The fanzines collection displayed on this site was compiled by three teams that took different approaches in the OCR (Optical Character Recognition) process. OCR uses a scanned image of text, such as a page from a fanzine, and converts it into the text that the image represents.
Comic Fandom Monthly Oct 1971 No 2's Process
The group responsible for Comic Fandom Monthly used a program called freeOCR. The program first downloaded the .tif files of the fanzine before running the OCR application through each of the image files to produce .txt files. The .txt files were then compared to the images and mistakes were manually corrected.
Initially, an OCR was performed on pages from The Myriad using Islandora. This method proved to be ineffective, with several errors resulting. Though these errors are not uncommon in word-recognition software and employing the Islandora OCR software was deemed user-friendly and generally suitable, the results were considered to be sub-standard by the development team. Transcribing each page manually was deemed more efficient than the error detection and corrections involved with performing an OCR.
A Companion in Zeor 2's Process
The team responsible for converting A Companion in Zeor 2 chose to manually type the text from the image of each page. Originally, the team started with the intention of using the OCR capablities provided by Adobe Acrobat. A combination of Adobe Photoshop and Adobe Acrobat was used in the scanning and OCR process. As the only team with scans that included two pages from the zine on one image, they found that Acrobat was successful with reading the text as two pages separately instead of reading the two pages together from left to right. But the greatest challenge laid with the actual text recognition. A combination of factors such as typography and fading led to the inadvertent misreading of the image's text; therefore the produced text contained a multitude of small errors. Rather than proofreading and editting when necessary, they found that simply typing out each page by hand was a more efficient method to produce the desired result.
Trivial Pursuits #148's Process
The member responsible for Trivial Pursuits #148 used Acrobat's OCR function and extracted the text into a .txt file. This method was shown to be ineffective, resulting in several errors. The OCR software is incapable of deciphering typewriter font as well as the writing inside of the comics. The OCR software tried to read all of the text inside of images and was not able to translate it properly, due to the font imitating a handwritten style.
Starwind Science Fiction and Fantasy's Process
The team member responsible for Starwind's OCR process used Adobe Acrobat Pro DC's automatic OCR program. It was efficient and quick but rather inaccurate, especially when images or stylized text was involved. It was handy in that it could easily export the original file as a searchable pdf, plain text, or other sorts of documents making it a good way to create any file form needed.