32	6	40	5.43
48	7	37	5.31
64	7	41	5.14
96	7	42	5.20
107	7	35	5.20

N-sweep for e over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	8.95	7.05	11.10	7.42
2	7.05	5.84	8.29	5.25
3	7.15	5.72	8.51	4.29
4	6.47	5.79	7.28	3.71
5	6.23	5.45	7.34	3.32
6	6.01	5.29	6.92	3.03
8	5.68	5.08	6.19	2.62
10	5.71	5.21	6.16	2.35
12	5.54	5.08	6.02	2.14
14	5.51	5.05	5.91	1.98
16	5.44	5.08	5.81	1.86
20	5.43	5.12	5.77	1.66
24	5.31	5.07	5.56	1.52
28	5.30	5.11	5.51	1.40
36	5.29	5.05	5.56	1.24
48	5.26	5.08	5.43	1.07
64	5.20	5.08	5.31	0.93
84	5.20	5.12	5.29	0.81
107	5.20	5.20	5.20	0.72

p, 13 instances

Trace from the first N instances of p
N	Cards	Nodes	Edge RMS µm
1	1	171	9.44
2	2	137	8.87
3	3	135	7.76
4	3	117	6.73
5	3	111	6.96
6	3	96	6.94
8	3	80	7.79
10	4	85	7.08
12	4	76	7.38
13	4	79	7.10

N-sweep for p over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	10.63	8.85	14.00	8.11
2	8.88	7.62	10.59	5.73
3	8.02	6.93	9.25	4.68
4	7.95	7.08	8.75	4.06
5	7.69	6.99	8.86	3.63
6	7.55	6.91	8.30	3.31
8	7.41	6.77	8.37	2.87
10	7.27	6.75	7.81	2.56
12	7.14	6.90	7.39	2.34
13	7.10	7.10	7.10	2.25

o, 46 instances

Trace from the first N instances of o
N	Cards	Nodes	Edge RMS µm
1	1	126	12.15
2	1	110	8.99
3	2	109	5.74
4	3	94	4.28
5	3	87	4.27
6	3	79	3.31
8	3	76	3.22
10	4	72	2.72
12	4	68	2.56
16	5	57	2.62
20	5	60	2.56
24	7	65	2.45
32	7	52	2.32
46	7	53	2.21

N-sweep for o over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	8.13	6.37	10.91	7.91
2	5.91	4.64	7.17	5.59
3	4.96	3.90	6.21	4.57
4	4.37	3.60	5.04	3.95
5	3.94	3.18	4.53	3.54
6	3.69	3.08	4.44	3.23
8	3.50	2.81	4.20	2.80
10	3.23	2.70	3.87	2.50
12	2.93	2.47	3.39	2.28
14	2.83	2.36	3.24	2.11
16	2.69	2.31	3.09	1.98
20	2.65	2.26	2.99	1.77
24	2.48	2.21	2.81	1.61
28	2.44	2.26	2.63	1.49
36	2.29	2.14	2.45	1.32
46	2.21	2.21	2.21	1.17

i, 33 instances

Trace from the first N instances of i
N	Cards	Nodes	Edge RMS µm
1	1	79	11.22
2	2	59	8.61
3	3	46	8.52
4	4	42	8.00
5	4	36	8.06
6	4	34	8.35
8	4	25	8.43
10	5	29	7.93
12	7	34	8.46
16	7	31	8.13
20	7	27	7.98
24	7	28	7.97
32	7	26	8.09
33	7	27	8.13

N-sweep for i over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	11.30	9.60	13.62	8.00
2	9.91	8.67	11.39	5.66
3	9.16	8.12	10.17	4.62
4	8.89	8.06	9.74	4.00
5	8.75	8.05	9.75	3.58
6	8.34	7.69	8.97	3.27
8	8.53	7.85	9.21	2.83
10	8.35	7.83	8.92	2.53
12	8.30	7.95	8.64	2.31
14	8.29	7.94	8.64	2.14
16	8.25	7.87	8.66	2.00
20	8.16	7.93	8.45	1.79
24	8.16	7.97	8.35	1.63
28	8.12	7.98	8.26	1.51
33	8.13	8.13	8.13	1.39

u, 31 instances

Trace from the first N instances of u
N	Cards	Nodes	Edge RMS µm
1	1	112	13.96
2	2	101	8.81
3	3	76	8.17
4	4	65	8.09
5	4	65	7.54
6	4	56	7.09
8	5	53	6.46
10	5	49	6.67
12	6	40	6.78
16	6	45	6.79
20	6	49	6.81
24	7	44	6.49
31	7	40	6.69

N-sweep for u over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	10.54	7.87	13.89	8.39
2	9.03	7.60	10.98	5.93
3	7.99	6.98	9.40	4.84
4	7.76	6.60	8.95	4.19
5	7.50	6.61	8.40	3.75
6	7.25	6.63	7.78	3.42
8	7.17	6.45	7.77	2.96
10	7.05	6.60	7.58	2.65
12	6.96	6.59	7.42	2.42
14	6.85	6.50	7.30	2.24
16	6.81	6.46	7.13	2.10
20	6.79	6.51	7.04	1.88
24	6.73	6.60	6.90	1.71
28	6.69	6.57	6.82	1.58
31	6.69	6.69	6.69	1.51

y, 12 instances

Trace from the first N instances of y
N	Cards	Nodes	Edge RMS µm
1	1	150	9.34
2	2	133	7.92
3	2	112	7.63
4	2	98	8.03
5	3	99	8.02
6	3	94	8.02
8	3	83	8.53
10	3	87	8.50
12	3	93	8.46

N-sweep for y over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	11.30	9.68	13.44	7.92
2	10.13	8.76	11.67	5.60
3	9.11	8.12	10.21	4.57
4	8.87	8.10	9.49	3.96
5	8.97	8.25	9.69	3.54
6	8.70	8.15	9.20	3.23
8	8.58	7.97	9.02	2.80
10	8.53	8.25	8.88	2.51
12	8.46	8.46	8.46	2.29

h, 14 instances

Trace from the first N instances of h
N	Cards	Nodes	Edge RMS µm
1	1	138	9.07
2	2	117	7.98
3	2	90	7.69
4	2	85	7.78
5	3	77	7.69
6	3	68	7.58
8	4	67	7.52
10	4	69	7.35
12	5	57	6.85
14	5	55	6.84

N-sweep for h over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	11.24	9.07	13.48	9.29
2	9.36	7.73	10.67	6.57
3	8.12	7.37	8.88	5.36
4	7.81	7.15	8.68	4.64
5	7.45	6.89	8.08	4.15
6	7.42	6.93	8.08	3.79
8	7.18	6.83	7.55	3.28
10	6.97	6.73	7.16	2.94
12	6.91	6.81	7.02	2.68
14	6.85	6.85	6.85	2.48

full stop, 15 instances

Trace from the first N instances of full stop
N	Cards	Nodes	Edge RMS µm
1	1	21	6.17
2	2	19	7.39
3	2	13	7.00
4	2	11	7.92
5	3	9	8.06
6	4	9	7.71
8	5	6	8.32
10	6	8	8.19
12	7	6	8.38
15	7	6	8.06

N-sweep for full stop over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	9.31	6.17	11.18	4.95
2	8.86	7.27	10.46	3.50
3	8.44	6.98	9.79	2.86
4	8.36	7.26	9.18	2.47
5	8.07	7.40	8.69	2.21
6	8.28	7.57	9.10	2.02
8	8.10	7.56	8.62	1.75
10	8.20	7.66	8.75	1.56
12	8.08	7.73	8.39	1.43
14	8.06	7.91	8.25	1.32
15	8.06	8.06	8.06	1.28

h, italic, 19 instances

Trace from the first N instances of h, italic
N	Cards	Nodes	Edge RMS µm
1	1	176	15.58
2	2	155	15.24
3	2	146	14.42
4	3	138	14.55
5	3	130	14.29
6	3	135	14.31
8	5	120	14.19
10	5	115	14.35
12	5	105	14.43
16	5	117	14.41
19	5	112	14.49

N-sweep for h, italic over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	16.35	14.90	18.02	7.62
2	15.21	14.14	16.46	5.39
3	15.05	13.90	15.90	4.40
4	14.80	13.88	15.66	3.81
5	14.70	13.98	15.48	3.41
6	14.67	13.94	15.55	3.11
8	14.49	13.97	15.00	2.70
10	14.58	14.15	15.12	2.41
12	14.66	14.37	15.05	2.20
14	14.53	14.24	14.89	2.04
16	14.53	14.34	14.75	1.91
19	14.49	14.49	14.49	1.75

l, italic, 17 instances

Trace from the first N instances of l, italic
N	Cards	Nodes	Edge RMS µm
1	1	109	13.79
2	1	86	13.97
3	2	80	11.45
4	3	77	9.47
5	3	74	8.18
6	3	78	8.04
8	4	75	8.82
10	4	67	9.42
12	5	72	10.15
16	5	66	9.91
17	5	68	9.86

N-sweep for l, italic over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	12.56	9.57	16.86	8.18
2	11.97	8.61	15.57	5.78
3	10.24	8.27	12.85	4.72
4	9.80	8.02	11.42	4.09
5	9.88	8.41	11.28	3.66
6	10.13	8.63	11.57	3.34
8	10.12	8.77	11.35	2.89
10	10.08	9.34	11.12	2.59
12	9.67	8.97	10.21	2.36
14	9.78	9.31	10.30	2.19
16	9.89	9.65	10.22	2.04
17	9.86	9.86	9.86	1.98

t, italic, 41 instances

Trace from the first N instances of t, italic
N	Cards	Nodes	Edge RMS µm
1	1	79	10.81
2	2	68	10.21
3	3	69	9.06
4	4	69	10.02
5	4	54	9.94
6	4	49	9.56
8	4	52	9.13
10	5	49	9.11
12	5	49	9.13
16	5	47	9.29
20	5	51	9.78
24	5	45	9.51
32	5	47	9.20
41	5	41	9.39

N-sweep for t, italic over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	11.79	9.77	15.27	7.32
2	11.08	9.18	13.31	5.18
3	9.98	8.85	11.79	4.23
4	10.22	8.97	11.83	3.66
5	9.72	8.80	10.73	3.27
6	9.80	8.66	10.95	2.99
8	9.63	8.64	10.66	2.59
10	9.66	8.84	10.51	2.32
12	9.61	8.94	10.28	2.11
14	9.64	8.90	10.37	1.96
16	9.56	8.95	10.13	1.83
20	9.58	9.04	10.04	1.64
24	9.48	8.97	9.96	1.49
28	9.41	9.08	9.70	1.38
36	9.40	9.23	9.57	1.22
41	9.39	9.39	9.39	1.14

o, italic, 33 instances

Trace from the first N instances of o, italic
N	Cards	Nodes	Edge RMS µm
1	1	121	7.13
2	2	108	5.17
3	3	108	4.80
4	4	100	4.84
5	4	87	4.83
6	4	90	5.23
8	5	79	3.80
10	5	68	4.23
12	5	72	4.34
16	5	64	3.71
20	5	51	3.32
24	5	46	3.23
32	5	45	3.33
33	5	44	3.21

N-sweep for o, italic over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	8.77	7.13	10.70	8.21
2	6.34	5.11	8.01	5.81
3	5.63	4.58	7.50	4.74
4	4.96	4.18	5.84	4.11
5	4.74	3.84	5.85	3.67
6	4.44	3.90	5.44	3.35
8	4.13	3.30	4.77	2.90
10	3.76	3.23	4.43	2.60
12	3.80	3.27	4.48	2.37
14	3.80	3.28	4.68	2.19
16	3.44	3.10	3.77	2.05
20	3.46	3.09	3.79	1.84
24	3.31	3.07	3.64	1.68
28	3.28	3.11	3.47	1.55
33	3.21	3.21	3.21	1.43

u, italic, 12 instances

Trace from the first N instances of u, italic
N	Cards	Nodes	Edge RMS µm
1	1	170	17.71
2	2	138	15.48
3	3	114	15.43
4	3	115	14.24
5	3	119	14.12
6	3	113	14.07
8	4	111	14.35
10	4	98	14.53
12	4	96	14.41

N-sweep for u, italic over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	16.32	14.35	18.37	8.14
2	15.43	14.01	16.68	5.76
3	15.06	14.07	16.00	4.70
4	14.77	13.98	15.70	4.07
5	14.64	13.93	15.20	3.64
6	14.65	14.04	15.37	3.32
8	14.46	14.08	14.84	2.88
10	14.41	14.26	14.54	2.57
12	14.41	14.41	14.41	2.35

a, italic, 31 instances

Trace from the first N instances of a, italic
N	Cards	Nodes	Edge RMS µm
1	1	160	11.93
2	2	126	12.04
3	3	116	11.31
4	3	112	11.49
5	3	109	11.37
6	3	107	10.85
8	4	96	9.82
10	4	106	9.90
12	4	87	9.94
16	5	88	9.59
20	5	71	9.59
24	5	83	9.78
31	5	60	9.78

N-sweep for a, italic over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	12.35	9.94	14.05	7.66
2	11.12	9.93	12.43	5.42
3	10.38	9.44	11.31	4.42
4	10.67	9.61	12.16	3.83
5	10.34	9.62	11.07	3.43
6	10.22	9.50	11.06	3.13
8	10.06	9.31	10.70	2.71
10	10.04	9.48	10.61	2.42
12	9.95	9.45	10.72	2.21
14	9.90	9.58	10.31	2.05
16	9.81	9.44	10.18	1.92
20	9.77	9.48	10.13	1.71
24	9.81	9.62	10.03	1.56
28	9.75	9.59	9.91	1.45
31	9.78	9.78	9.78	1.38

c, italic, 14 instances

Trace from the first N instances of c, italic
N	Cards	Nodes	Edge RMS µm
1	1	108	9.87
2	2	88	6.67
3	2	77	6.80
4	3	73	7.01
5	3	63	7.08
6	3	52	6.16
8	3	49	6.26
10	3	53	6.14
12	4	46	6.54
14	5	61	6.54

N-sweep for c, italic over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	9.45	7.94	10.45	7.10
2	7.84	6.78	8.94	5.02
3	7.78	6.60	8.99	4.10
4	7.44	6.34	8.68	3.55
5	7.05	6.18	8.00	3.17
6	6.92	6.02	8.10	2.90
8	6.74	6.19	7.43	2.51
10	6.59	6.17	7.11	2.24
12	6.62	6.31	6.88	2.05
14	6.53	6.53	6.53	1.90

s, italic, 31 instances

Trace from the first N instances of s, italic
N	Cards	Nodes	Edge RMS µm
1	1	103	8.99
2	2	91	6.21
3	3	66	6.87
4	4	57	6.90
5	4	57	7.18
6	4	54	7.51
8	5	59	7.13
10	5	43	7.22
12	5	42	7.55
16	5	37	6.91
20	5	39	6.37
24	5	37	6.43
31	5	28	6.24

N-sweep for s, italic over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	9.68	7.33	11.83	7.60
2	8.18	7.10	9.72	5.37
3	7.62	6.31	8.74	4.39
4	7.05	6.29	8.16	3.80
5	6.84	6.07	7.62	3.40
6	6.92	6.14	7.88	3.10
8	6.68	6.24	7.50	2.69
10	6.49	5.84	7.12	2.40
12	6.43	5.89	7.02	2.19
14	6.39	5.86	6.81	2.03
16	6.41	5.93	6.76	1.90
20	6.33	5.91	6.77	1.70
24	6.27	6.05	6.49	1.55
28	6.23	6.10	6.34	1.44
31	6.24	6.24	6.24	1.36

e, italic, 55 instances

Trace from the first N instances of e, italic
N	Cards	Nodes	Edge RMS µm
1	1	119	8.22
2	2	105	7.36
3	3	92	7.15
4	3	96	7.40
5	3	83	7.93
6	4	73	7.39
8	4	74	6.89
10	4	74	6.57
12	5	76	6.51
16	5	56	6.68
20	5	57	6.57
24	5	56	6.60
32	5	57	6.46
48	5	48	6.49
55	5	45	6.55

N-sweep for e, italic over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	10.07	8.27	12.29	7.76
2	8.38	7.17	9.86	5.49
3	8.18	7.32	9.44	4.48
4	7.45	6.69	8.25	3.88
5	7.32	6.25	8.60	3.47
6	7.27	6.33	8.16	3.17
8	7.14	6.46	8.12	2.74
10	6.88	6.26	7.54	2.45
12	6.83	6.34	7.46	2.24
14	6.79	6.38	7.20	2.07
16	6.73	6.34	7.20	1.94
20	6.75	6.34	7.20	1.74
24	6.62	6.30	7.03	1.58
28	6.66	6.36	6.88	1.47
36	6.61	6.39	6.78	1.29
48	6.57	6.45	6.71	1.12
55	6.55	6.55	6.55	1.05

f, italic, 12 instances

Trace from the first N instances of f, italic
N	Cards	Nodes	Edge RMS µm
1	1	137	15.09
2	2	115	10.75
3	3	114	12.52
4	3	101	14.56
5	3	97	15.05
6	3	101	13.38
8	4	98	13.32
10	4	95	13.96
12	4	87	14.28

N-sweep for f, italic over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	17.46	12.87	23.29	10.28
2	15.16	11.79	17.80	7.27
3	15.07	11.45	17.97	5.93
4	14.31	11.47	17.09	5.14
5	14.42	12.62	16.88	4.60
6	14.23	12.42	16.36	4.20
8	14.11	12.96	15.61	3.63
10	14.38	13.65	15.31	3.25
12	14.28	14.28	14.28	2.97

d, italic, 18 instances

Trace from the first N instances of d, italic
N	Cards	Nodes	Edge RMS µm
1	1	151	16.41
2	1	148	15.60
3	2	140	14.21
4	3	144	12.40
5	4	129	12.38
6	4	119	11.19
8	5	117	12.21
10	5	108	12.45
12	5	101	12.21
16	5	101	12.50
18	5	95	12.68

N-sweep for d, italic over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	14.86	12.06	17.27	7.92
2	13.50	10.28	15.57	5.60
3	13.57	11.68	15.38	4.57
4	12.97	10.71	14.72	3.96
5	12.67	11.08	14.05	3.54
6	12.73	11.28	14.21	3.23
8	12.87	11.48	13.88	2.80
10	12.78	12.04	13.53	2.50
12	12.68	12.11	13.17	2.29
14	12.72	12.29	13.32	2.12
16	12.65	12.38	12.91	1.98
18	12.68	12.68	12.68	1.87

r, italic, 24 instances

Trace from the first N instances of r, italic
N	Cards	Nodes	Edge RMS µm
1	1	95	18.82
2	2	80	16.01
3	2	72	15.88
4	3	67	15.59
5	4	66	16.34
6	4	67	16.58
8	4	57	16.96
10	4	51	17.03
12	4	55	17.61
16	5	47	18.19
20	5	46	17.85
24	5	49	17.64

N-sweep for r, italic over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	18.51	16.08	20.39	5.91
2	17.99	15.79	20.51	4.18
3	18.10	16.76	19.36	3.41
4	17.96	16.73	19.08	2.96
5	17.89	16.82	19.01	2.64
6	17.74	16.64	18.61	2.41
8	17.66	16.78	18.42	2.09
10	17.61	16.90	18.46	1.87
12	17.70	17.04	18.34	1.71
14	17.51	16.90	17.97	1.58
16	17.64	17.30	18.08	1.48
20	17.65	17.33	17.87	1.32
24	17.64	17.64	17.64	1.21

n, italic, 27 instances

Trace from the first N instances of n, italic
N	Cards	Nodes	Edge RMS µm
1	1	157	18.89
2	2	129	17.38
3	2	116	16.38
4	2	116	16.96
5	2	110	16.29
6	3	107	16.14
8	5	99	15.81
10	5	101	15.49
12	5	92	15.45
16	5	94	15.54
20	5	84	15.52
24	5	84	14.82
27	5	92	14.98

N-sweep for n, italic over random subsets, µm
N	Edge RMS	p10	p90	Noise model
1	16.92	15.17	18.62	7.96
2	15.84	14.17	17.57	5.63
3	15.59	14.39	17.23	4.59
4	15.52	14.17	16.84	3.98
5	15.33	14.12	16.63	3.56
6	15.10	14.24	15.93	3.25
8	15.08	14.33	15.70	2.81
10	15.05	14.52	15.71	2.52
12	15.13	14.56	15.73	2.30
14	15.08	14.42	15.80	2.13
16	15.06	14.64	15.57	1.99
20	14.98	14.61	15.33	1.78
24	15.00	14.82	15.21	1.62
27	14.98	14.98	14.98	1.53

Observed spacing

Gaps between adjacent letters measured on the cards, in font units (1 fu is about 3.1 µm), against the designed gap with and without its kern value.

168 pairs, most measured first
Pair	Samples	Designed gap	Kern	Mean deviation before kern	After kern	sd
re	51	40	0	1.8	1.8	4.9
en	40	49	0	4.4	4.4	4.4
at	40	45	0	3.8	3.8	6.1
ea	32	77	0	1.5	1.5	4.6
ur	29	42	0	5.3	5.3	6.1
tu	28	32	0	1.5	1.5	4.4
cr	28	51	0	6.3	6.3	4.3
nt	24	38	0	6.4	6.4	5.3
er	24	54	0	2.4	2.4	7.1
le	18	71	0	1.4	1.4	6.4
ta	18	58	0	2.1	2.1	4.9
ar	18	47	0	3.0	3.0	5.7
it	17	43	0	2.3	2.3	7.7
te	16	56	0	0.7	0.7	4.9
on	16	60	0	1.7	1.7	4.7
an	16	42	0	2.2	2.2	4.4
to	16	56	0	0.9	0.9	4.8
ke	15	40	0	-1.6	-1.6	7.4
nd	14	58	0	-0.3	-0.3	9.0
et	14	52	0	5.9	5.9	6.6
il	14	51	0	2.5	2.5	8.1
ou	14	62	0	0.9	0.9	4.8
co	14	72	0	5.6	5.6	4.8
un	14	37	0	2.5	2.5	4.3
ne	12	61	0	2.1	2.1	4.7
ok	12	75	0	1.2	1.2	4.7
ll	12	56	0	8.5	8.5	4.9
or	10	65	0	-0.7	-0.7	3.4
wi	10	34	0	1.5	1.5	2.5
rd	10	37	0	0.1	0.1	6.4
th	10	23	0	4.6	4.6	4.0
es	10	73	0	-2.1	-2.1	2.6
ol	10	71	0	2.3	2.3	5.1
rn	8	14	30	33.4	3.4	6.9
ts	8	54	0	2.7	2.7	3.3
me	8	65	0	-0.7	-0.7	6.0
t.	8	97	0	3.7	3.7	6.8
of	8	66	0	-3.0	-3.0	5.3
lu	8	47	0	-0.4	-0.4	3.3
ha	7	61	0	2.5	2.5	3.6
ve	7	44	-20	-23.5	-3.5	4.7
in	7	40	0	2.5	2.5	5.2
yo	7	44	-15	-11.9	3.1	2.9
s,	7	74	0	5.8	5.8	5.9
la	6	73	0	4.4	4.4	3.3
tt	6	33	0	-0.2	-0.2	3.4
el	6	60	0	4.0	4.0	4.7
d.	6	106	0	2.6	2.6	5.4
ow	6	51	-15	-17.0	-2.0	3.4
wn	6	25	0	2.1	2.1	4.1
ac	6	69	0	-4.2	-4.2	5.8
ss	6	82	0	1.7	1.7	4.9
xi	6	35	0	2.8	2.8	5.3
ro	6	40	0	5.1	5.1	5.0
ca	6	74	0	5.2	5.2	3.4
ed	6	72	0	5.8	5.8	6.1
di	6	48	0	3.1	3.1	5.0
ie	6	66	0	0.8	0.8	7.5
bl	6	70	0	0.2	0.2	2.5
is	6	64	0	-3.9	-3.9	5.0
ue	6	63	0	2.7	2.7	3.4
n.	6	102	0	6.9	6.9	9.6
ti	6	39	0	-0.6	-0.6	3.2
rs	5	38	20	17.0	-3.0	4.2
tr	5	35	0	0.7	0.7	3.6
ge	5	47	0	3.1	3.1	3.5
om	5	65	0	1.3	1.3	3.2
ba	4	87	0	-1.4	-1.4	4.6
nl	4	46	0	3.2	3.2	2.8
rm	4	19	30	31.3	1.3	3.0
st	4	61	0	-2.9	-2.9	4.8
ev	4	34	0	3.6	3.6	4.4
si	4	67	0	3.9	3.9	4.7
ex	4	42	0	-1.0	-1.0	3.3
mo	4	65	0	-6.6	-6.6	8.6
em	4	54	0	2.5	2.5	3.3
av	4	27	0	4.2	4.2	4.6
ra	4	42	0	2.5	2.5	7.2
ch	4	39	0	4.4	4.4	4.3
pp	4	56	0	3.9	3.9	3.1
pe	4	88	0	0.3	0.3	3.5
Re	4	20	0	-3.4	-3.4	3.0
hi	4	42	0	-1.6	-1.6	2.9
al	4	53	0	5.3	5.3	3.7
’’	4	128	-60	5.3	65.3	64.3
lf	4	51	0	1.0	1.0	3.0
+4	4	104	0	13.2	13.2	4.8
Fi	3	47	0	-1.1	-1.1	2.8
ck	3	61	0	6.4	6.4	3.6
he	3	59	0	-4.1	-4.1	0.5
4/	3	43	0	6.5	6.5	0.6
sn	2	58	0	-6.2	-6.2	3.4
ni	2	44	0	-12.3	-12.3	3.4
ld	2	68	0	0.5	0.5	0.0
d,	2	55	0	0.1	0.1	0.0
tl	2	41	0	-2.6	-2.6	3.4
no	2	61	0	0.8	0.8	0.0
ma	2	67	0	-1.8	-1.8	3.4
Ev	2	21	0	0.6	0.6	0.0
vo	2	44	-20	-25.7	-5.7	3.4
ks	2	38	0	-2.8	-2.8	3.9
bn	2	59	0	-0.6	-0.6	3.9
Ib	2	34	0	1.2	1.2	3.9
lt	2	48	0	2.7	2.7	3.9
Al	2	31	0	-3.5	-3.5	3.9
ad	2	65	0	-6.6	-6.6	3.9
As	2	44	0	-1.0	-1.0	3.9
as	2	66	0	-3.7	-3.7	0.0
sa	2	86	0	3.3	3.3	3.9
ya	2	46	-10	-3.0	7.0	3.9
gr	2	26	20	20.8	0.8	0.0
Th	2	23	0	4.5	4.5	3.9
t,	2	46	0	0.8	0.8	0.0
Ex	2	29	0	-1.5	-1.5	3.9
ns	2	59	0	-4.5	-4.5	0.0
se	2	84	0	-6.3	-6.3	0.0
os	2	84	0	1.4	1.4	0.0
py	2	47	0	3.7	3.7	3.9
3/	2	64	0	11.9	11.9	0.0
ee	2	75	0	4.2	4.2	3.4
ef	2	55	0	0.6	0.6	0.0
‘‘	2	128	-60	-65.7	-5.7	0.0
sh	2	51	0	7.9	7.9	3.4
6/	2	53	0	2.6	2.6	0.0
9/	2	64	0	15.2	15.2	3.4
ak	2	57	0	12.1	12.1	0.0
/3	2	43	0	15.9	15.9	3.4
/6	2	64	0	8.5	8.5	3.4
/9	2	53	0	5.9	5.9	3.4
Wh	2	11	0	4.0	4.0	0.0
Qu	2	32	0	3.5	3.5	0.0
ui	2	46	0	20.1	20.1	3.4
/+	2	66	0	6.9	6.9	3.4
l,	2	61	0	10.5	10.5	0.0
Il	2	55	0	-1.0	-1.0	3.5
Gl	2	35	0	1.4	1.4	0.0
be	2	85	0	-6.4	-6.4	0.0
lc	2	72	0	-4.0	-4.0	3.5
Ha	2	76	0	2.6	2.6	0.0
io	2	66	0	-1.5	-1.5	0.0
ze	2	60	0	1.0	1.0	3.5
ec	2	76	0	-4.5	-4.5	0.0
us	2	61	0	-7.0	-7.0	3.5
az	2	39	0	4.4	4.4	0.0
It	2	47	0	-0.1	-0.1	3.5
tm	2	35	0	-9.2	-9.2	3.5
nc	2	62	0	2.5	2.5	0.0
Cl	2	88	0	-6.2	-6.2	0.0
ew	2	40	0	8.3	8.3	0.0
wo	2	51	-20	-19.4	0.6	3.3
Dr	2	69	0	2.7	2.7	3.3
aw	2	33	0	8.6	8.6	0.0
Wi	2	27	0	4.0	4.0	0.0
l.	2	112	0	14.0	14.0	3.8
iz	2	37	0	16.8	16.8	0.0
za	2	62	0	7.0	7.0	0.0
Co	2	103	0	0.2	0.2	3.8
pa	2	90	0	-2.0	-2.0	3.8
ys	2	42	0	4.2	4.2	0.0
ip	1	34	0	1.0	1.0	0.0
ri	1	23	20	23.8	3.8	0.0
ik	1	55	0	-0.5	-0.5	0.0
ir	1	45	0	-5.9	-5.9	0.0
La	1	23	0	0.7	0.7	0.0
ki	1	23	0	0.7	0.7	0.0
ho	1	59	0	-4.5	-4.5	0.0
mb	1	29	0	-5.3	-5.3	0.0
/4	1	38	0	12.4	12.4	0.0

Limits

One scanner. Seven pilot cards. Optical blur of 100 to 125 µm. The ink spread is relative to the pooled edge level, not an absolute physical measurement. Stack-against-design numbers depend on one fitted outline.