HDTV using MPlayer

In my desktop computer, I’ve got an nvidia GeForce 6150 PCI-E video card. It has a built-in mpeg2 decoding helper in it. Using Linux, this is called XvMC, or Xvideo Motion Compensation. I finally got it partially usable.

I’m currently watching the news on WJBF, the only channel I get b/c I haven’t fully setup my antenna. Using xvmc is fairly simple:

mplayer -vc ffmpegmc12 -vo xvmc -framedrop dvb://WJBF-DT

I add -framedrop to keep the video and audio sync’ed.  Mplayer will rant about waiting for retrace. Using OpenGL is actually better, for me at least.

mplayer -vo gl2 -framedrop dvb://WJBF-DT

It dropped 14 frames at the beginning, but hasn’t dropped another in 10 minutes. XvMC would have crapped out by now. A/V sync is still perfect. I’ve even got it full screen and it’s resizing from 1280×720 (720 HD resolution) to 1440×900.

X11 Forwarding

A few days, I moved my desktop computer out into the living room. The computer has a tuner card that can receive digital broadcasts. I don’t own a TV, so this is the closest I’ve got to one. I still have to set up lircd before I can use a remote, but I wanted something a bit easier

Today, I discovered a new trick … SSH X11 Forwarding. Now, I can play video on my desktop right from my netbook. It’s wonderful and very simple!

In the following instructions, the “A” refers to the computer playing the video and “B” refers to any other computer.

  1. Make sure “X11Forwarding Yes” is in your sshd_config file on computer A. Restart sshd if you had to uncomment/add it.
  2. On B, run `ssh -XC A`
  3. In the ssh session, type `DISPLAY=:0.0`.
  4. Now run mplayer or any other video player in the  ssh session and it will play on computer A.

If you want  to have a program run on computer A and use the screen on computer B, the process is very similar:

  1. Make sure “X11Forwarding Yes” is in your sshd_config file on computer A. Restart sshd if you had to uncomment/add it.
  2. On B, run `xhost A`
  3. On B, run `ssh -XC A`
  4. Now run mplayer or any other video player in the  ssh session and it will play on computer A.

Step 2 authorizes use of the current display by computer A and only needs to be run once. Also, note that the display variable doesn’t need to be set when using B‘s screen.

Mplayer works wonderfully when playing on my desktop, plus I control it via the ssh session. It has tons of keyboard shortcuts. The ones I use the most are space for pause and the arrow keys for skipping around.

PVR Update

In anticipation of TV shows coming back on, I’ve been tweaking my PVR scripts. These record shows from the digital airwaves right onto my computer. I’ve separated the script into two parts: recording and encoding. Each is a simple one line shell script that can be called easily from the command line or from a cron. The record.sh script is:

/usr/bin/mencoder dvb://$1 -ovc copy -oac copy -o "$2.mpeg" -endpos $3

All this does is copy the video and audio streams directly to the disk. The three parameters are  the station name, as stated in ~/.mplayer/channels.conf, the filename and the end position. This end position can be the number or seconds to record, or a time format of hh:mm:ss. This makes recording fairly simple within the cron.

The encode.sh script contains:

/usr/bin/mencoder -vf scale=640:-2 -ovc xvid -xvidencopts bitrate=1700:threads=2 \
 -oac copy "$1" -o "$1.avi"

This, slightly more complex use of mencoder, will scale the video to 640 wide, the height calculated to keep the aspect ratio. The video is encoded into xvid at 1700kbps using 2 threads. To speed up the encoding process, the audio is not re-encoded, but copied as is.

You will need to play with the number of xvid threads to find the best fps. I have a dual-core and it encodes the fastest with 2 threads. It does need to be an integer, and most dual-cores will work best at 2.

Then, two lines are required in the cron to record and encode the program:

59 19 * * 1 /dvr/record.sh WAGTNBC /dvr/shows/heroes 01:03:00
05 20 * * 1 /dvr/encode.sh /dvr/shows/heroes.mpeg

ffmpeg Cheat Sheet

ffmpeg is a video encoder for linux. This tool is very versatile and can do pretty much any kind of video processing all from the console.

As with any linux tool, there a tons of bells and whistles. This makes the learning curve kind of steep, but with a bit of determination anyone can make sense of it. Without further ado, here’s the cheat sheet:

ffmpeg -i clip1.avi -i clip2.avi -vcodec copy -acodec copy new-clip.aviCopies clip1 and clip2 into new-clip.avi
ffmpeg -i clip.avi -vcodec libxvid -b 800000  -acodec libmp3lame -ab 128 new-clip.aviEncodes a clip with the xvid codec at 800Kbps and mp3 audio at 128Kbps.
ffmpeg -i clip.avi -t 00:05:00 -vcodec copy -acodec copy new-clip.aviCopies the first 5 minutes from clip.avi into new-clip.avi
ffmpeg -i clip.avi -t 00:05:00 -ss 60 -vcodec copy -acodec copy new-clip.aviCopies 5 minutes from clip.avi into new-clip.avi, skipping the first minute.
ffmpeg -i clip.avi -s 640:576 -vcodec libxvid -b 1200000 -acodec copy -o new-clip.aviResizes the video to 640×576, then encoding video using xvid at 1200Kbps and copying the audio directly
ffmpeg -i clip.avi -target ntsc-dvd -b 5000000 dvd-clip.mpegEncodes clip.avi into a dvd-compatible mpeg at 5000Kbps

Mencoder Cheat Sheet

Mencoder is a video encoder for linux. It is part of the mplayer package which also includes a video player. This tool is quite versatile and can do pretty much any kind of video processing all from the console.

As with any linux tool, there a tons of bells and whistles. This makes the learning curve kind of steep, but with a bit of determination anyone can make sense of it. Without further ado, here’s the cheat sheet:

mencoder clip1.avi clip2.avi -ovc copy -oac copy -o new-clip.aviCopies clip1 and clip2 into new-clip.avi
mencoder clip.avi -ovc xvid -xvidencopts bitrate=800  -oac lamemp3 -lameopts cb:br=128 -o new-clip.aviEncodes a clip with the xvid codec at 800Kbps and mp3 audio at 128Kbps.
mencoder clip.avi -ovc xvid -xvidencopts bitrate=800:threads=2  -oac lamemp3 -lameopts cbr:br=128 -o new-clip.aviSame as above, but uses 2 threads for the xvid encoding.
mencoder clip.avi -endpos 00:05:00 -ovc copy -oac copy -o new-clip.aviCopies the first 5 minutes from clip.avi into new-clip.avi
mencoder clip.avi -endpos 00:05:00 -ss 00:01:00 -ovc copy -oac copy -o new-clip.aviCopies 5 minutes from clip.avi into new-clip.avi, skipping the first minute.
mencoder dvb://WJBF-DT -endpos 01:0000 -ovc copy -oac copy -o ugly.betty.mpegTunes to WJBF-DT, records 1 hour of that station to ugly.betty.mpeg
mencoder clip.avi -vf scale=640:-2 -ovc xvid -xvidencopts bitrate=1200:threads=2 -oac copy -o new-clip.aviResizes the video to 640 wide, keeping the aspect ration, then encoding video using xvid at 1200Kbps and copying the audio directly
mencoder tv:// -tv driver=v4l2:width=640:height=480:device=/dev/video0 -ovc lavc -o >(tee webcam-`date +%Y-%m-%d-%H.%M.%S`.avi | mplayer -cache 128)Records the webcam to “webcam-yyyy-dd-mm-hh.mm.ss.avi”, where that is the date, and display it to the screen while recording.
mencoder -idx clip.avi -ovc copy -oac copy new-clip.aviFixes the AVI index of clip.avi, the output being new-clip.avi
mencoder clip.avi -vf cropdetect -o /dev/nullDetects what cropping is needed
mencoder clip.avi -of mpeg -mpegopts format=dvd -lavcopts vcodec=mpeg2video:vbitrate=5000:acodec=ac3 -o dvd-clip.mpegEncodes clip.avi into a dvd-compatible mpeg at 5000Kbps
mencoder -dvd 2 -ovc lavc -lavcopts vcodec=mpeg4 -oac copy -o dvd.aviRips a DVD to dvd.avi

Using Subversion for Server Backups

Last month, I lost the template for Barry Bakes. I was ultimately to blame for it for 2 reasons:

  1. I had altered the default template without copying everything to a new folder.
  2. I had no server backups running.

After a few iterations, I came up with an ingenious way to use subversion and rsync to handle my backups and not waste megabytes every day from duplication. Also, this method doesn’t leave you with a folder full of backup folders. There are 2 folders, the svn repository and the rsync folder.

To do this, you must generate dsa  or rsa keys for passwordless logins.

To set it up svn, you must do the following:

svn create /path/to/svn/repository
rsync -azvv --progress server:/path/ /local/path/
cd /parent/of/local/path/
svn import folder file:///path/to/svn/repository -m 'Initial Import'
rm folder -rf
svn co file:///path/to/svn/repository

Note that everything in bold must be replaced with your specific values. Once that is done, run this shell script once a day to commit changes to subversion:

#/bin/sh

/usr/bin/rsync -azvv --delete --progress server:/path/ /local/path/
cd /parent/of/local/path/
/usr/bin/svn status |
	/bin/grep '^?' |
	/bin/sed 's/^?       /svn add "/g' |
	/bin/sed 's/$/"/g' |
	/bin/bash
/usr/bin/svn status |
	/bin/grep '^!' |
	/bin/sed 's/^?       /svn delete "/g' |
	/bin/sed 's/$/"/g' |
	/bin/bash
/usr/bin/svn commit -m '`date +%Y-%m-%d`'
/usr/bin/svn update

That will update and commit the changes to subversion. It comments the date in yyyy-mm-dd format for each commit. Plus, now you can checkout any date (revision) you want!

Happy Backup

Laptop Fn Keys

On a laptop, there are a few keys that have a blue icon on them. These are function keys. They work when you hold down the blue “Fn” key, which is normally next to the Ctrl key.

There are tons of Fn keys. Not all will have the same icons on them. If you want to know what yours do exactly, there should be an easy to read table in your manual. If you don’t have the manual or can’t find it, check the manufacturer’s website.

Standard Fn Keys

There are volume and mute keys . Look for a speaker. One will be crossed through or X’ed out, that’s mute. Volume up will have a lot of sound waves coming out. Volume down will have small sound waves.

Also, there are two brightness keys. These look like a small computer screen with an icon in the middle. To brighten the screen, hold down Fn and press the one with the larger icon. Use the smaller icon to make it darker.

Another key will control the monitor. It should have two screens on it with a slash between. This will flip back and forth between the laptop screen and the monitor.

There’s usually a Number Lock. Sometimes, you’ll get a Scroll Lock or System Request, even though neither of those keys is really used today.

Bonus Fn Keys

There are a few keys that aren’t on all laptops. One is the wifi key. This will look like an antenna with waves coming off either side. It turns the wifi on and off.

You may have a bluetooth key. It will have that trademark bluetooth icon. You can guess what that one does.

Another key is the sleep key. The icon varies greatly. It may be some Z’s on it, but it will definitely remind you of sleep.

HTML Post #7: CSS & Browsers

In closing up the HTML series, we’re gonna cover stylesheets and differences between browsers.

Stylesheets

This is just a text document that holds all your style information. It is stored in a separate file so the webpages are smaller and keep traffic down on the webserver and the internet.

To set the style of all p tags, you would say:

p {
    color: green;
    text-align: center;
    background: red;
}

That would make everything inside a p tag green, centered, and have a red background. You could do the same with a div tag, table tag, img tag, or any other tag. You can also use a class name:

.christmas {
    color: green;
    text-align: center;
    background: red;
}

Any tag that has class=”christmas” would have those styles. You can combine the two:

p.christmas {
    color: green;
    text-align: center;
    background: red;
}

Now, any p tag that has class=”christmas” will have the style.

The “p.christmas” in that last example is called the selector. There are plenty more types of selectors, including “ancestor descendant” and “parent > child”. Here is more information about selectors and style properties.

Linking

Putting this in the head section will link the page with a stylesheet at http://www.example.com/style.css:

<link rel="stylesheet" href="http://www.example.com/style.css" type="text/css" />

You can even set a print css, only used when printing the page:

<link rel="stylesheet" href="http://www.example.com/print.css" type="text/css"
 media="print" />

A document can have multiple stylesheets. The second one will supercede the first if there’s any duplication. There is one exception. If the first file says this:

p.error {
    color: red !important;
}

Then the second file will have to use “!important” if it is to supercede the color.

Browser Differences

From time to time, you’ll run across something that looks fine in one browser and totally funky in another. If you run into this, do a quick search or two. Odds are that someone else has run into it too.

Many times, the answer will be an IE conditional statement or some sort of CSS Hack. The IE conditional works like this:

<!--[if IE]>
<p>According to the conditional comment this is Internet Explorer</p>
<![endif]-->

It’s really one big html comment! IE will show it. Firefox and the rest will ignore it. You can even say:

<link rel="stylesheet" href="http://example.com/overall.css" type="text/css" />
<!--[if IE 6]>
<link rel="stylesheet" href="http://example.com/ie6.css" type="text/css" />
<![endif]-->

That will tell only IE 6 to use an additional stylesheet.

<link rel="stylesheet" href="http://example.com/ie.css" type="text/css" />Linking

HTML Post #6: Forms

An HTML form is pretty much anything that requests the user for information. They can be used for just about anything from leaving a comment on a blog to uploading a picture to facebook. Even logins are very simple HTML forms.

Here’s an example login, in html:

<p>Please log in.</p>
<form action="http://example.com/login.php" method="post">
Username: <input name="username" /><br />
Password: <input name="password" type="password" /><br />
<input type="submit" value="Log In" />
</form>

The form tag is where the magic happens. It has an action attribute which is where the form sends the information. The form can use one of two methods to send the information: post or get. Post is more secure, and prefered when doing logins.

There are three inputs. The name attribute of the input is sent back to the web server as the name of the data. The type is the type of input, text being the default. You can also use “password”, “submit”, “file”, “checkbox”, “hidden”, or “radio”.

Also, note that all inputs for the form are contained in the form element.

The previous code makes this form:

Please log in.

Username:

Password:

Get Vs. Post

The main difference between Get and Post is how the data is sent. Get data is encoded into the URL. For example, http://www.google.com/search?q=compguyaug is the URL when searching for “compguyaug” on google. This makes get form results bookmarkable, in most cases. Therefore, get works great for searches and data lookups.

Post sends the data separate from the URL. This makes it more secure and should be used for logins and adding/editing/updating any data. That way, the end user can’t bookmark adding a certain customer and keep opening that bookmark and unintentionally adding that customer.

Note on Files

When creating a form for file uploads, make sure you use post. Also, a special encoding type must be specified:

<form enctype="multipart/form-data" method="post"
 action="http://example.com/uploader.php">
...

Also, some servers have a limit as to how large of a file can be uploaded.