I recently read an article about MPEG-2 in the New York TImes (should not need registration). I sent the author, David Pogue, a note trying to help explain things. He seemd to appreciate it, so I thought I would share it with you all.

Mr. Pogue,

I read your article on MPEG-2 and can empathize with your travails. I
am a professional television engineer and have been dealing with video
format issues for quite a while.

MPEG-2 is not a format that will be gaining in the marketplace, it is on
the decline. MPEG-2 is one of the older video formats in use today. It
was developed in the mid-1990s as a revolutionary new way to digitize
video. The idea was that you could roughly predict the next frame of
video in a sequence. If you could get the sender and the receiver to
predict the same next frame, you have the sender send to the receiver
only the ACTUAL CHANGE from the prediction. This would save a lot of
bandwidth. There had been a older version called MPEG 1, but it didn't
work so well at the high bitrates that broadcasters use.

The concept worked, the broadcasters said hooray, and began implementing
MPEG 2 systems in professional broadcast, transmission systems like
DirecTV, and DVDs. There were limitations to the format, but since it
was only intended for well-financed professionals to use, no one worried.

The main drawback was all this prediction stuff. It's fine for playing
back video at normal speed, but it gets quite mind-bending to think
about playing it in reverse or editing it. Trust me in saying it is a
Hard Thing To Do.

To solve this editing problem, an alternate method (sometimes called
DCT) for encoding was used. The method used did not use prediction,
thereby solving some of the problems. It did create new ones though.
To get the same image quality you needed a higher bandwidth. It's all
about the tradeoffs.

To compare, a DVD uses MPEG-2 at 4-6 Mb/s and a MiniDV player uses a
form of DCT at 25 Mb/s.

(By the way, DCT stands for Discrete Cosine Transform with means little
to anyone without an engineering or math degree. It's pretty much like
a JPEG image with the lower the resolution, the blockier the picture.)

Then the dot-boom happened with the desire for video over the internet
started driving money into the problem. Lots of money. This money
poured into next-generation forms of encoding and you get the family of
Windows Media, MPEG-4, H.264, etc. All more advanced video encoding
methods that work well as long as you have plenty of processing power to
do the math. Microsoft and Apple are pushing these technologies because
they have the computing power to utilize them and get the benefits of
low bandwidth and high picture quality.

Fast forward to today and you see manufacturers trying get into the home
video market as inexpensively as possible. The MPEG 2 chips are old
(therefore inexpensive) and they utilize less bandwidth than the
comparable DCT type encoding and they don't require computing power
since the encoding/decoding is done on a chip, not in CPU. Put this
together and you have and inexpensive video camera that make
hard-to-edit video.

The trouble with converting MPEG-2 to another format lies in the fact
that MPEG-2 requires a license to legally use the codec. Imagine that
you are a software maker and you want to be able to convert MPEG-2 to
another format as a feature. To do this, you need to pay the MPEG LA
for a license to do this. Usually it's a per copy license and believe
me, that adds up. The license cost is the reason you see MPEG-2 as an
additional costs and often why you need special, separate software to
play DVDs on a computer if the MPEG-2 codec is not included with the
native media player software.

There you go, a long unsolicited reply to your article. I hope it
shines a bit of light on the topic.

-Michael