# Turns out a picture is worth at least 2,380 words

The old cliché “a picture is worth a thousand words” has been around forever so there must be some truth to it. But only a certain kind of obsessive compulsive neurotic (often our favourite people) would go to the trouble to do the maths. David Ash, a consultant at Camwood, the application migration experts, is just such a neurotic.

As it turns out, a picture is worth at least 2,380 words (if ‘worth’ can be correlated with ‘weight in bits’  – a large, but not entirely implausible, ‘if”). Here are David’s workings out:

Average Image Size
JPEG: 11.9K
As determined by Optimization Week (yes, there really is something to optimize every week)

Average Word Length
5.12 Characters
According to the recent (presumably unpaid) work of some Scrabble-happy academic.

Calculations:
11.9K = 11.9 x 1024 bytes = 12185.6 characters

12185.6 characters / 5.12 characters per average word = 2,380 words.

Conclusion:
Even a measly 8-bit colour picture actually paints 2,380 words.
If the picture is made of pixels of greater colour depth, then multiply accordingly:

16-bit picture = 4,760 words
24-bit picture = 7,140 words
32-bit picture = 9,520 words

In other words, as David wriggled free from his straitjacket for long enough to state, “Even the lowest-grade picture actually paints twice as many words as originally thought.” David goes on to point out that the Chinese version of the proverb actually comes closest:  “One picture is worth ten thousand words.

Note to Velocity design team: ths does not mean your salaries should be 10,000 times higher than the copywriters. No one ever seaid the world was fair.

Thanks for the… uh… insight, David.

——

Photo: mindmapinspiration – flickr creative commons

### 2 Responses to “Turns out a picture is worth at least 2,380 words”

1. David Ash

I just had to say…the MindMap graphic itself is worth just over 50,132 words . Yes there is an Internet Kiosk in this soft-walled room.
Dash

2. Doug Kessler

After a tortured night at the chalk board, Dash has issued this update to his calculations:

“I noticed just now that there is a technical flaw my calculations, regarding the ‘deeper colour depth’ extrapolations. The 8-bit calcs are fine. I promise I didn’t do this on purpose!

The flaw is quite subtle, and maybe it’s better to leave it in the blog just to see if anyone else out there is interested enough to comment on it. In any case, I am neurotic enough to describe the flaw below for your records!

An image is comprised of a given number of bits/bytes as you note. It’s the picture’s byte-size that matters, not the X and Y dimensions. So, a 200 x 200 pixel picture at 8-bit colour would occupy 40,000 bytes, but if you convert that same picture to 16-bit colour, it would become 80,000 bytes. This is because the number of pixels is not changing.

The subtlety is that increasing the colour depth would make the byte-size bigger but the picture’s XY dimensions would remain the same (200 x 200). Relating to the blog, once the reader visualises an 11.9K picture in their mind and then imagines it at different colour depths, they are misled in a way because they would likely believe that the same picture is becoming ‘worth more words’ just based on colour depth, without necessarily realising that the picture’s size is actually changing. This makes the extrapolations invalid from certain purist contexts.

I never liked purists anyway.”

Yeah, us too, Dash…