|
Reduced
file size is important because it speeds up the transfer of
information from point to point, whether from digital camera to
memory card, card to computer, or computer to any other destination.
Without the efficiencies made possible by the JPEG format,
transferring images from point to point would take much more time,
and require substantially more storage space to contain them.
If efficient transfer and storage of large digital files were not
an issue, the JPEG format might never have been invented. But it is
an issue, and when properly used the compromises are vastly
outweighed by the benefits. The key is knowing how and when to use
the JPEG format so that image quality is compromised as little as
possible.
JPEG is a compressed format. The image data in a JPEG file is
manipulated in order to reduce file size. The reduction can be
substantial; 10, 20, even 100 times smaller than the original
uncompressed image file. The 'cost' to image quality ranges from
undetectably small, to rendering an image virtually useless. The
goal in working with JPEG files, especially as it relates to high
quality photographic imaging, is to keep the cost to image
quality in the 'undetectably small' column. This is easily
accomplished by using JPEG at the right time and in the
right way.
It helps to understand how the JPEG format actually works, and
how it differs from an uncompressed format like TIFF or PSD
(Photoshop file format). The data in an uncompressed image file is
stored pixel-by-pixel; each pixel is accompanied by a unique set of
data describing it. This is true whether an image has 10,000
different color pixels, or 10,000 pixels that are identical, or
nearly identical, in color; each pixel has a unique set of data
describing it.
The JPEG format takes a much different approach to storing pixel
data. Rather than assigning a unique set of data to each pixel, it
maps the location of pixels that are identical (or near
identical) in color and stores that information just once. When a
JPEG file is opened, the location map is used to assign colors back
to each pixel. This is why an unopened JPEG file might report a file
size of 1MB and when opened in Photoshop report a size of 6MB.
To illustrate the efficiency of this strategy, consider this: a
one-color picture containing 10,000 pixels can accurately be
described with a single set of data - the data for one color. An
uncompressed file would contain 10,000 sets of data. That's how JPEG
realizes such huge reductions in file size.
Of course photographic images are made up of many hundreds to
many millions of different colors, but the JPEG compression is the
same: identify the location of like colors, map their location, and
store that information just once for each color.
If you shoot in JPEG format with your digital camera, this
explains why file sizes vary so much from shot to shot. Simple
scenes, say a blue sky over the ocean, result in a much smaller file
size than a mountain meadow full of wildflowers.... even though the
number of pixels used is exactly the same. The greater the number of
colors contained in a scene, the larger the JPEG file.
The terms used up to this point, 'nearly identical color' and
'like colors', are important to our understanding. The JPEG format
allows us to choose a level of compression from 'low' to 'high'. The
higher the compression, the smaller the resulting file. In order to
achieve the highest possible compression, the JPEG format takes a
little liberty in what it considers to be a 'single' color. Colors
that are 'close' to being the same are grouped together and defined
by one set of data saying they 'are' the same. This can save a
considerable amount of space. At 'low' compression, JPEG takes fewer
liberties in grouping. Thus, the lower the compression, the higher
the accuracy of the colors.
Color grouping is illustrated nicely in the photograph of the
flower at the top of this article. Depending on the level of
compression, the result ranges from visibly undetectable to grossly
unacceptable.
Variable compression levels serve to address many different
needs, from extremely small files required for use on the
Web, to extremely high quality files used for photographic printing.
In Photoshop, the range of compression is on a scale from 1 to 12.
Somewhat confusingly, 1 equals maximum compression and 12 equals
minimum compression. There is no single right way to use compression, just different ways
to use it for each type of application.
This is a brilliant strategy for reducing file size while
maintaining image quality, but not without flaws. The biggest issue
is what happens when a JPEG file is repeatedly opened, edited, and re-saved.
In many cases, each time a JPEG file is re-saved, the compression scheme kicks in
anew. Two very close colors from the first JPEG save,
might be grouped into a single color on the next. Do this
enough times, and image quality will degrade.
Because image quality can take a hit with each save, and repeated
JPEG saves ultimately can degrade an image, it's wise to use JPEG
when necessary and other, non-compressed formats like TIFF or PSD when it's not. It's
not possible to define one set of rules that apply to every
workflow or application, but it is possible to state one general
rule: Don't re-save a JPEG unless it is necessary.
There are some exceptions. Depending on what type of editing has
been done to an image, a re-save may not cause quality loss. Simply
rotating an image 90 degrees is an example. There are others. But
unless there is a need to edit and re-save in JPEG, an uncompressed
format like TIFF is a better choice for images that are in the
process of being edited.
A typical workflow, and the one we use for image editing, is
this: Files received in JPEG format are opened and immediately
re-saved as TIFF's or PSD's. All subsequent editing work and
re-saves are done in one of these un-compressed file formats. These
are considered 'working files'. The working files are not re-saved as
JPEGS until editing is complete and they are ready to enter into
production for printing, where efficiency becomes a
priority. Following this workflow prevents a file from being
re-saved as a JPEG more than once after it is received.
Every workflow that includes handling of JPEG files should be
managed in
a similar fashion. JPEG should be used when required due to transfer
speed or storage space, and avoided at other times. Use of the JPEG
format has many advantages and when used carefully, has its rightful
and respected
place in the imaging world.
It is worth noting that original camera files, regardless of file
format, should always be kept in their original unedited form,
saved/archived to CD or DVD disk. This gives you future access to
the original unedited files. Simply copying or moving a JPEG file
(or any file for that matter) from one location to another has no
effect on the data.
For an in-depth, scientific look at JPEG, see this article in
Wikipedia:
http://en.wikipedia.org/wiki/JPEG |
