ld decode utilities

ld-compress

ld-compress is a script to simplify the compression of .lds (raw LaserDisc RF files) into .ldf images.

ld-decode fully supports FLAC compressed files as input. Files can be suffixed with .ldf as shown here, or .raw.oga. ld-decode will automatically uncompress the input file during processing.

To compress a .lds file simply use:

ld-compress <filename>.lds

This script will write a .ldf compressed version of the .lds file to the directory it's called from.

Enabling GPU acceleration for ld-compress

ld-compress also supports GPU acceleration via FLACCL. This requires an OpenCL compatible GPU, and and configuring the environment to support this. Instructions:

Check if your system supports openCL:

sudo apt install clinfo
clinfo

If no devices are listed, you may need to install the relevant drivers for your GPU. For example, for modern integrated Intel graphics:

sudo apt-get install intel-opencl-icd

Next, setup mono, as flaccl is a .NET application

sudo apt install mono (or download for your system from https://www.mono-project.com/download/stable/)

Download the latest CUETools release zip (https://github.com/gchudov/cuetools.net/releases/), and extract to /opt/flaccl

Add the relevant line to the mono configuration file, /etc/mono/config on most systems, within <configuration>:

<dllmap os="osx" dll="opencl.dll" target="/System/Library/Frameworks/OpenCL.framework/OpenCL"/>
<dllmap os="linux" dll="opencl.dll" target="libOpenCL.so" />

The target library name may change depending on your environment.

Finally, to compress an .lds file with GPU acceleration, use:

ld-compress -a <filename>.lds

Flaccl does not presently support ogg file containers, so the the output will be with the file extension of .flac.ldf to distinguish from traditionally compressed captures.

The full list of command line options is as follows:

Usage: /usr/local/bin/ld-compress [-c] [-a] [-u] [-v] [-p] [-h] [-l <1-12>] [-g] file(s)

Modes:
-c Compress (default): Takes one or more .lds files and compresses them to .ldf files in the current directory.
-u Uncompress: Takes one or more .ldf/.raw.oga files and uncompresses them to .lds files in the current directory.
-a GPU Acceleration.  Uses OpenCL or CUDA to accelerate encoding. See https://github.com/happycube/ld-decode/wiki/ld-decode-utilities
-v Verify: Returns md5 checksums of the given .ldf/.raw.oga files and their contained .lds files for verification purposes.
Options
-p Progress: displays progress bars - requires pv to be installed.
-h Help: This dialog.
-l Compression level 1 - 12 (1 - 11 for GPU encoding). Default is 11 (10 for GPU). 6 is recommended for faster but fair compression.
-g Use .raw.oga extension instead of .ldf when compressing.

ld-ldf-reader

ld-ldf-reader is a streaming decompressor for .ldf files that can jump to a specific location in the file. It is very lightly modified ffmpeg library sample code and is mostly used by ld-decode to support seamless seeking inside .ldf files, therefore it only outputs via standard output.

NOTE: This program only exists because of an ffmpeg CLI program bug preventing seeking deep into a file, so it may (or may not) go away after the move to Ubuntu 20.04.

usage: ld-ldf-reader infile [seek location] | (consumer)

ld-cut

ld-cut is a utility for cutting samples from raw RF LaserDisc captures (useful to create samples of trouble-areas when issue reporting), and can now also be used to compress .lds files. The utility allows you to seek and specify start and end frames similar to the main ld-decode application.

usage: ld-cut [-h] [-s start] [-l length] [-S seek] [-E end] [-p] [-n]
              infile outfile

Extract a sample area from raw RF laserdisc captures. (Similar to ld-decode,
except it outputs samples)

positional arguments:
  infile                source file
  outfile               destination file (recommended to use .lds or .ldf suffixes)

optional arguments:
  -h, --help            show this help message and exit
  -s start, --start start
                        rough jump to frame n of capture (default is 0)
  -l length, --length length
                        limit length to n frames
  -S seek, --seek seek  seek to frame n of capture
  -E end, --end end     cutting: last frame
  -p, --pal             source is in PAL format
  -n, --ntsc            source is in NTSC format

Using ld-cut, you can do parallel .ldf encodings (optionally targeting different directories) using shell scripting pretty easily:

for i in f1.lds f2.lds f3.lds f4.lds; do (ld-cut $i /someotherdirectory/`basename -s .lds $i`.ldf &); done