Before there was Photoshop

Discussion in 'Documentary/Journalistic' started by AlanLichty, Dec 12, 2018.

  1. AlanLichty

    AlanLichty Moderator

    Joined:
    Aug 16, 2017
    Messages:
    7,492
    Likes Received:
    5,895
    Location:
    Vancouver, Washington
    Home Page:
    This borders on a Once Upon a Time tale but in the early 1980's we weren't using the term Digital Imaging to describe computer graphics. There were quite a number of folks working in the field of computer graphics but the notion of a hand held digital camera was a pipe dream at best. I think about how much things have changed every time I grab a slider in Lightroom or Photoshop to edit an image. These were not the Good Old Days(TM).

    I was hired by the Utah Archaeological Center to facilitate computing tasks in the late 1970's after funding for Middle East work dried up. One of the early projects was to develop methods of using remote sensing data for predictive models of archaeological site locations primary based on Landsat data. Landsat 3 was the primary data set we used which did actually use CCD (Charge Coupled Device) sensors similar to our modern camera sensors but were rather cumbersome with profoundly limited resolution. Landsat scenes were purchased from NASA on 9 track tapes and the ground resolution was roughly 75 meters.

    Through grant funding the University of Utah Anthropology department gave me the task of building a computational facility capable of allowing the department faculty to utilize computer graphics in their research. This is a set of photos of that facility and some of the output we generated. Absolutely everything that was generated was homegrown FORTRAN software I wrote back then with the exception of a set of modules we got from NASA to perform analysis on the Landsat data.

    What did it take to process computer graphics? In our case I bought a DEC VAX 11/730 computer, a Kennedy 4500 9-track tape drive, a Ramtek graphics processor and monitor, and a Matrix Instruments camera system to capture the output. Total cost for the hardware came in at around $60,000.

    VAX11-730.jpg

    The graphics processor is between the tape drive and the tapes. The graphics system also included a color monitor that could display a 1280x1024 image (yup - just short of a one megapixel screen). We also added a 60" Summagraphics digitizing table so we could build maps and record areas that had been surveyed off of USGS Quad sheets.

    MatrixCamera.jpg

    The monitor is in the center and the camera system is just to the right of that. The camera system had a small version of the 1280x1024 monitor that displayed images in black and white and had a color wheel so we could get color output by taking a triple exposure with the red, green, and blue parts of the color wheel. I went through a lot of rolls of film getting the triple exposure values calibrated. The attached camera was an Olympus SLR with a fixed focus lens that was calibrated for the small internal monitor. The camera system was custom built to allow it to synchronize with the Ramtek graphics system.

    Computing power was the biggest obstacle - sampling a Landsat scene could take hours of CPU time which I had to do on one of the larger campus computers as our little VAX wasn't up to that kind of number crunching with only 2MB of RAM.

    Examples of what kinds of images we were able to produce as graphics output - I used a FORTRAN subroutine library (Plot79) developed by the Univ of Utah Math department to create vector based graphics and developed a series of programs that allowed our faculty to digitize maps for lectures and publications both for archaeology as well as biological anthropology. All of the output examples shown below are 1 megapixel images.

    This first image was a map of a living bedouin black tent encampment in Petra showing the outlines of the tent living area (light blue) and the distribution of trash and belongings. The intent here was to record an example of how a tent site is used as a model for how to interpret artifact distributions in archaeological sites. Note that the living area itself is the clean spot inside of the tent outline.

    BeitGoblan.jpg

    The next one shows blood haptoglobin genetic samples for North American Indian tribes with the mask of a North America map for locational orientation. This was done entirely in custom FORTRAN code. Things I can do in minutes in Photoshop these days took days of coding back in the mid 80's.

    Haptoglobin.jpg

    I loaded up a portion of a Landsat scene so I could try my hand at false color composites because I thought satellite images were seriously cool. The part that gets rather convoluted - this takes digital output from a satellite which is processed in our computers so we could display it on a screen and then take an analog photo of the image on slide film so I could turn around and scan it into a digital image to show you here :D

    This is a scene in southern Utah with the Green River area in the upper left and Moab down in the lower right. I-70 is visible heading east out of the Green River area at the top. The false color composites typically used the IR band so vegetated areas show up in red.

    LandsatIII.jpg

    For the math geeks I also have a copy of the output from one of the demos in the Plot79 subroutine library - in this case a mathematical function mapped onto polar coordinates:

    Plot79.jpg

    Thanks if you stuck around to the end of this :)
     
    Kurt Harrigan and BarryHamilton like this.
  2. Ben Egbert

    Ben Egbert Forum Helper
    Staff Member

    Joined:
    Aug 19, 2017
    Messages:
    7,119
    Likes Received:
    4,403
    Location:
    American Fork Utah
    Home Page:
    Wow Alan, Fortran? I doubt if anyone uses that anymore. I learned Cad/Cam on a Vax 750 and thought that was pretty cool at the time. I recognize the hardware and yes, we saved our files on tape, 1980.

    You had a really interesting job and great job here showing this bit of history. I am reminded of the movie Apollo 13 when the astronauts asked for a check of their calculations, and the guys at mission control all got out slide rules.
     
  3. AlanLichty

    AlanLichty Moderator

    Joined:
    Aug 16, 2017
    Messages:
    7,492
    Likes Received:
    5,895
    Location:
    Vancouver, Washington
    Home Page:
    I certainly hope Fortran has become a dead language - I wrote thousands of lines of code using that language and honestly hope I never see any of it again :D

    I started out in computers working for the College of Science (Math, Chemistry, Physics, and Biology) computing site with a DEC 20-60. Everything was done in Fortran with those departments. We also had several VAX computers - a VAX 11/750 for handling satellite uplinks for sending data to a Cray site in San Diego, a VAX 11/780 for Physics, and then a VAX 8650 that replaced the DEC 20. We also had a Sun 3/280 server and lots of Sun 3 workstations. C took over as the programming language of choice once the unix workstations took over the desktops. PCs weren't allowed on the network.
     
  4. Ben Egbert

    Ben Egbert Forum Helper
    Staff Member

    Joined:
    Aug 19, 2017
    Messages:
    7,119
    Likes Received:
    4,403
    Location:
    American Fork Utah
    Home Page:
    I learned basic and later c++
     
  5. JimFox

    JimFox Moderator
    Staff Member

    Joined:
    Jul 19, 2017
    Messages:
    10,326
    Likes Received:
    6,936
    Location:
    SoCal
    Home Page:
    Very cool Alan. It's great to see that history, you have documented it well. It was fun seeing the photographs and the outputs.

    I was programming at that time in Fortran, Pascal and Assemble and Machine languages. Also there were several different version of Basic that I had learned and used. Probably like you guys, it was all self taught.
     
    AlanLichty likes this.
  6. AlanLichty

    AlanLichty Moderator

    Joined:
    Aug 16, 2017
    Messages:
    7,492
    Likes Received:
    5,895
    Location:
    Vancouver, Washington
    Home Page:
    The NASA software I used for the Landsat based predictive modeling (ELAS - Earth Laboratory Application System) had a core set of modules that were written for a 16 bit PDP-11 computer. I was running the number crunching code on a 36 bit DEC 20 mainframe and had to rewrite all of the core I/O modules using assembly language to create Fortran callable subroutines.

    I wasn't entirely self taught - I was mentored by some highly skilled programmers at the College of Science facility including the author of the Plot79 subroutine library. He had some tools to analyze other people's Fortran programs and flag all of the bad practices and input/output code that wasn't portable between different computer architectures. Those programming tools would spank me for bad coding practices and made me learn how to write well structured applications.
     
    Ben Egbert likes this.

Share This Page

  • About Us

    Welcome to Focal World. As a Guest you have access to view the resources here. But it is totally FREE to Register and participate in the forums. If you find FocalWorld useful, please support it be Upgrading your Account to a Premium Account located in your Profile page. It is only $15 a year and will give you added ability of uploading photos directly from your computer and also the allows you to create your own Galleries. We will be sharing the Daily Featured Photos on FaceBook, Instagram and in a Pbase Gallery to help share FocalWorld with others and to draw in new users as they view the incredible photos shared here. All photos share will include the name of the photographer who took them. If you do not want your photo to be shared in this method please let me know. But hopefully everyone will participate in helping to spread the word of FocalWorld in this manner. Thanks, Jim Fox