aboutsummaryrefslogtreecommitdiff
path: root/README
blob: 589d961daa7faf557f127473e8492669dec4b1a1 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
PDP - Pure Data Packet v0.12.2
a packet processing library for pure data

Copyright (c) by Tom Schouten <pdp@zzz.kotnet.org>

This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.

The GNU Public Licence can be found in the file COPYING


------------------------------------------------------------------

This external pd library is a framework for image/video processing (and 
other raw data packets) that is fast, simple and flexible and as much 
as possible compatible with the pd messaging system.

This distro contains the main pdp library and 3 extension libs:
* pdp_scaf (in directory scaf/)
* pdp_opengl (in directory opengl/)
* pdp_guile (in directory guile)

Features:

* packet formats: greyscale and YCrCb encoded images,
  binary cellular automata (CA), textures, render buffers,
  matrices.
* sources: noise, video4linux and quicktime (with audio), CA, plasma
* sinks: xvideo display, sdl, glx display
* filters: convolution, biquad time, biquad space, CA
* warping: rotate, zoom, stretch
* transforms: colour translation/scaling, grey->palette
* add, mul, mix, random pixel mix
* utility objs: packet register, snapshot, trigger
* packet delay line, loop, ..
* and more.. (see doc/reference.txt)

Optional features

* cellular automata simulator and effects processor with
  built in forth extension language (only for mmx)
* opengl 3d processing (gem like, but built around packets)
* embedded guile interpreter

See the README files in the scaf/ opengl/ and guile/ dirs for more
info on the extension libraries.


Requirements:

* pd
* linux
* libgsl
* a video4linux device for video input. special support for 
philips webcam included.
* libquicktime (not quicktime4linux!) for quicktime playback.
* libpng for png image loading/saving
* an X display with XVideo extension, glx or SDL for video display.


Documentation:

When you use "make install" to install pd, all docs will be in
the pd doc directory. Right click on an object and select help
for a help patch. The pdp directory in pd's doc contains a
reference.txt file which lists all objects and abstractions and
an introduction and example section. 

NOTE: all help patches use a pdp_help_input and pdp_help_output
abstraction. you can edit these to choose which input or output
object you want for the help patches.


Building:

./configure
make

Options for configure:
--enable-mmx		compile with mmx support
--enable-pwc		force pdp_v4l to use philips web cam

If pd is not installed in /usr/local you'll have to
specify the prefix on the configure command line with

./configure --prefix=/prefix

type

make install

To install pdp in $prefix/lib/pd: the library in externs/
the abstractions in extra/ and the documentation in doc/5.reference
and doc/pdp

You can also try the "buildall" script. This will configure, build
and install the whole distribution. It works for intel/mmx.

NOTE: If you're not using the standard pd install location, and have
everything in a single tree, just untar pdp next to the pd source tree,
or set the prefix dir to the location of the pd source.


Bugs:

See the TODO file. Apart from the few items listed on top of this
file, pdp is fairly stable: i am unaware of crash bugs. If you encounter 
a crash please consider sending a bug report.


PiDiP Is Definitely In Pieces

Have a look at Yves Degoyon's PiDiP library, the must have addition
to the basic pdp library. It contains a lot of extra effects (from 
EffecTV and FreeJ), a quicktime recording object, streaming modules,
ascii art objects ... (the list is growing rapidly ;)
Yves also did a gem2pd and a pdp2gem object to connect pdp and gem
together.

http://ydegoyon.free.fr/


Acknowledgements

PDP is no longer a one man show. Many thanks to the people who 
directly or indirectly contributed to this project. Either by 
writing a giant extension lib (thanks Yves) or by contributing code 
snippets (Thanks Martin, CK), or by giving mental support, feedback, 
bug reports and ideas. 

I also wish to thank the GEM crew. This project borrows a lot of ideas 
and some code from the GEM project. And while i'm at it, i wish to
thank everyone who contributed code to the PD project as a whole and
the people that are very helpful at answering questions on the mailing
list.

Last but not least i wish to thank Miller for making the PD project
open source. If that didn't happen i wouldn't know what i'd be doing
now :)


Some Remarks: 

* New versions of this package can be found at
http://zwizwa.fartit.com/pd/pdp
Experimental relases are in the subdirectory test/

* If you have libquicktime and quicktime4linux installed, this can 
give problems with the configure script. Consider removing 
quicktime4linux. (libquicktime is a drop-in replacement).

* The modules that depend on system libraries only get compiled when
the configure script can locate them. If compilation or linking fails 
please let me know. 

* Packets can be processed in a low priority thread with a dropping 
mechanism to prevent overload or audio drops, or they can be processed 
in the main pd thread. This can be set/unset by sending a "thread x"
message to pdp_control. Processing in pdp thread is off by default.
Use the pd thread for rock solid video timing. Increase the audio 
latency to avoid dropouts. If you want low audio latency, and don't
care about a dropped video frame here and there, switch on the pdp
thread. Note that when using the pdp thread, the control flow is no
longer depth first. Additinal delays will be introduced.

* Pdp runs on osx, using an x server (apple or fink) and libquickime 
(fink). pdp_xv does not work yet. Once apple includes xv support in
their x server, it should. Until then, use pdp_glx for output.

* Some quicktime remarks: pdp_qt~ is still a bit experimental, you might 
try pdp_yqt in yves' lib to see if it works better for you. Also, it 
seems that libquicktime does not work with compressed headers. The 
suggested codec is jpeg (photo) and uncompressed audio with the same 
samplerate as the one pd is using. You can use other codecs, but i've
found this one to perform the best. (hint: compile libquicktime with
mmx jpeg support) Try to avoid the mpga codec. It is haunted.

* The reason i use YV12 and not RGB is simple: it's faster, and for 
linear operations it doesn't make much difference. Most camera's and
codecs use a luma/chroma system, and most video cards support hardware
accellerated display. Note that in YUV, the u (Cb) and v (Cr) components 
can be negative, so saturation after multiplication with a positive 
(or negative) value produces not only black and white, but also red, 
green, blue and their complements. In short, in YUV, nonlinear operations 
have a different effect than in RGB. Another thing: all the spatial 
operations are not implemented "correctly" because of the subsampling 
involved.The image packets are 16bit/component planar. This is to 
simplify  the mmx code and to have a little more headroom and precision 
for processing.

* Since version 0.11 there is a type system used for identifying a
packet type and performing type conversion. A type is represented
by a symbol, with subtypes separated by the "/" character. I.e.
"image/grey/320x240" is a grey 16 bit/component image with dimensions
320 by 240. For conversions a wildcard can be specified. I.e.
"image/*/*" which matches image type packets with all encodings
and dimensions.

* Since version 0.12 there are 2 different image data types: "image/*/*"
and "bitmap/*/*". The image type is the native 16 bit/component pdp
type. It supports the subtypes "image/grey/*": one channel grey
scale images, "image/YCrCb/*": luma/chroma packets with subsampled
chroma planes and "image/multi/*": a multi channel planar format (i.e.
to store 3 or 4 channel rgb or rgba data). Almost all processors
use this data type. The bitmap type is an intermediate type which
can contain standard fourcc encoded 8 bit/component images i.e.
"bitmap/rgb/*", "bitmap/yv12/*", "bitmap/grey/*". Currently no processors
support this type natively, so you have to use pdp_convert to convert
them to "image/*/*".

* Writing new objects is not much different than normal pd externs.
Pdp just adds some functions to the pd kernel for package allocation,
manipulation and communication. The api is very raw to keep it as
simple as possible. Since 0.10 there is a pdp_base object you can
derive your objects from. Have a look at pdp_add, pdp_gain and 
pdp_noise to see how it works. There is not much documentation, but
i do my best to keep things clean so the code and the header files
should get you started. You can always just send me an email for info,
or ask on the pd-dev list. Since version 0.12 there is a simple forth
scripting language in pdp. This is far from finished, but it should
make pdp programming a lot easier.

* Philips webcam detection is now automatic. However this requires a 
fairly recent kernel (one which includes the pwc driver version 8.6 
or higher). If you have a pwc and pdp_v4l does not display a
"pwc detected" message, you can conifigure with --enable-pwc. This
forces pdp_v4l to use the pwc api extensions.



Directory structure:

abstractions/		some abstractions that use the pdp objects
doc/introduction/	getting started with pdp
doc/examples/		some example patches
doc/objects/		pd style reference documentation
include/		header files
modules/		pd object code
opengl/			opengl extension lib (experimental)
scaf/			CA extension lib
guile/			scheme extension lib
system/			core pdp system




Bugreports, feature suggestions, code, documentation, example patches
and general comments and questions are welcome.

Have Fun,

Tom


last modified: 2003/07/20