WORKED EXAMPLE · SCANNED PDF REDACTION

Scanned 1983 NASA press kit redacted, then partly un-redacted

A communications team had an 8-page scan of a 1983 NASA press kit to send to a design agency. They asked to black out every phone number and every person’s name, burned into the scan, while job titles, agencies and companies stayed readable.

read the scan and removed 26 names and 5 phone numbers from the page images and from the text, touching nothing else. Then legal relaxed the rule, and went back to the original scan to make a second version with the 13 senior officials visible again.

  • 8scanned pages, old OCR text
  • 31redactions burned into the scan
  • 0redacted names left in the text
  • 13names restored from the original

Open the real files

These are the actual files from the run. Download them, zoom in, and try to copy a name out.

Where the file comes from: NASA Technical Reports Server, NASA News Release 83-44, "Press Kit: Project RCA (Satcom 1R)", 31 March 1983. A US Government work, public use permitted.

01 · THE STARTING FILE

Eight scanned pages with a noisy text layer

Every page is a black-and-white scan. The PDF also carries text from old OCR, and that text is noisy: it reads “Joseph 6. Mahon” for Joseph B. Mahon and “Wi11iam Palme” for William Palme. The press contacts are on page 2, the NASA/Industry Team list is on pages 6 and 7, and phone numbers are on pages 1, 2 and 8.

A black box drawn over a scan hides nothing: the picture is still underneath, and so is the text. A real redaction has to change the pixels and delete the words.

Image from the source documentPDF page 2 of the 1983 press kit: two press contacts with their phone numbers, then the release
Page 2: the letterhead number and the two press contacts.Open the original
Image from the source documentPDF page 6 of the 1983 press kit: the NASA Headquarters and Goddard team list, names on the left, titles on the right
Page 6: the NASA/Industry Team list, names on the left, titles on the right.Open the original

02 · REQUEST 1

"Redact every phone number and every person’s name"

What the comms team typed

We're sending this 1983 NASA press kit (Satcom 1R launch) to our design agency as reference for a heritage campaign. Before it goes, please redact:

1. Every phone number: the 'AC 202 755-8370' letterhead number on pages 1 and 2, the two press-contact numbers on page 2, and the recorded-updates number in the note to editors on the last page.
2. The names of individual people: the two press contacts at the top of page 2 and every name in the NASA/Industry Team list on PDF pages 6-7 (NASA Headquarters, Goddard, Kennedy and RCA American Communications), including prefixes like 'Lt. Gen.' and 'Dr.'.

Job titles, company and agency names and places stay readable, including company names that contain a person's name (e.g. Charles Stark Draper Labs on page 8). It's a scan, so the redactions must be burned into the page image, not just boxes on top.

What did

  • Read all 8 pages and found each name and number on the page, including the ones the old OCR had misspelt.
  • Redacted 26 names, with their prefixes (“Lt. Gen.”, “Dr.”, “, Jr.”), and 5 phone numbers: 31 redactions on pages 1, 2, 6, 7 and 8.
  • Burned each one into the scan. The black is part of the page picture now, and the words are deleted from the text.
  • Left every job title, agency, company and place readable, including Charles Stark Draper Labs on page 8. Pages 3 to 5 were not touched.

Before and after: drag to compare

Page 6 in version 1, every name blacked out, titles readablePage 6 as scanned, names visibleoriginal scanversion 1
Page 6. Drag the handle: the names go, the titles and section heads stay.

Before and after: drag to compare

Page 2 in version 1, the contacts and numbers blacked outPage 2 as scannedoriginal scanversion 1
Page 2: the letterhead number, both press contacts and their numbers.
Picture of the delivered fileThe note to editors on page 8 with its phone number blacked out
Page 8: the recorded-updates number in the note to editors.

We checked the delivered file: copying out its text finds none of the 26 surnames and none of the 5 numbers, every title and company is still there, it still has 8 pages of the same size, and pages 3 to 5 are the original scans, unchanged.

One thing we would tidy: on page 2 the boxes follow the old OCR’s line boxes, which are a little taller than the type, so they nick the line below; and the word “(Phone:” went with the numbers.

03 · REQUEST 2

"Senior officials can stay visible"

What the comms team typed

Legal has relaxed the rule: senior officials can stay visible. Working from the original scan, un-redact the people in the team list whose title contains 'Director', 'Administrator' or 'President' (so Deputy Director, Network Director, Associate Administrator and Vice President all count).

Everyone else, the two press contacts on page 2 and all phone numbers stay redacted. Save it as a new version and list who is now visible.

What did

  • Started again from the original upload, not from version 1: a burned-in redaction cannot be undone, and it did not need to be.
  • Applied the new rule: the 13 people whose title has Director, Administrator or President in it stay visible; the other 13, the two press contacts and all 5 phone numbers are redacted again.
  • Saved it as a new version. Version 1 is still there as the earlier version.
  • Listed who is now visible, section by section, with their titles.

Before and after: drag to compare

Page 6 in version 2, the directors and administrators visiblePage 6 in version 1, every name blacked outversion 1version 2
Page 6: NASA Headquarters and Goddard.

Before and after: drag to compare

Page 7 in version 2, the directors and presidents visiblePage 7 in version 1, every name blacked outversion 1version 2
Page 7: Kennedy and RCA American Communications.

Visible in version 2 13

  • Lt. Gen. J. A. AbrahamsonAssociate Administrator for Space Flight
  • Joseph B. MahonDirector, Expendable Launch Vehicle Program
  • Robert E. SmylieAssociate Administrator for Space Tracking and Data Systems
  • Dr. Noel W. HinnersDirector (Goddard)
  • John J. QuannDeputy Director
  • William C. KeathleyDirector, Project Management
  • Ralph BanningNetwork Director
  • Richard G. SmithDirector (Kennedy)
  • Thomas S. WaltonDirector, Cargo Operations
  • Charles D. GayDirector, Expendable Vehicles Operations
  • Dr. James J. TietjenPresident (RCA)
  • John ChristopherVice President of Technical Operations
  • Joseph SchwarzeDirector, Space Systems

Still redacted 13

  • press contact, page 2
  • press contact, page 2
  • Manager, Delta Program
  • Acting Delta Project Manager
  • Deputy Delta Project Manager
  • Manager, Delta Mission Analysis and Integration
  • RCA Satcom 1R Mission Integration Manager
  • RCA Satcom 1R Network Support Manager
  • Chief, Automated Payloads Division
  • Chief, Delta Operations Division
  • Spacecraft Coordinator
  • Manager, Advanced Space Programs
  • Manager, Spacecraft Engineering

We checked it: the 13 names that came back are the original scan’s own pixels, with no trace of a box. “Joseph B. Mahon” reads correctly on the page even though the old OCR text still says “6.”. Copying out the text finds the 13 senior names and none of the other 13 or the phone numbers.

Redaction that changes the file, not just the view

Most tools draw boxes on top. takes the words out of the page and out of the text, on scans too, and keeps your original, so a changed rule is a new version instead of a redo by hand.

KEEP READING

Read another worked example

See all 19 worked examples All examples for Marketing & communications