hpr-knowledge-base/hpr_transcripts/hpr0026.txt

Episode: 26
Title: HPR0026: Intro to codecs
Source: https://hub.hackerpublicradio.org/ccdn.php?filename=/eps/hpr0026/hpr0026.mp3
Transcribed: 2025-10-07 10:26:34

---

MUSIC
Welcome to Hacker Public Radio. My name is Klaatu. I'm here to talk about
Codex, Video Codex, and Audio Codex alike. This is going to probably end up
being sort of an in-depth series since this is such a huge subject. I want to
instill a little bit of an understanding in you about how Codex work and kind
of what they do and why they exist at all. But I also want to maybe talk
about each or some of the major Codex out there and maybe some of some of the
features of each and why they exist, where they come from, things like that,
whether they're free, non-free, stuff like that. And of course the other side of
I mean the different sides of Codex are the compression and the decompression of
the content. So I'd like to talk a little bit about that. So initially I think
what is a Codex? That's something that a lot of people don't quite understand.
A lot of people know empirically that they need a Codex because when they click
on a file or they go to a site, the video doesn't play and it tells them to
install a Codex or a plug-in for their browser. So a lot of people realize that
there's a need for them to install a Codex. But not many people really know
what a Codex is. The word Codex is actually it's like modem or something like
that. It's not really, you know, we say it as a word, it's actually two words
put together. It's code and decode, CO for code, DEC for decode. And the idea
behind a Codex simply is at least in its most pure form and I'll explain what
I mean by that in a little while. In its most pure form, a Codex simply
exists because video does not originate on your computer. Obviously what we
see in the real world is not digital. Something needs to make them digital. And
this is where Codex come in. So a Codex is simply a mathematical method of
taking a signal and turning it into a digital stream of information. So Codex
are actually at work, hard at work on your camera well before you ever get to
your computer. I mean obviously your digital video cameras are, you know,
obviously they have embedded computers in them. So when you're taking video,
whether it's mini DV or HDV or AVCHD or whatever, the video is being
processed by a codec and it's being stored in a certain format and then
you're transferring that to your computer to either edit it or to store it or
to convert it to DVD or whatever you're going to do without video. So obviously
Codex have to exist. If we want to have video on our computers, the question
is do they have to exist in the way that they exist? And the answer to that is
probably not. Codex are also a big business. There are companies out there that
are in the business of having created a Codex and having come up with a
scheme to make people come to them to be able to use that Codex. It's not always
the end user that needs to go to that company. Sometimes we do in order to
download the Codex especially in Linux where there are very few Codex included
with our operating system. But even, you know, I mean if you've paid for your
operating system, if you've paid for either Windows or Mac OS 10, someone from
one of those two operating systems has gone to these companies, signed a license
agreement, paid whatever amount of money needs to be paid in order to get that
Codex to be bundled with that operating system, and of course you're paying
that eventually when you're paying for the operating system. So there's a
business to Codex and it's a really profitable business. And interestingly enough,
a lot of people believe incorrectly that there are so many Codex out there
because each Codex is designed for a certain use. You know, one Codex must
be better than another for such and such a job, whilst another Codex must be
better than the other one for this job. That's usually not the case. While
certain Codex sometimes do have a kind of specialty or a, you know, maybe they've
been created with the intent to sell it to a certain market. It really is
just, it's usually just that. It's just someone started a company, got some
Codex together, had them make a Codex and geared it toward a certain market
because they see that there's a need for video delivery. So Codex can be really,
really uncompressed. And that's simply, like I say, that's the most pure form of it.
Uncompressed video would just be a really, really, really big file that you'd
need a really nice computer system to be able to even play back to watch. More
typically, what we find in video cameras, especially, for instance, something like
mini-dv or motion JPEG, which you'll find in some of the digital still cameras
that, you know, also have the video camera function. Not the iPhone, of course,
because the iPhone, as advanced as it is, doesn't have that function. But other
cameras, other phones even have little video cameras in them. Motion JPEG. So
that's a very, very compressed picture because obviously if you're taking it on
your phone, it's not as if though you have that much, not that much space on it
to hold, you know, video. So it's very, very compressed. Mini-dv, which is, I guess,
kind of right now it's the lowest consumer video format, is actually quite
compressed as well. And you start to see that when you bring it into different
color correction programs, starting, you know, you try to kind of isolate
certain things. It gets kind of difficult to do that with such compression
because you're losing a lot of color depth and things like that. And then
there's obviously, you know, there's like HDV, which is kind of in between
mini-dv and HD. And that's not quite as compressed, but it's compressed. So you've
got, you know, the codec by nature is to have the video on the computer. And then
there's different kinds of, there's different kinds of compression that those
codecs are applying to that video. And usually that serves a very useful
function. Like I say, if you want to take video on your camera and you don't want
to lug around a big, you know, ridiculously large digital still camera with, you
know, like an eight gig hard drive attached to it or something like a, you
know, I mean, then then you're going to have to compress that video. If you've
got a system that you want to edit video on and your system doesn't have, you
know, 10 processors and eight gigs of RAM and a gig of video RAM, then you're
going to need to compress that video so that your system doesn't totally choke
on the video that you're feeding it and demanding that it plays back to you in
real time so that you can get a feel for it so you can edit it. So the, the
codecs that compress things force a certain amount of, it's a balancing act.
Obviously, the artist part of the person taking the video ideally would not
have to compress the video at all. It would be a, a pure stream of, of basically
exactly what the camera could capture. There would be no compression. It would
just be a perfect image, a digital translation of what we see with our own eyes
basically. Realistically, on a technical side, that's just not going to happen
either because the system that you're running it on isn't going to be able to
handle it or the system that you're delivering it to isn't going to be able to
handle it. So think about internet video delivery. There's no way that you're
going to be able to send someone uncompressed video, you know, when they go
and download your video netcast or your, your movie, your open source movie or
something like that. It's just not going to be possible. You have to compress it
so that they can download it within a reasonable amount of time. But you have to
compress it only so much and you want to obviously leave it looking as good as
possible. And so it gets really, really tricky and there's a certain art and
there's a certain science to compression and it is fairly complicated to
compression. It's something that you really kind of have to get to know, but it
really helps to know what exactly is going on when you're doing the
compression. And that'll help you wrap your mind around how to go about
compressing video. So assuming we've gone out and gotten some video and we come
back and we've edited it or whatever we're going to do with it, we now, we
now need to compress it so that we can send it out to other people so that they
can actually watch it. And whether you're going out to DVD or whether you're
going to offer it as a download over the internet or you're going to compress
it onto a video CD or you're going to put it on your own media player, your
in 800 or something like that, you have to keep in mind your what you're
delivering to because of course it couldn't be that simple. Every different
device needs a different kind of compression method or could benefit from
having its own compression method. The process of encoding video or compressing
video is to take the video in its native codec whatever it was captured in into
the camera and now you're taking it and you're going to transcode it into a
format that is more ideal for whatever delivery method you've chosen. You can
do this on Linux on the command line. I think there are some GUI tools as well.
I usually just do it on the command line. However you do it you you really need
to know what the different what the important variables are because honestly
there are only a couple of a really important variables and if you understand
what they do you'll have you'll have part of what you need to understand how
you're going to go about compressing the video. The interesting thing about
compression is that only half of it is understanding what needs to be done. The
other half is actually looking at the video that you are compressing on a
case-by-case basis and kind of analyzing it with your own eyes and and kind of
processing in your own mind how that video is going to treat or how your
compression method is going to treat that video. So the idea here is that if you
have a video of people blowing up buildings and shooting lasers and
swinging around lightsabers that's going to need a pretty that's a lot of
information happening in that video frame and if you really start thinking about
video frames traditionally I think we all think of it kind of as we know film
is you know as a series of images being shown to us you know a series of
still images being shown to us at 29 or 30 or 24 frames per second and it's
simply the the illusion of motion that we have from seeing all these still
images and while technically it's it's sort of the same thing in video it's
actually a lot more helpful to think of it more like how the computer sees it.
The computer doesn't really see it as a set of still images the computer sees
it as a block of pixels and each of those pixels has a value each pixel has a
Luma value and a Chroma value the Luma value is the brightness or darkness of
that pixel and the Chroma value is the hue the color of that pixel those two
things in each little pixel is what the computer is looking at and then
think about how many pixels there are in even a standard definition frame
typically it's 720 pixels by 480 which is basically more math than I can do in
my head right now but that's a lot of pixels you know it's a lot of pixels for
the computer to think about and that is 720 by 480 approximately 30 times a
second so now you've got even more for the computer to consider so if you think
about your frames of video as just a box of pixels of ever changing pixels you
kind of just analyze the picture on that basis so if you've got people swinging
your own lightsabers you've got pixels that are changing value both Luma and
Chroma value just a lot in one second you know that lightsaber is gonna go
from one end of the string to the next that's a streak of light that that
goes across the frame but then the pixels that it has passed through are going to
go back to their old value after it passes through that area you know so and
then you've got people around in that frame and you've got sparks flying around
and you've got all that kind of stuff going on that's a lot of information and
that that kind of compression you're not going to be able to compress it that
much and be able to retain clarity and you know the high quality of image on the
other hand if you if I was filming myself right now or video taking myself
right now all we would have in the frame was me talking into a microphone and
really nothing else you know there wouldn't be a whole lot of movement I'm
not I'm not really moving around it's just my lips removing and if I even could
seal my lips behind the microphone the video camera wouldn't even see
that much and so it would it would practically be like a still image and so
that's that's a lot of that's a lot of pixels that basically aren't changing
value and that kind of stuff is very easy for the computer to compress and it's
very easy to have a you know to give that kind of frame a very small amount of
what's called bitrate so I've probably given you enough to think about for this
episode I'm going to wrap it up now and I'm gonna do a second episode in which we
will go through what exactly is happening during during compression and what
those important variables to think about are so that you can kind of get a
sense of what you should be thinking about when you set about compressing
video and then probably in the third and probably final episode I will tackle a
couple of different codecs and talk about what they all are and why they
exist and who owns them and what we should do to deal with them
thank you for listening to H.P.R.R. sponsored by caro.net so head on over to
H.P.R.O.N.C. all of us in here