Working around typographic transcription errors

[bugbear]

Searching OCR can be incredibly difficult sometimes, particularly where the print is very faint, so as well as those listed by BugBear I also look out for the following:

cl instead of d
c instead of e
li instead of h
h instead of b
ll instead of H

Good stuff - thank you!

 BugBear

[mrsruz]

Local accents can account for errors too.
I have found Halls entered as Harle in Tyneside & my Brummie ancestors often missed the H from the beginning of a name, or added it when they wanted to appear Posh  eg Halice for Alice.
Pays to know your dialects.

[rayard]

I've had Anna For Hannah.
Searching surname "Yapp" is tricky, "Japp Jopp Topp Tapp" etc. Even when I spoke to someone at the Register Office they said it looked like "Japp."
 (On that certificate they had even written the wrong names in for the groom but he signed it anyway.)
rayard.

[Billyblue]

The problem with rn and m occurs when sans serif typeface is used, eg Arial

When you use a serif font, it looks a bit better  rn & m as there is a minute gap between the r and the n.
Another way to get around it is to go into Font in your program, find character spacing and apply 'expanded' to the 2 letters.
Very fiddly!

Of course this doesn't work with stuff that's already printed and we are just trying to read it.
But sometimes if you can magnify it enough you will see whether it is 2 letters or just 1.

Dawn M

[g eli]

The Capital T, S,F,L, seem to be the most common options to try.
eck often eek
Liz

[clairec666]

Often the loop of a letter "e" isn't very... well... loopy, and it can be confused for a "c" or an "i". Or a double "e" confused for a "u".

[bugbear]

Applying some of the suggestions in this thread to a search in the BNA for "jewellery"

(search term, count)
jewellery   2462893
tewellery   7647
lewellery   6455
jewehery   1589
fewellery   960
sewellery   465
lewehery   51
tewehery   36
fewehery   7
sewehery   1

  BugBear

[andrewalston]

The problem with rn and m occurs when sans serif typeface is used, eg Arial

When you use a serif font, it looks a bit better  rn & m as there is a minute gap between the r and the n.

Of course virtually every old newspaper used a serif font. They also used ink which ran when printed quickly, and that is of course the way newspapers are produced.

When examining a scanned image of an old page, I am often surprised at how good a job an OCR program has made.

I am also amazed how easily our brains cope with sorting out the mess when we read this stuff. We automatically filter the text through our knowledge of the language and the context, mostly without conscious effort.

[Maiden Stone]

t and k
Census transcription of occupation  boot binder  as  book  binder .