Added some notes on DVD-o-matic's files to the manual.

[dcpomatic.git] / doc / manual / dcpomatic.xml

<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE book [
<!ENTITY % sgml.features "IGNORE">
<!ENTITY % xml.features "INCLUDE">
<!ENTITY % dbcent PUBLIC "-//OASIS//ENTITIES DocBook Character Entities V4.5//EN"
   "/usr/share/xml/docbook/schema/dtd/4.5/dbcentx.mod">
%dbcent;
<!ENTITY % extensions SYSTEM "extensions.ent">
%extensions;
]>
<book xmlns="http://docbook.org/ns/docbook" version="5.0" xml:lang="en">

<bookinfo>
<title>DCP-o-matic</title>
<author><firstname>Carl</firstname><surname>Hetherington</surname></author>
</bookinfo>

<chapter xmlns="http://docbook.org/ns/docbook" version="5.0" xml:lang="en">
<title>Introduction</title>

<para>
Hello, and welcome to DCP-o-matic!
</para>

<section>
<title>What is DCP-o-matic?</title>

<para>
DCP-o-matic is a program to generate <ulink
url="http://en.wikipedia.org/wiki/Digital_Cinema_Package">Digital
Cinema Packages</ulink> (DCPs) from DVDs, Blu-Rays, video files such as MP4
and AVI, or still images.  The resulting DCPs will play on modern digital
cinema projectors.
</para>

<para>
You might find it useful to make DVDs easier to present, to encode
independently-shot feature films, or to generate local advertising for
your cinema.
</para>

</section>

<section>
<title>Licence</title>

<para>
DCP-o-matic is licensed under the <ulink url="http://www.gnu.org/licenses/old-licenses/gpl-2.0.html">GNU GPL</ulink>.
</para>

</section>

</chapter>

<chapter xmlns="http://docbook.org/ns/docbook" version="5.0" xml:lang="en">
<title>Installation</title>

<section>
<title>Windows</title>

<para>
To install DCP-o-matic on Windows, simply download the installer from
<ulink url="http://carlh.net/software/dcpomatic">http://carlh.net</ulink>
and double-click it.  Click through the installer wizard, and
DCP-o-matic will be installed onto your machine.
</para>

<para>
If you are using a 32-bit version of Windows, you will need the 32-bit
installer.  For 64-bit Windows, either installer will work, but I
suggest you used the 64-bit version as it will allow DCP-o-matic to
use more memory.  You may find that DCP-o-matic crashes if you run
many parallel encoding threads (more than 4) on the 32-bit
version.
</para>

</section>

<section>
<title>Ubuntu Linux</title>

<para>
You can install DCP-o-matic on Ubuntu 12.04 (&lsquo;Precise
Pangolin&rsquo;) or 12.10 (&lsquo;Quantal Quetzal&rsquo;) using
<code>.deb</code> packages: download the appropriate package from
<ulink
url="http://carlh.net/software/dcpomatic">http://carlh.net</ulink> and
double-click it.  Ubuntu will install the necessary bits and pieces
and set DCP-o-matic up for you.
</para>

</section>

<section>
<title>Other Linux distributions</title>

<para>
Installation on non-Ubuntu Linux is currently a little involved, as
there are no packages available (yet); you will have to compile it
from source.  If you are using a non-Ubuntu distribution, do let me
know via the <ulink url="mailto:dcpomatic@carlh.net">mailing
list</ulink> and I will see about building some packages.
</para>

<para>
The following dependencies are required:
<itemizedlist>
<listitem><ulink url="http://ffmpeg.org/">FFmpeg</ulink></listitem>
<listitem><ulink url="http://www.mega-nerd.com/libsndfile/">libsndfile</ulink></listitem>
<listitem><ulink url="http://www.openssl.org/">OpenSSL</ulink></listitem>
<listitem><ulink url="http://www.openjpeg.org/">libopenjpeg</ulink></listitem>
<listitem><ulink url="http://www.imagemagick.org/script/index.php">ImageMagick</ulink></listitem>
<listitem><ulink url="http://www.boost.org/">Boost</ulink></listitem>
<listitem><ulink url="http://www.libssh.org/">libssh</ulink></listitem>
<listitem><ulink url="http://www.gtk.org/">GTK</ulink></listitem>
<listitem><ulink url="http://www.wxwidgets.org/">wxWidgets</ulink></listitem>
<listitem><ulink url="http://carlh.net/software/libdcp/">libdcp</ulink></listitem>
</itemizedlist>
</para>

<para>
Once you have installed the development packages for the dependencies,
download the source code from <ulink
url="http://carlh.net/software/dcpomatic">http://carlh.net</ulink>,
unpack it and run the following commands from inside the source
directory:
</para>

<programlisting>
./waf configure
./waf build
sudo ./waf install
</programlisting>

<para>
With any luck, this will build and install DCP-o-matic on your system.  To run it, enter:
</para>

<programlisting>
dcpomatic
</programlisting>

<para>
in a shell.
</para>

</section>
</chapter>

<chapter xmlns="http://docbook.org/ns/docbook" version="5.0" xml:lang="en">
<title>Creating a video DCP</title>

<para>
In this chapter we will see how to create a video DCP using
DCP-o-matic.  We will gloss over some of the finer details, which are
explained in later chapters.
</para>

<section>
<title>Creating a new film</title>

<para>
Let's make a very simple DCP to see how DCP-o-matic works.  First, we
need some content.  Download the low-resolution trailer for the open
movie <ulink url="http://sintel.org/">Sintel</ulink> from <ulink
url="http://ftp.nluug.nl/ftp/graphics/blender/apricot/trailer/Sintel_Trailer1.480p.DivX_Plus_HD.mkv">their
website</ulink>.  Generally, of course, one would want to use the
highest-resolution material available, but for this test we will use
the low-resolution version to save everyone's bandwidth bills.
</para>

<para>
Now, start DCP-o-matic and its window will open.  First, we will
create a new &lsquo;film&rsquo;.  A &lsquo;film&rsquo; is how DCP-o-matic refers to
a piece of content, along with some settings, which we will make into
a DCP.  DCP-o-matic stores its data in a folder on your disk while it
creates the DCP.  You can create a new film by selecting
<guilabel>New</guilabel> from the <guilabel>File</guilabel> menu, as
shown in <xref linkend="fig-file-new"/>.
</para>

<figure id="fig-file-new"> 
  <title>Creating a new film</title> 
  <mediaobject>
    <imageobject> 
      <imagedata fileref="screenshots/file-new&scs;"/>
    </imageobject> 
  </mediaobject>
</figure>

<para>
This will open a dialogue box for the new film, as shown in <xref
linkend="fig-video-new-film"/>.
</para>

<figure id="fig-video-new-film"> 
  <title>Dialogue box for creating a new film</title> 
  <mediaobject>
    <imageobject> 
      <imagedata fileref="screenshots/video-new-film&scs;"/>
    </imageobject> 
  </mediaobject>
</figure>

<para>
In this dialogue box you can choose a name for the film.  This will be
used to name the folder to store its data in, and also as the initial
name for the DCP itself.  You can also choose whereabouts you want to create
the film.  In the example from the figure, DCP-o-matic will create a
folder called &lsquo;DCP Test&rsquo; inside my home folder (carl) into which it
will write its working files.
</para>

<para>
If you always create your DCPs in a particular folder, you can use
DCP-o-matic's <guilabel>Preferences</guilabel> to make life a little
easier by setting the default folder that DCP-o-matic will offer in this dialogue.
See <xref linkend="ch-preferences"/>.
</para>

</section>

<section>
<title>Selecting content</title>

<para>
The next step is to set the content that you want to use.  Click the
content selector, as shown in <xref
linkend="fig-click-content-selector"/>, and a file chooser will
open for you to select the content file to use, as shown in <xref
linkend="fig-video-select-content-file"/>.
</para>

<figure id="fig-click-content-selector">
  <title>Opening the content selector</title> 
  <mediaobject>
    <imageobject> 
      <imagedata fileref="screenshots/click-content-selector&scs;"/>
    </imageobject> 
  </mediaobject>
</figure>

<figure id="fig-video-select-content-file"> 
  <title>Selecting a video content file</title> 
  <mediaobject>
    <imageobject> 
      <imagedata fileref="screenshots/video-select-content-file&scs;"/>
    </imageobject> 
  </mediaobject>
</figure>

<para>
Select your content file and click <guilabel>Open</guilabel>.  In this
case we are using the Sintel trailer that we downloaded earlier.
</para>

<para>
When you do this, DCP-o-matic will take a look at your file.  After a
short while (when the progress bar at the bottom right of the window
has finished), you can look through your content using the slider to
the right of the window, as shown in <xref linkend="fig-examine-thumbs"/>.
</para>

<figure id="fig-examine-thumbs"> 
  <title>Examining the content</title>
  <mediaobject>
    <imageobject> 
      <imagedata fileref="screenshots/examine-thumbs&scs;"/>
    </imageobject> 
  </mediaobject>
</figure>

<para>
Dragging the slider will move through your video.  You can also click
the <guilabel>Play</guilabel> button to play the content back.  Note
that there will be no sound, and playback might not be entirely
accurate (it may be slightly slower or faster than it should be, for
example).  This player is really only intended for brief inspection of
content; if you need to check it more thoroughly, use another player
such as <ulink url="http://projects.gnome.org/totem/index.html">Totem</ulink>, <ulink url="http://www.mplayerhq.hu/design7/news.html">mplayer</ulink> or <ulink url="http://www.videolan.org/vlc/index.html">VLC</ulink>.
</para>

</section>

<section>
<title>Setting up</title>

<para>
Now there are a few things to set up to describe how the DCP should be
created.  The settings are divided into four tabs: film, video, audio and subtitles.
</para>

<section>
<title>Film tab</title>

<para>
The &lsquo;film&rsquo; tab contains settings that pertain to the whole film, as shown in <xref linkend="fig-film-tab"/>.
</para>

<figure id="fig-film-tab"> 
  <title>Film settings tab</title>
  <mediaobject>
    <imageobject> 
      <imagedata fileref="screenshots/film-tab&scs;"/>
    </imageobject> 
  </mediaobject>
</figure>

<para>
The first thing here is the name.  This is generally set to the title
of the film that is being encoded.  If <guilabel>Use DCI
name</guilabel> is not ticked, the name that you specify will be used
as-is for the name of the DCP.  If <guilabel>Use DCI name</guilabel>
is ticked, the name that you enter will be used as part of a
DCI-compliant name.
</para>

<para>
Underneath the name field is a preview of the name that the DCP will
get.  To use a DCI-compliant name, tick the <guilabel>Use DCI
name</guilabel> checkbox.  The DCI name will be composed using details
of your content's soundtrack, the current date and other things that
can be specified in the DCI name details dialogue box, which you can
open by clicking on the <guilabel>Details</guilabel> button.
</para>

<para>
If the DCP name is long, it may not all be visible.  You can see the
full name by hovering the mouse pointer over the partial name.
</para>

<para>
The <guilabel>Trust content's header</guilabel> button starts off
checked, and this means that DCP-o-matic will use the content's header
information to determine its length.  If, for some reason, this header
length is wrong, uncheck the <guilabel>Trust content's
header</guilabel> button and DCP-o-matic will run through the content
to find its exact length.  This may take a while for large pieces of content.
</para>

<para>
Next up is the content type.  This can be
&lsquo;feature&rsquo;, &lsquo;trailer&rsquo; or whatever; select the
required type from the drop-down list.
</para>

<para>
The <guilabel>trim frames</guilabel> settings allow you to trim frames
from the beginning and end of the content; any trimmed frames will not
be included in the DCP.
</para>

</section>

<section>
<title>Video tab</title>

<para>
This tab contains settings related to the picture in your DCP, as shown in <xref linkend="fig-video-tab"/>.
</para>

<figure id="fig-video-tab"> 
  <title>Video settings tab</title>
  <mediaobject>
    <imageobject> 
      <imagedata fileref="screenshots/video-tab&scs;"/>
    </imageobject> 
  </mediaobject>
</figure>

<para>
The first option on this tab is the format.  This will govern the
shape that DCP-o-matic will make your image into.  Select the aspect
ratio that your content should be presented in.  The &lsquo;4:3 within
Flat&rsquo; and &lsquo;16:9 within Flat&rsquo; settings will put the
image at the specified ratio within a Flat (1.85:1) frame, so that you
can project the DCP using your projector's Flat preset.
</para>

<para>
The remaining options can often be left alone, but may sometimes be
useful.  The &lsquo;crop&rsquo; settings can be used to crop your
content, which can be used to remove black borders from round the
edges of DVD images, for example.  The specified number of pixels will
be trimmed from each edge, and the content image in the right of the
window will be updated to show the effect of the crop.
</para>

<para>
The &lsquo;filters&rsquo; settings allow you to apply various video
filters to the image.  These may be useful to try to improve
poor-quality sources like DVDs.  We will discuss filtering later in the manual.
<!-- XXX: link -->
</para>

<para>
The &lsquo;scaler&rsquo; is the method that will be used to scale up
your content to the required size for the DCP, if required.  We will
discuss the options in more detail later; Bicubic is a fine choice in
most situations.
<!-- XXX: link -->
</para>

<para>
The &lsquo;colour look-up table&rsquo; specifies the colour space that
your input content will be expected to be in.  If in doubt, leave it
set to &lsquo;sRGB&rsquo;.
</para>

<para>
Finally, the &lsquo;JPEG2000 bandwidth&rsquo; setting changes how big the final
image files used within the DCP will be.  Larger numbers will give
better quality, but correspondingly larger DCPs.  The bandwidth can be
between 50 and 250 megabits per second (MBps).
</para>

</section>

<section>
<title>Audio tab</title>

<para>
This tab contains settings related to the sound in your DCP, as shown in <xref linkend="fig-audio-tab"/>.
</para>

<figure id="fig-audio-tab"> 
  <title>Audio settings tab</title>
  <mediaobject>
    <imageobject> 
      <imagedata fileref="screenshots/audio-tab&scs;"/>
    </imageobject> 
  </mediaobject>
</figure>


<para>
&lsquo;Audio Gain&rsquo; is used to alter the volume of the
soundtrack.  The specified gain (in dB) will be applied to each sound
channel before it is written to the DCP.
</para>

<para>
If you use a sound processor that DCP-o-matic knows about, it can help
you calculate changes in gain that you should apply.  Say, for
example, that you make a test DCP and find that you have to run it at
volume 5 instead of volume 7 to get a good sound level in the screen.
If this is the case, click the <guilabel>Calculate...</guilabel>
button next to the audio gain entry, and the dialogue box in <xref
linkend="fig-calculate-audio-gain"/> will open.
</para>

<figure id="fig-calculate-audio-gain"> 
  <title>Calculating audio gain</title>
  <mediaobject>
    <imageobject> 
      <imagedata fileref="screenshots/calculate-audio-gain&scs;"/>
    </imageobject> 
  </mediaobject>
</figure>

<para>
For our example, put 5 in the first box and 7 in the second and click
<guilabel>OK</guilabel>.  DCP-o-matic will calculate the audio gain
that it should apply to make this happen.  Then you can re-make the
DCP (this will be reasonably fast, as the video data will already have
been done) and it should play back at the correct volume with 7 on
your sound-rack fader.
</para>

<para>
Current versions of DCP-o-matic only know about the Dolby CP750.  If
you use a different sound processor, and know the gain curve of its
volume control, <ulink url="mailto:cth@carlh.net">get in
touch</ulink>.
</para>

<para>
&lsquo;Audio Delay&rsquo; is used to adjust the synchronisation
between audio and video.  A positive delay will move the audio later
with respect to the video, and a negative delay will move it earlier.
</para>

<para>
By default the <guilabel>Use content&lsquo;s audio</guilabel> button
will be selected.  This means that the DCP will use one of the
soundtracks from your content file; you can select the soundtrack that
you wish to use from the drop-down box.
</para>

<para>
Note that if your content's audio is mono, DCP-o-matic will place it
in the centre channel in the DCP.
</para>

<para>
Alternatively, you can supply different sound files by clicking the
<guilabel>Use external audio</guilabel> button and choosing a WAV file
for any channels that you want to appear in the DCP.  These files can
be any bit depth and sampling rate, and will be re-sampled and
bit-depth converted if required.
</para>

</section>
<section>
<title>Subtitles tab</title>

<para>
This tab contains settings related to subtitles in your DCP, as shown in <xref linkend="fig-subtitles-tab"/>.
</para>

<figure id="fig-subtitles-tab"> 
  <title>Subtitle settings tab</title>
  <mediaobject>
    <imageobject> 
      <imagedata fileref="screenshots/subtitles-tab&scs;"/>
    </imageobject> 
  </mediaobject>
</figure>

<para>
DCP-o-matic will extract subtitles from the content, if present, and
they can be &lsquo;burnt into&rsquo; the DCP (that is, they are
included in the image and not overlaid by the projector).  Note that
DVD and Blu-Ray subtitles are stored as bitmaps, so it is not possible
(automatically) to use non-burnt-in subtitles with these sources.
Select the <guilabel>With Subtitles</guilabel> checkbox to enable
subtitles.  The <guilabel>offset</guilabel> control moves the
subtitles up and down the image, and the <guilabel>scale</guilabel>
control changes their size.
</para>

<para>
Future versions of DCP-o-matic will hopefully include the option to
use text subtitles (as is the norm with most professionally-mastered
DCPs).
</para>

</section>
</section>

<section>
<title>Making the DCP</title>

<para>
Now that we have set everything up, choose <guilabel>Make
DCP</guilabel> from the <guilabel>Jobs</guilabel> menu.  DCP-o-matic
will encode your DCP.  This may take some time (many hours in some
cases).  While the job is in progress, DCP-o-matic will update you on
how it is getting on with the progress bar in the bottom of its window, as shown in <xref linkend="fig-making-dcp"/>.
</para>

<figure id="fig-making-dcp">
  <title>Making the DCP</title>
  <mediaobject>
    <imageobject> 
      <imagedata fileref="screenshots/making-dcp&scs;"/>
    </imageobject> 
  </mediaobject>
</figure>

<para>
When it has finished, the DCP will end up on your disk inside the
film's folder.  You can then copy this to a projector via a USB
stick, hard-drive or network connection.  See <xref
linkend="ch-files"/> for details about the files that DVD-o-matic creates.
</para>

<para>
Alternatively, if you have a projector or TMS that is accessible via
SCP across your network, you can upload the content directly from
DCP-o-matic.  See <xref linkend="sec-tms-upload"/>.
</para>

</section>
</chapter>


<chapter xmlns="http://docbook.org/ns/docbook" version="5.0" xml:lang="en">
<title>Creating a still-image DCP</title>

<para>
DCP-o-matic can also be used to create DCPs of a still image, perhaps
for an advertisement or an on-screen announcement.  This chapter shows you
how to do it.
</para>

<para>
As with video DCPs, the first step is to create a new
&lsquo;Film&rsquo;; select <guilabel>New</guilabel> from the
<guilabel>File</guilabel> menu and the new film dialogue will open as
shown in <xref linkend="fig-still-new-film"/>.
</para>

<figure id="fig-still-new-film"> 
  <title>Dialogue box for creating a new film</title> 
  <mediaobject>
    <imageobject> 
      <imagedata fileref="screenshots/still-new-film&scs;"/>
    </imageobject> 
  </mediaobject>
</figure>

<para>
Enter a name and click <guilabel>OK</guilabel>.  Then we set up the
content; click the content selector as before, and this time we will
choose an image file, as shown in <xref
linkend="fig-still-select-content-file"/>.
</para>

<figure id="fig-still-select-content-file"> 
  <title>Selecting a still content file</title> 
  <mediaobject>
    <imageobject> 
      <imagedata fileref="screenshots/still-select-content-file&scs;"/>
    </imageobject> 
  </mediaobject>
</figure>

<para>
Setting up for a still image DCP is somewhat simpler than for a video;
the tabs are all the same, but many options are removed and a few are added.
</para>

<para>
As with video, you can select a content type and the format (ratio)
that your image should be presented in.  It will be scaled and padded
to fit the selected ratio, but in such a way that the pixel aspect
ratio is preserved.  In other words, the image will not be stretched,
merely scaled; if you want to stretch your image, you will need to do
so in a separate program before importing it into DCP-o-matic.  You
can also crop your image, if you so choose, and then set a duration
(in seconds) that the image should appear on screen.
</para>

<para>
Still-image DCPs can include sound; this can be added from the
<guilabel>Audio</guilabel> tab.  If your specified duration is shorter
than the audio, the audio will be cut off at the duration; if it is
longer, silence will be added after your audio.
</para>

<para>
Finally, as with video, you can choose <guilabel>Make DCP</guilabel>
from the <guilabel>Jobs</guilabel> menu to create your DCP.  This will
be much quicker than creating a video DCP, as DCP-o-matic only needs
to encode a single frame which it can then repeat.
</para>

</chapter>


<chapter xml:id="ch-preferences" xmlns="http://docbook.org/ns/docbook" version="5.0" xml:lang="en">
<title>Preferences</title>

<para>
DCP-o-matic provides a few preferences which can be used to modify its
behaviour.  This chapter explains those options.
</para>

<section>
<title>The preferences dialogue</title>

<para>
The preferences dialogue is opened by choosing
<guilabel>Preferences...</guilabel> from the <guilabel>Edit</guilabel>
menu.  The dialogue is shown in <xref linkend="fig-prefs"/>.
</para>

<figure id="fig-prefs"> 
  <title>Preferences</title> 
  <mediaobject>
    <imageobject> 
      <imagedata fileref="screenshots/prefs&scs;"/>
    </imageobject> 
  </mediaobject>
</figure>

<section>
<title>TMS setup</title>

<para>
The first part of the dialogue gives some options for specifying
details about your TMS.  If you do this, and your TMS accepts SSH
connections, you can upload DCPs directly from DCP-o-matic to the TMS.
This is discussed in <xref linkend="sec-tms-upload"/>.
</para>

<para>
<guilabel>TMS IP address</guilabel> should be set to the IP address of
your TMS, <guilabel>TMS target path</guilabel> to the place that DCPs
should be uploaded to (which will be relative to the home directory of
the SSH user).  Finally, the user name and password are the
credentials required to log into the TMS via SSH.
</para>
</section>

<section>
<title>Threads</title>

<para>
When DCP-o-matic is encoding DCPs it can use multiple parallel threads
to speed things up.  Set this value to the number of threads
DCP-o-matic should use.  This would typically be set to the number of
processors (or processor cores) in your machine.
</para>

</section>

<section>
<title>Default directory for new films</title>

<para>
This is the directory (folder) which DCP-o-matic will suggest initially as a place to put new films.
</para>

</section>

<section>
<title>A/B options</title>

<para>
These options are for DCP-o-matic's special mode of making A/B
comparison DCPs for checking the performance of video filters.  Their
use is described in <xref linkend="sec-ab"/>.
</para>

</section>

<section>
<title>Encoding servers</title>

<para>
If you have spare machines sitting around on your network not doing
much, they can be pressed into service to speed up DCP encodes.  This
is done by running a small server program on the machine, which will
encode video sent to it by the &lsquo;master&rsquo; DCP-o-matic.  This
option is described in more detail in <xref linkend="sec-servers"/>.
Use these preferences to specify the encoding servers that should be
used.
</para>

</section>

</section>
</chapter>

<chapter xmlns="http://docbook.org/ns/docbook" version="5.0" xml:lang="en">
<title>Advanced topics</title>

<para>This chapter describes some parts of DCP-o-matic that are
probably not essential, but which you might find useful in some
circumstances.
</para>

<section>
<title>Filtering</title>

<para>
DCP-o-matic offers a variety of filters that can be applied to your
video content.  You can set up the filters by clicking the
<guilabel>Edit</guilabel> button next to the filters entry in the
setup area of the DCP-o-matic window; this opens the filters selector
as shown in <xref linkend="fig-filters"/>.
</para>

<figure id="fig-filters"> 
  <title>Filters selector</title> 
  <mediaobject>
    <imageobject> 
      <imagedata fileref="screenshots/filters&scs;"/>
    </imageobject> 
  </mediaobject>
</figure>

<para>
After changing the filters setup, you will need to regenerate the DCP
to see the effect on the cinema screen.  The preview in DCP-o-matic
will update itself whenever filters are changed, though of course this
image is much smaller and of lower resolution than a projected image!
</para>

</section>

<section>
<title>Scaling</title>

<para>
If your source material is not of the DCI-specified size, or if it
uses non-square pixels, DCP-o-matic will need to scale it.  The
algorithm used to scale is set up by the <guilabel>Scaler</guilabel>
entry in the film setup area.  We think &lsquo;Bicubic&rsquo; is the
best all-round option, but tests are ongoing.
</para>

</section>

<section xml:id="sec-tms-upload">
<title>TMS upload</title>

<para>
If you have configured details of a TMS in the preferences dialogue
(<xref linkend="ch-preferences"/>) you can upload a completed DCP
straight to your TMS buy choosing <guilabel>Send DCP to TMS</guilabel>
from the <guilabel>Jobs</guilabel> menu.
</para>

</section>


<section xml:id="sec-ab">
<title>A/B comparison</title>

<para>
When evaluating the effects of different filters or scalers on the
image quality, A/B mode might be useful.  In this mode, DCP-o-matic
will generate a DCP where the left half of the image uses some
&lsquo;reference&rsquo; filtering and scaling, and the right half of
the image uses a different set of filters and a different scaler.
This DCP can then be played back on a projector and the image quality
evaluated.
</para>

<para>
To enable A/B mode, click the A/B checkbox in the setup area of the
DCP-o-matic window.  When you generate your DCP, the left half of the
screen will use the filters and scaler specified in the <xref
linkend="ch-preferences">preferences</xref> dialogue, and the right
half will use the filters and scaler specified in the film setup.
</para>

</section>

<section xml:id="sec-servers">
<title>Encoding servers</title>

<para>
One way to increase the speed of DCP encoding is to use more
than one machine at the same time.  An instance of DCP-o-matic can
offload some of the time-consuming JPEG2000 encoding to any number of
other machines on a network.  To do this, one &lsquo;master&rsquo;
machine runs DCP-o-matic, and the &lsquo;server&rsquo; machines run
a small program called &lsquo;servomatic&rsquo;.
</para>

<section>
<title>Running the servers</title>

<para>
There are two options for the encoding server;
<code>servomatic_cli</code>, which runs on the command line, and
<code>servomatic_gui</code>, which has a simple GUI.  The command line
version is well-suited to headless servers, especially on Linux, and
the GUI version works best on Windows where it will put an icon in the
system tray.
</para>

<para>
To run the command line version, simply enter:
</para>

<programlisting>
servomatic_cli
</programlisting>

<para>
at a command prompt.  If you are running the program on a machine with
a multi-core processor, you can run multiple parallel encoding threads
by doing something like:
</para>

<programlisting>
servomatic_cli -t 4
</programlisting>

<para>
to run 4 threads in parallel.
</para>

<para>
To run the GUI version on windows, run the &lsquo;DCP-o-matic encode
server&rsquo; from the start menu.  An icon will appear in the system
tray; right-click it to open a menu from whence you can quit the
server or open a window to show its status.
</para>

</section>
<section>
<title>Setting up DCP-o-matic</title>

<para>
Once your servers are running, you need to tell your master
DCP-o-matic instance about them.  Start DCP-o-matic and open the
<guilabel>Preferences</guilabel> dialog from the
<guilabel>Edit</guilabel> menu.  At the bottom of this dialog is a
section where you can add, edit and remove encoding servers.  For each
encoding server you need only specify its IP address and the number of
threads that it is running, so that DCP-o-matic knows how many
parallel encode jobs to send to the server.
</para>

<para>
Once this is done, any encodes that you start will split the workload
up between the master machine and the servers.
</para>

</section>
<section>
<title>Some notes about encode servers</title>

<para>
DCP-o-matic does not mind if servers come and go; if a server
disappears, DCP-o-matic will stop sending work to it, and will check
it every minute or so in case it has come back online.
</para>

<para>
You will probably find that using a 1Gb/s or faster network will
provide a significant speed-up compared to a 100Mb/s network.
</para>

<para>
Making changes to the server configuration in the master DCP-o-matic
will have no effect while an encode is running; the changes will only
be noticed when a new encode is started.
</para>

</section>
</section>

</chapter>

<chapter xml:id="ch-files" xmlns="http://docbook.org/ns/docbook" version="5.0" xml:lang="en">
<title>Generated files</title>

<para>
DCP-o-matic generates a number of files as it makes a DCP.  <xref
linkend="fig-file-structure"/> shows the files that might be generated
after you have created a DCP for a film called &lsquo;DCP Test&rsquo;.
</para>

<figure id="fig-file-structure"> 
  <title>Creating a new film</title> 
  <mediaobject>
    <imageobject> 
      <imagedata fileref="diagrams/file-structure&scs;"/>
    </imageobject> 
  </mediaobject>
</figure>

<para>
The <code>DCP Test</code> folder is the one that you specify when you
select the <guilabel>New Film</guilabel> option from DCP-o-matic's
menu.  Everything is stored inside this folder.
</para>

<para>
DCP-o-matic generates some working files as it goes along.  These are as follows:
<itemizedlist>

<listitem><code>log</code> is a list of notes that DCP-o-matic makes as it goes
along.  This can be useful for debugging purposes if something goes
wrong.</listitem>

<listitem><code>metadata</code> stores the settings that you have made
for this film: things like cropping, output format and so on.</listitem>

<listitem><code>video</code> is where DCP-o-matic writes the DCP's
video data as it encodes it.</listitem>

<listitem><code>analysis</code> is used to keep the results of audio analysis runs.</listitem>

<listitem><code>info</code> contains details of each video frame that
DCP-o-matic has written so far.  This is used when an encoding
operation is interrupted and DCP-o-matic must resume it.</listitem>
</itemizedlist>
</para>

<para>
Following this is the DCP itself:
<code>DCP-TEST_EN-XX_UK-U_51_2K_CSY_20130218_CSY_OV</code>.  This
contains some small XML files, which describe the DCP, and two large
MXF files, which contain the DCP's audio and video data.  This folder
(<code>DCP-TEST_EN-XX_...</code>) is what you should ingest, or pass
to the cinema which is showing your DCP.
</para>

</chapter>


</book>