[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: Barcode performance
I've followed with great interest over the past week, the dialog on barcode
performance. Unfortunately, I've not had a chance to respond until now.
This being one of my main consulting areas for Kodak Professional Motion
Imaging, I'd like to add a few facts, figures and observations. I
apologize for the lengthy reply, but hopefully, it will provide some useful
As many of you point out, generally there's little difficulty reading
Keykode numbers on negative films using modern barcode readers. You should
be able to set the sensitivity of a modern Keykode reader and expect to get
accurate barcode readings without further adjustment for any camera stock
in the same format. Of course, reader sensitivity may need to be adjusted
going from 35mm to 16mm films and from negative to reversal stocks.
Early Keykode Numbers and Readers:
As some of you have also pointed out, the edgeprint density on older
Eastman 35mm B&W negative, particularly 5222 manufactured shortly after
Kodak introduced Keykode numbers late 1989, required a higher reader light
intensity. In fact, the edgeprinting on all Kodak films first coded with
Keykode numbers was denser than today. Back then, readers were not readily
available to determine if a different edgeprint exposure would be needed.
It was decided best to maintain previous standards based on optimum human
As Keykode readers became more prevalent and users provided feedback, we
learned that the barcode exposure on those early stocks was greater than it
needed to be. Also, some of the first Keykode readers lacked the higher
intensity required to read the denser code.
Five years ago, Kodak reduced the edgeprint exposure on all Eastman color
stocks to improve machine readability, but not so much as to impair human
readability. The min - max density tolerances were narrowed for greater
consistency across all product lines. The same parameters were applied to
Eastman B&W films in early 1993.
However, B&W camera films are "push-processed" much more frequently than
color films, and a few transfer facilities continued to report problems
reading the barcode. So, in the spring of '95, responding to user
requests, Kodak reduced the latent edgeprint exposure on its B&W films a
Push - Pull Processing:
I'm sure most of you are aware that push- or pull-processing can adversely
effect barcode readability on any film. The picture area and the edges
receive the same process. Therefore, if a cinematographer intentionally
underexposes two-stops and asks for push-processing, the Keykode and
human-readable numbers can be up to four times denser than normal.
However, with manual adjustment, most of the newer readers should be able
to track the dense barcode. Going the other way -- thin and faint barcode
-- makes accurate reading MORE difficult. Too thin is generally worse than
Barcode on Print and Intermediate Films (IP and IN):
If it's the original Keykode numbers on intermediate film you're after, no
problem, given normal processing. Latent edgeprinting on these films
follows the same tight standards as those for Kodak camera films. It's
when print-through Keykode numbers are needed that the rub can come.
Here's where the lab must start playing a more prominent role. From some
of the film samples I've seen and heard about, it would appear that the
rheostats controlling the edgeprint lights on some of the old B&H printers
haven't been adjusted since these veritable old workhorses were installed
in the mid-'50s or before. And some of the optics may not have been
cleaned since then either.
Neg cutters have suffered for years with marginal edgeprinting, but somehow
they've learned how to decipher faint and blurry numbers. Barcode readers
are not so intuitive. With more print and master positive being
transferred, particularly in the features business, good edgeprinting and
barcode readability have become much more important We're hoping film labs
will see the need to improve their edgeprinting.
It's not that difficult, really. The barcode density range on print can be
quite wide and still very readable. Here are some guidelines:
35mm & 65mm edgeprint - densitometer reading status A red density:
Acceptable range 1.3 - 2.5 Optimum 1.9
16mm edgeprint Acceptable range 1.8 - 2.6 Optimum 2.2
That's a density range of 800% for 35mm and 65mm edgeprint, and almost 400%
More Than Density:
Important as it is, proper density is only one of the factors effecting
barcode readability. Uniform exposure, edge contrast, symbol contrast and
modulation are also very important, whether it's barcode on film or on a
box of corn flakes. These ubiquitous elements can be precisely analyzed
using a screen-capture oscilloscope, reading the waveform output of the
If It Looks Good, Chances Are It Is Good:
But, you can learn a lot about good and bad barcode by simply looking at
the film on a light table with a good loupe. Is the density of the bars
fairly uniform across their height, laterally from film edge inward? Any
encroaching edge fog? (Can be particularly troublesome on the tiny 16mm
barcode.) Are the edges of the bars sharply defined? Good contrast
between bars and spaces? If there's a lack of contrast, ask the lab to
read the density of the spaces as well as the bars. (Density patches
provided on all Kodak films for this purpose.) A red density difference of
1.6 or greater on print is ideal. However, I've seen good readability with
less contrast if overall density and other factors are on target.
We've conducted extensive edgeprint tests at the Kodak R&D labs, commercial
labs in New York, Virginia and Los Angeles, and determined that Keykode
numbers imaged on print stock and intermediates can be read over a wider
range of reader intensities than most original negative. We've also run
tests proving that Keykode numbers read well through several generations --
IP to IN to print. In fact, the Kodak Keykode verification films -- 35mm,
16mm and super 16mm -- have the original Keykode and human-readable numbers
imaged through three generations. We had to produce them this way in order
to create all the picture elements needed to test the accuracy of the total
system -- barcode reader, proper frame/field digitization and
frame-accurate film cut lists from video editing systems.
A long dissertation, but one I hope you've found somewhat helpful. Your
comments and questions are most welcome.
Don Ver Ploeg