By integrated systems, I mean systems that offer the whole picture-telegraphic package with minimal human intervention : scanning (also called exploration), coding (and possibly counting), production of an intermediate record (e.g., punched tape), transmission, and reconstitution at the receiving end. The systems described above either required the human eye for analysis or were silent on the analysis itself, focussing instead on the elementary signs for reconstitution.

Some of these systems did not involve an intermediate record, but rather synchronous transmission/receiving. Korn and Thorne Baker, for example, preferred these so-called "direct" methods, akin to analog telephony in the kinds of circuits they required.

Moreover, there appears to have been a general movement toward direct methods in which alternating or pulsating signals were transmitted over phone lines. One objective was simplication of the facsimile process, so that "comparatively untrained personnel" could use it. Intermediate perforated tape records did not fit into this objective. And so we have McFarlane and Hansen 1939 US 2,164,038, describing a system that was essentially already in use at that time. In some sense, they were leaving something like PCM coding of pictures -- that is, quanitizing samples -- behind, at about the same time Reeves conceptualized Pulse Code Modulation for telephony.

Direct methods are mentioned mostly in passing below. I focus rather on "indirect" methods involving analysis of pixel-like zones into classes of grayscale values, transmission of that coded information, and reconstitution into "elementary signs" whether they be printed or photographic dots of different sizes, or some other form. The expression "indirect" might also be characterized as a condition of "off line," in which analysis, coding and preparation of perforated tape intermediate records happened separately from transmission.

We begin a survey of "automated" systems with this first Ives patent for picture analysis. It is transitional, because an operator must manually key an encoding device -- the tape perforator -- as he or she monitors successive readings from a galvanometer. This vestige of earlier ideas would be gone in Horton Ives and Long 1926/1,606,227, which claimed to be a "working whole suitable for commercial transmission of pictures."

In image below right can be seen an observation microscope "used for inspecting the light valve aperture for adjusting purposes." This device was probably a part of the working prototype machine, and provides yet another vestige of human analysis of the image -- here only to calibrate the galvanometer dial or the relays.
 

Korn and the "indirect method"

Thorne Baker (1927) writes :

"A further method of employing a code was adopted recently by Korn, who devised an instrument for rapidly assigning to the unit tones of a photographic image the appropriate code letters, and a second instrument rather on the lines of a typewriter for reconstructing the image from the code telegram. The latter could, of course, be sent equally well by wire or wireless, and an actual example of such a code transmission from the United States to Italy is seen in Fig. 11. The code method has appealed to the imagination of many inventors, but it seems doubtful whether in the long run it can vie with more direct processes." (20)

The same image is presented in Korn's own Bildtelegraphie (1927), in chapter 11 devoted to such "indirect methods". Chapter 11 of that book runs as follows (my rough Englishing, with thanks to Dorothee Weinlich in Berlin) :

I would like further to briefly mention methods by which pictures may be transmitted by cipher telegrams and punched tape. Individual picture elements are photoelectrically scanned and translated into ciphers. For example, a light picture element is assigned the letter "A", a dark element the letter "Z", and the remaining graduated tones are assigned their respective letters, which are then transmitted telegraphically.

One can also use photoelectric cells to generate differerent current strengths for the different brightnesses of each picture element; these in turn can actuate the different letters of a printer for the automatic dispatch of a telegram. The telegram is sent in the usual way, by cable or wireless, and the picture is reconstituted at the receiving end. The same can happen with the use of punched tape; the recomposition of the image is done with a simple typewriter, where the typebars bear not individual letters but small squares or rectangles whose dimensions range from large to small, each representing the tone assigned to its letter. For a "Z", a small rectangle is printed that fills out a grid space. A "Y" will be a somewhat smaller rectangle, and so on down down to the smallest point. The "A" bar will leave an empty space. Thus does one print one picture element after another, following the received telegram, to build up the same picture that was originally scanned and transformed into a cipher telegram.

In this way, I effected a wireless transmission of a photograph over the ocean, from Rome via Nauen and Bar Harbor to New York, in 1922.

I prepared a similar method also for black and white, obtaining the advantage that a single letter can represent five successive picture elements, and so reduce telegram wordcount down to one fifth.

These indirect methods offer obvious advantages over direct methods: (a) scanning and reproduction at the receiver end are done off-line, and do not tie up long-distance circuits or wireless stations, and (b) they require only the usual telegraph working. There are no requirements for synchrony between sending and receiving station, and they take advantage advantage of the broad reach of ordinary telegraph service. One would ordinarily prefer direct picture transmission where possible, because it is more economical and can be done in shorter time. But where direct phototelegraphy is uncertain or impossible, e.g., because of distance, then the indirect method should be used.

Bartlane patents GB

1921 GB 166,917
Improvements in and relating to the production of facsimiles of documents, pictures, drawings and the like.
"Presuming the code comprises letters and numbers, and reads, for instance, A 4 B 2 D 1 E 6 etc., the successive areas will be exposed as follows, 4 areas tone A, 2 areas tone B, 1 area tone D, 6 areas tone E, etc." (2/64ff). The tape operates a shutter device that can either provide multiple exposures of equal lengths, or explosures of varying duration. Drawings relate primarily to the shutter mechanism.

1923 GB 195,577
Improvements in and relating to telautography or phototelegraphy "In picture transmission to obtain gradation of tone in the reproduction, a plurality of representations from the original, each corresponding with a given fraction or range of tone, are transmitted, and the corresponding received representations are used to form a composite reproduction of the original. An intermediate record of the tone variations, such as a perforated tape, may be used, and such record may be effected by the method and means described in Specification 197,402."
substantially same as 1927 US 1,653,425

1923 GB 197,402
Improvements in and relating to telautography

1924 GB 209,816
Improvements in and relating to Telautography or Phototelegraphy
(optical system, using elaborate lenses and five-unit tape as a kind of light valve; same as US 1924 1,674,881)

1924 GB 226,244
Improvements in and relating to Telautography of [sic] Phototelegraphy
(multiplexes telegraph lines, or uses telephone transmission)

1926 GB 262,943
Improvements in and relating to telautography or photo-telegraphy
readers scan slightly off in successive scans, so as not to miss thin features
similar to 1930 US 1,758,388

Bartlane patents US

1923 US 1,454,719
Facsimile transmission system
substantially same as 1922 GB 182,928
A representation of the matter to be transmitted, such as a half-tone or other etched plate or an impression in wax such as is used in phonograph records, is used to control production of sound signals corresponding with the elements to be sent, and the sound signals are transmitted by line or wireless telephony. At the receiving end, the signals are used to control the reproduction. (from UK abstract)
The sound produced by the relative motion between the surface on which the representation is formed and the jewel or needle will vary in character according to the size of the dots when the representation is in the form of an etched metal plate..." (1:47ff)
no coding

1925 US 1,545,301
Telegraphy combined with phototelegraphy or telautography
devoted to allocation of image and message signals in various channels of multiplex systems, for efficient use of circuits

1927 US 1,653,425
Telautography
substantially same as 1923 GB 195,577
Here is a clear statement of the initial Bartlane system, in which four or five different representations of a single image are made, each comprising a different tonality (deepest tone, next lighter tones, still lighter tones) which are coded and transmitted separately, for "composite reproduction" at the receiving station. Four "examples" are given :

1. (1) four representations are formed on zinc plate (by coating with an insulating material), wound round spindle, explored. signals stored on perforated tape.
2. Record stored on tape. Perforations themselves control exposure.
3. Scan of a half tone or line screen reproduction, code record stored on perforated tape.
4. Similar to example three, except that it incorporates a Western Electric 6A machine perforator.

1928 US 1,674,881
Telautography or phototelegraphy
The perforated tape itself provides light valves (in concert with lenses and apertures, "such apertures being wider at one end 10a than at the other 10, the narrower end being adjacent that over which the hole representing the lightest tine will pass." (1:81ff)
substantially same as 1924 GB 209,816

1929 US 1,731,518
Phototelegraphy

1930 US 1,758,388
involves five successive scans of same image, with a "progressive 'lateral' variation of the reader points" in order to capture subtle details. perforated tape is light valve.
substantially same as 1926 GB 262,943

1930 US US 1,754,688
Machine for reproducing and transmitting pictures
 

The following patents are grouped together, out of chronology, because each of them directly relate to coding and efficient use of tape -- that is to say, data compression.
 

1929 US 1,731,345
Method of and tape for reproducing pictures or the like

The first -- 1929 US 1,731,345 -- sets the stage. A first intermediate record is normal five-unit punched tape. As explained in the note to Fig. 10, it can describe as many as 32 tone values, but 10 are shown here. This perforated tape is fed through an apparatus shown in section in Fig. 4, which contains condenser lenses and the apertures (5), which can be adjusted to yield the "area relationships" shown in Figs. 6 and 7. It is the combination of these five apertures, with their varying widths, that yield up to 32 gradations of tone.

Note that each row of code in the perforated tape contains information for one "small area" at a time. One hundred such "pixels" in a row would, therefore, require 100 rows of 50-unit code. Subsequent Bartlane patents provide for "counting" of adjacent small areas with the same tone value, as a savings in tape running time (and hence, cost).

1933 US 1,910,556
Mechanism for producing a perforated tape for reproducing pictures of the like
This patent and 1,910,586 focus on coding of the grayscale images in their successive "zones." Neither is concerned with the actual method of scan, whether relief or photocell, but with the translation of that data into code. The present patent does provide for the use of the punched tape for exposure of light, through 1, 2, 3, 4 or 5 of its apertures, onto photocemical paper.
no counter

1933 US 1,910,586
Telautography or Phototelegraphy
Incorporates counter.
"The object of the present invention is in general to enable the time occupied in transmitting a record of the picture to be reduced... the results of analysis of the original to be transmitted are divided into two portions, one indicative of the tones and the other expressing the number of times each tone occurs in succession." (1:47ff)
Several means of coding this information, on either one or two intermediate records -- if two, the primary intermediate record for "tone perforations" and the secondary intermediate record for "count perforations." (14:5ff) Provides a thorough exploration of getting the most out of five-unit tape, particularly with regards to the tradeoff of tones versus count. See for example 4:33ff (and Fig. 4) where six tone combinations are indicated by three positions, leaving two positions available for count (maximum 4); versus Fig. 6 four tone combinations (two holes) leaving room for eight counts (in three holes).
"Fig. 7 is an elevation of a Western Electric 5B start-stop distributor modified to act as a counter and sequence switch for an apparatus to automatically punch a secondary intermediate record from a primary intermediate record." (3:4ff) compare to Watson and Weaver, "Electrical Transmission of Pictures," 1926 US 1,602,469

1935 US 2,021,474
Phototelegraphy
Presents punched tape intermediate record as a means of controlling against unintentional fluctuations or fading attendant encountered in "direct" photo-telegraphy and television (calls to mind the same objective in Reeves's 1942 patent for pulse code modulation in telephony).
 

By ingenious means described in
1923 GB 197,402
1923 GB 195,577 and
1930 US 1,758,388
each segment of five units on the black (not white!) perforated tape functions as a series of five light valves, allowing the passage of up to five beams of light in combinations of 1, 2, 3, 4 and 5 open perforations, for exposure on photographic paper. The system incorporates a series of lenses, all training the light source whether various or single onto the same target point. The point here is five tones -- this would later be extended to 16.

The analysis of the image

1,758,388 (1930) introduces a "progressive 'lateral' variation of the reader points...from what may be termed the standard position." The Bartlane system creates five prints, each of which is exposed "for varying lengths of time to the action of light... one of these prints, which may be called the first print, and which is exposed for the shortest period of time, only records the brightest tones of the picture to be reproduced, the second print records the tones of the first print and in addition the next brightest tones, etc., until the last print records ubstantially all of the tones of the picture to be transmitted. As stated in [195,577], these prints are formed upon a metal plate by means of a film of hardened gelatin, which acts as an efficient insulator."

In order to secure perfect recording or transmission of the tone values of a print, it would be necessary for the pitch of the said helical path to be equal to the width or diameter of the reader point. This, however, is impossible for a number of practical reasons, and in actual practice it has been found necessary to impart a definite pitch to the said helical path so that the reader points do not pass over each and every consecutive small area of each print. This results in a noticeable loss of detail when the reader points simultaneously contact with corresponding small areas of the respective prints, which has been the practice up to the present time.

According to this invention, the reader points are placed so as to simultaneously contact with corresponding small areas of the respective prints, these different small areas comprising a consecutive series.

For example...

1,758,388 is, apparently, the same as 1925 GB 262,943 -- which however has different (more specific in some aspects, less in others) language but includes no drawings. [this is a problem that is addressed as well by Ranger in his important 1926 paper] See a paper by Wise and Coggeshall in Proceedings of the IRE (May 1941), usefully summarized here.
 

A primary objective of this patent is the elimination of human agency in coding and decoding, by means of sounds from intermediate records on wax cylinders. It does not describe in detail the analysis or initial "exploration" of the original image, for encoding.

"A further object of the invention is to provide an improved method of securing the desired graduation... This phase of the invention comprises forming a plurality of representations from the original, differing one from the other in that they each correspond to given fraction of the tones or ranges of tone in the original.

"Thus one representation may comprise only the deepest tones or full shadows of the original, a second may include only the next lighter tones or both these tones and the full shadows, a third including still lighter tones and so on.

"Four or five such representations will usually be found sufficient, and by their aid a plurality of prints may be produced in register on a single surface in order to provide a final result which shall sufficiently for most practical purposes agree with the gradation in the original." (2:104ff)

Moreover, "The object of the employment of the intermediate record is to facilitate the correction or errors which may arise..." (5:25ff)
 

The elementary signs here are six dots, of six diameters. Presumably a seventh signal -- either "a" or "i" on the Wheatstone tape Fig. 6 -- would cause a blank (white) dot space. Their progressively stepped diameters distinguish these dots from half-tone dots, whose diameters can be any size.

We come here nearly full circle to the ideas embodied in Montagna, Mortier and Ellero, and as far back as Carl Fasol who achieved wonders with but four or five dot diameters.

This patent is substantially the same as 1929/US/1,713,266.
 

Several patents relate to counting elements of a similar shade, and coding that count in combination with the respective shades. Gaynor (1922/UK/190,522) describes a cipher code system in which "the last nine consonants (N, P, Q, R, S, T, V, X and Z)" indicate the number of times a value is repeated along a scanning line. Of course, it takes five letters (each one composed of a five unit code) to yield a five-letter cipher, making this a relatively inefficient means. See also Bartlane 1921/166,917 and 1933/1,910,586 (which allows for up to 32 tone values).

I show Fig. 2 because it so clearly expresses the relationship of the five unit code (with its 32 possible combinations) and efforts to exploit it to economize on tape length (hence transmission time and expense).

Five-unit tape limits the number of elemental signs that can be easily used describe a picture. That limit is not absolute, however; see for example Bartlane 1930/1,758,388 and 1934/GB/404,715. Their 1934 specification explores the tradeoffs in five unit code of, say, six tone combinations (three holes) and maximum four positions (two holes). Electrical Research Products (1934/GB/404,715) provides for 16 values (but does not address "count").
 

Scrambles pictures for secure transmission, by (1) employing variations in speeds of scanning (first line, second line and so on) and (2) interlacing (move ahead four, move back eight); requires synchronization between sending and receiving machines.

Other example of image scrambling devices include 1922 GB 153,597, 1931 GB 354,249 and the much-later 1951 GB 655,320. Scrambling devices are similar in some ways to (1) condensers, that convert figure series into alphabetical ciphers, and (2) to various encryption devices that emerged out of multiplex printing telegraph technology (Hebern, Sanford Vernam, A G Damm, Ingenieursbureau "Securitas", etc.). (e.g., 1922 GB 153,597, 1931 GB 354,249 or the later 1951 GB 655,320)
 

At receiving end, vibrations control light beam through lens light-valves onto photographic paper. Perhaps the most important feature of this patent is the synaesthetic idea of translating a picture into sound; other patents on this page translate a picture into code -- a form of ekphrasis.

Keen appears to have been an illustrator ca 1900-1920, and has several other patents under his name, including :
1916/1,176,147 Telegraphy (transmission of pictures, emphasis on diaphragm at receiving end)
1916/1,176,148 Telegraphy (full system, incorporating a slightly different diaphragm apparatus)
1918/1260393 Apparatus for producing animated drawings (ccl/352/87).