dvgrab version 1.01, August 04, 2001
(c) 2000-2001 Arne Schirmacher dvgrab@schirmacher.de

Changes in version 1.01

	- PAL/NTSC detection improved

	- minor changes to compile with g++ 3.0 (contributed by
	  Daniel Kobras)

	- buffer space for frames doubled: now 100 frames instead
	  of 50 (about 3-4 secs, 16 MByte RAM total)

	- bugfix: dvgrab will now reliably close the current
	  AVI file when aborted by ctrl-c (contributed by Glen
	  Nakamura)


Changes in version 1.0

	- adapted to libraw 0.9

	- packet assemble code modified - some camcorders send
	  only 492 bytes instead of 496


Changes in version 0.99

	- fixed a bug that caused dvgrab to crash when grabbing very
	  big files in AVI Type 2 format and NTSC mode

	- fixed the incorrect size of files in raw format (they always
	  had an extra 480 bytes)

	- fixed exception thrown when writing a 0 frames file

	- automake and debian files provided by Daniel Kobras


Changes in version 0.89:

	- added the raw1394_stop_iso_rcv call in the close routines


Changes in version 0.88:

	- support for 12 bit audio in AVI Type 2 files (courtesy
	  Mike Pieper)


Changes in version 0.87:

	- new feature (courtesy Stephane Gourichon): there is now a
	  parameter that allows to save every n-th frame only. Great
	  for monitoring purposes or when you need a rough overview of
	  the contents of a tape.

	- Type 2 audio has been slightly improved. Still no 12-bit
	  audio, but I'm getting close.

	- refactoring and cleanup of the code.

	- a few bug fixes: the --format raw mode now works again.


Changes in version 0.86:

	- some improvements in the Type 2 AVI audio code. Many
	  camcorders send audio data which is not defined in my 314m
	  specification, which caused clicks and distortions, because
	  the audio data was left out.
	  If you have extensive audio error messages, your camcorder
	  is using an unsupported audio mode, such as 12 bit audio data.

	- the --testmode parameter reads DV data from a testdata.raw
	  file created previously with the --format test parameter. No
	  camcorder is required.


Changes in version 0.85:

	- handling of additional file formats: 
		--format raw
		--format test


Changes in version 0.84:

	- major rewrite. dvgrab now uses a separate thread for
	  handling the ieee1394 interface. This improves the behavior
	  under load conditions very much.


Changes in version 0.83:

	- there is now audio support for --format dv2 AVIs.
	  Currently only 16 bit 48 kHz sound is supported.


Changes in version 0.82:

	- incorrect audio speed value fixed, overflow error for long
	  Type 2 AVIs fixed.

	- Added support for the --format raw data format.
	  This is just the raw DV data (120000 bytes for each NTSC
	  frame or 144000 bytes for each PAL frame) written to a
	  file. Can be used with the playdv program from the libdv
	  project.


Changes in version 0.81:

	- fixed a bug that prevented Type 1 AVIs from playing with
	  some players.


Changes in version 0.8:

	- the program can now write Type 1 and Type 2 DV AVI files.
	  The difference is that Type 1 contains only one data stream
	  which has audio and video data interleaved, whereas Type 2
	  AVIs contain an additional audio stream. Since the video
	  stream still has the audio data interleaved, this is some
	  waste of disk space, but MainActor can't read Type 1 files
	  so I had to add support for this other format too.

	  Note that this version currently writes empty audio tracks.
	  This means: no sound, but at least it loads nicely into
	  MainActor.

	  A free demo version of MainActor can be downloaded from:

	  http://www.mainconcept.com 

	- Because of the increasing number of supported file formats
	  I have changed the option for selecting them.

	  To write MainActor compatible AVI files use this command:

	  dvgrab --format dv2 outfile


Changes in version 0.7:

	- the previous algorithm for assembling the data packets did
	  not work for some camcorders. I rewrote it and the new
	  algorithm explicitly calculates the address in the frame
	  buffer for each data packet.

	- the program attempts to auto-detect NTSC and PAL format.
	  The --pal or --ntsc command line parameter is no longer
	  mandantory, but it will override the auto-detect feature if
	  applied. Use them if the auto-detect feature doesnt work
	  for you.

	- The --timestamp parameter will put the date and time of the
	  recording into the AVI file name. If it doesnt do that,
	  dvgrab cant find any date and time related packages in the
	  DV data.	  

	- I included the little utilities riffdump (prints out the 
	  structure of an arbitrary AVI file) and rawdump (prints
	  out an ASCII dump of a file created with the --raw
	  option)


Changes in version 0.6:

	- switched to C++ code. The reason is that C++ provides the
	  exception mechanism, which I use for a more reasonable
	  error reporting.
	  Also the whole concept of the RIFF/AVI file structure leads
	  to an object oriented approach.

	- The program has now an auto-split function (see
	  dvgrab --help): It will start a new file whenever (1) the
	  file size approaches 2 GByte (exceeds 0x70000000 bytes, to
	  be specific), (2) the frame count is exceeded, or (3) the
	  recording of a new scene is detected. The program will loop
	  forever until you press CTRL-C. The current AVI file will be
	  properly closed, so you dont lose any data (at least this
	  is the plan).

	- Closing the file is now reasonably fast.

	- there is some visual feedback if the device driver drops
	  data.


Changes in version 0.5:

	- implemented the extended AVI file format which allows for
	  files > 1 GByte. Note that Linux is currently still limited
	  to a maximum file size of 2 GByte, which corresponds to
	  approx. 14500 frames for PAL format or 17500 frames for
	  NTSC format.

	- more comments in the source code  


Changes in version 0.4:

	- implemented the NTSC file format.

	- there is now an error handler which intercepts the signals
	  1, 2, 3 and 15 and attempts to properly close the AVI
	  file. So you can safely press CTRL-C to stop capturing
	  without losing the data saved so far.


Changes in version 0.3:

	- implemented the raw capturing mode, wich simply saves
	  the raw DV data to a file. Useful for analyzing the data.

	- tested with files up to 2000 frames (288 MByte)


Changes in version 0.2:

	- a basic set of command line options. 
	  Run 'dvgrab --help' for a list.

	- more error checking

	- more comments

	- tested with the latest ohci driver
	  (no changes were necessary though)

Changes in version 0.1:

	- first version
