Hi,
Last week I updated txt2grd.py. Then got sidetracked. I've attached
the updated version, which I've tested for my needs and it appears to
produce the "right" thing.
However, the manual for the PSR-220 contains two lists of instruments:
The "panel" instruments and the "GM" instruments. From the manual:
> ABOUT PANEL VOICES AND GM VOICES
>
> Keep in mind that the PortaTone has two separate sets of Voices: 100
> Panel Voices and 129 GM (General MIDI) Voices. The GM Voices can also
> be used for optimum playback of GM-compatible song data. This means
> that any GM song data (played from > a sequencer or other MIDI device)
> will sound just as the composer or programmer intended. To change between
> Panel and GM Voices, use the Sub Menu in the Main, Dual, or Split Voice
> modes (see pages 11, 13, 15).
Elsewhere the manual says:
> (100 Panel Voices, 129 GM Voices (128 Voices + 1 Percussion Kit))
Since I'm really quite new to it all, I'm not sure which one to use.
Both? I don't know if there's a way to combine them in a single .rgd
file. (I'm thinking maybe I need to create three banks, "Panel", "GM"
and "Percussion Kit", in a plain-text file and then gzip it....)
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
#
# txt2rgd:
#
# DESCR: Converts from a text based voice list of an instrument/sound
# card to the XML mased *.rgd format understood by Rosegarden
#
# USAGE: txt2rgd <input file> <output file>
#
# where the input file contains the voice list in text format
# and the output file is the name/location of the *.rgd file
# to be generated
#
# The format of the imput file is any reasonable space or comma
# separated (CSV) list of parameters:
#
# a. comments start with '#' and continue to the end of the line
# and may occur anywhere
# b. blank lines (or containing whitespace and/or comments only)
# are ignored
# c. any line containing data must have 5 fields (+ comments,
# ignored):
#
# <user marker> <MSB> <LSB> <Prog#> <Voice> # comment
#
# where:
#
# <user marker> is a field used to help data entry,
# could be anything, e.g. voice number
# but must not contain whitespace, ','
# or '#'. This field is ignored in the output
# <MSB> MSB (Most Significant Byte) of the bank
# number [range: 0 - 127]
# <LSB> LSB (Least Significant Byte) of the bank
# number [range: 0 - 127]
# <Prog#> Program number of the voice witin the
# current bank [range: 1 - 128]
# <Voice> The name of the voice as desired
# it is best to type it in as specified by
# the manufacturer or the GM/XG/GS standard
# if available
#
# MSB and LSB are 0 for the GM only instruments.
#
# Given the notes 2 & 3 some manual intervention on the output
# may be required to achieve best results.
#
# NOTES: 1. This script may use features available only in Python 3.0
# and above
# 2. The device 'id' and 'name' tags are hardwired
# (apparently the id must be 0)
# 3. If possible we try to guess a bank name according to XG
# otherwise the bank name is autogenerated as "<MSB>-<LSB>"
#
# BUGS: 1. The name of the voice can't contain the ',' and '#'
#
# WARNING: 1. The output file will be overwritten if it already exists
#
# VER: 0.4 (Released: 2022.01.02)
#
# AUTH: Ryurick M. Hristev <ryurick.hris...@canterbury.ac.nz>
# Kevin Cole <ubuntour...@hacdc.org>
#
# LICENSE: This program is too trivial to license but if you need one
# pick any from: Public Domain, BSD, LGPL or GPL
# at your convenience.
import sys
import os
import re
import gzip
if len(sys.argv) != 3:
print(f"\n Usage: {sys.argv[0]} <in_text_filename> <out_rdb_filename>")
print(f"\n See the comments inside the script for detailed usage.\n")
sys.exit(1)
in_file = sys.argv[1]
out_file = sys.argv[2]
# markers
line_no = 0
parse_err_no = 0
# will store the structure in a dict of dicts:
# {MSB: {LSB: {Prog#: Voice}}}
data = {}
# each input data line have: <user_field> MSB LSB Prog <Voice_name>
# (after whitespace stripping)
pattern = re.compile(
"^(?P<user>[^\s,]+)\s*(,|\s)\s*"
"(?P<MSB>\d{1,3})\s*(,|\s)\s*"
"(?P<LSB>\d{1,3})\s*(,|\s)\s*"
"(?P<Prog>\d{1,3})\s*(,|\s)\s*"
"(?P<Voice>.+)$"
)
# convert to XML friendly output
converter_table = {
"&": "&",
"<": "<",
">": ">",
'"': """,
"'": "'"
}
# generic bank names { MSB : { LSB : Name } }
# Sources: XG Specifications 1.26, XG Guidebook
# The '(?)' means that the docs do not specify an exact naming convention
# so it's my interpretation.
# Naming is very strongly biased towards XG
# From XG Extra Vol.1 No 6
# - The voices in banks 1 - 8 are essentially similar to their GM equivalent
# in bank 0, with only minor (though often highly effective) modifications.
# - The voices in banks 9 - 15 differ from their GM counterparts mainly by
# having different AEG (Amplifier Envelope Generator) settings.
# - The voices in banks 16 - 23 differ from their GM counterparts mainly by
# having different filter settings.
# - The voices in banks 24 - 31 differ from their GM counterparts mainly by
# having different FEG (Filter Envelope Generator) settings.
# - The voices in banks 32 - 39 differ from their GM counterparts mainly by
# layering two elements and then detuning the elements or changing their
# pitch relative to one another in various ways.
# - The voices in banks 40 - 47 differ from their GM counterparts mainly by
# layering two elements, with the new wave different from the basic one.
# - Banks 48 - 63 are reserved for future use.
# - The voices in banks 64 - 72 differ from their GM counterparts in that
# they use a different wavetable (though they are designed to create a
# similar kind of sound).
# - The voices in banks 96 - 101 often have little or no resemblance to
# their GM counterpart.
# - Banks 112 - 127 may be used by future XG instruments for the
# storage of user-created voices.
bank_names = {
0: {
0: "GM", # XG Spec
1: "KSP", # XG Spec
2: "KSP 2", # XG Guide (?)
3: "Stereo", # XG Spec
4: "Stereo 2", # XG Guide (?)
5: "Stereo 3", # XG Guide (?)
6: "Single Element", # XG Spec/XG Guide
8: "Slow Attack", # XG Spec/XG Guide
9: "Fast Attack", # XG Guide
10: "Long Release", # XG Guide
11: "Short Release", # XG Guide
12: "Fast Decay", # XG Spec
13: "Slow Decay", # XG Guide
14: "Double Attack", # XG Spec
16: "Bright", # XG Spec
17: "Bright 2", # XG Spec (?)
18: "Dark", # XG Spec
19: "Dark 2", # XG Spec (?)
20: "Resonant", # XG Spec
24: "Attack Transient", # XG Spec/XG Guide
25: "Release Transient", # XG Spec
26: "Sweep", # XG Guide
27: "Rezo Sweep", # XG Spec
28: "Muted", # XG Spec
32: "Detune 1", # XG Spec
33: "Detune 2", # XG Spec
34: "Detune 3", # XG Spec
35: "Octave Layered 1", # XG Spec/XG Guide
36: "Octave Layered 2", # XG Spec/XG Guide
37: "Fifth Layered 1", # XG Spec/XG Guide
38: "Fifth Layered 2", # XG Spec/XG Guide
39: "Bend Up/Down", # XG Spec/XG Guide
40: "Tutti", # XG Spec
41: "Tutti 2", # XG Spec (?)
42: "Tutti 3", # XG Spec (?)
43: "Velocity Switch", # XG Spec
43: "Velocity Switch 2", # XG Guide (?)
45: "Velocity Crossfade", # XG Guide
46: "Velocity Crossfade 2", # XG Guide (?)
64: "Other Wave",
}, # XG Spec
64: {0: "SFX"}, # XG Spec
127: {0: "XG Rhythm Kits"}, # XG Guide (?)
}
# Python Cookbook pg. 88
def multiple_replace(dict, text):
regex = re.compile("|".join(map(re.escape, list(dict.keys()))))
return regex.sub(lambda match: dict[match.group(0)], text)
IN_FH = open(in_file, "r")
for line in IN_FH.readlines():
line_no += 1
line = line.split("#")[0]
line = line.strip()
if len(line) == 0:
continue
matched_line = pattern.match(line)
# if there is no match then no group is generated (all or none)
try:
user = matched_line.group("user")
MSB = int(matched_line.group("MSB"))
LSB = int(matched_line.group("LSB"))
Prog = int(matched_line.group("Prog"))
Voice = matched_line.group("Voice")
except AttributeError:
print(f"\nInput error at line {line_no}, read: {line}")
print(f"Skipping line ... continuing parsing...")
parse_err_no += 1
continue
# make it XML friendly
Voice = multiple_replace(converter_table, Voice)
# Range check on MSB, LSB and Prog#
if MSB < 0 or MSB > 127:
print(f"\nInput error at line {line_no}, read: {line}")
print("Second parameter (MSB) is out of range, "
"must be between 0 and 127.")
print("Skipping line ... continuing parsing...")
parse_err_no += 1
continue
if LSB < 0 or LSB > 127:
print(f"\nInput error at line {line_no}, read: {line}")
print("Third parameter (LSB) is out of range, "
"must be between 0 and 127.")
print("Skipping line ... continuing parsing...")
parse_err_no += 1
continue
if Prog < 1 or Prog > 128:
print(f"\nInput error at line {line_no}, read: {line}")
print("Fourth parameter is out of range, "
"must be between 1 and 128.")
print("Skipping line ... continuing parsing...")
parse_err_no += 1
continue
# Some confusion here: the program change number starts at 1
# while the program id starts at 0; substract 1
Prog -= 1
# create embedded dicts as required
try:
data[MSB][LSB][Prog] = Voice
except KeyError:
try:
data[MSB][LSB] = {Prog: Voice}
except KeyError:
data[MSB] = {LSB: {Prog: Voice}}
IN_FH.close()
if parse_err_no != 0:
print(f"\nThere were {parse_err_no} error(s) in the input file.")
print("Please correct the error(s) before proceeding forward.")
print("No output was generated.\n")
sys.exit(1)
with gzip.open(out_file, "wb") as OUT_FH:
OUT_FH.write("""<?xml version="1.0" encoding="UTF-8"?>\n""".encode())
OUT_FH.write("""<!DOCTYPE rosegarden-data>\n""".encode())
OUT_FH.write("""<rosegarden-data>\n""".encode())
OUT_FH.write("""<studio thrufilter="0" recordfilter="0">\n\n""".encode())
# WARNING: id here must be 0!
OUT_FH.write("""<device id="0" name="Unnamed" type="midi">\n\n""".encode())
# placeholder
OUT_FH.write(""" <librarian name="Unknown" email="unknown"/>\n\n""".encode())
# Add instruments, required ('midi' type instruments start at 2000)
# *** These should no longer be necessary
# for i in range(16) :
# OUT_FH.write(f""" <instrument id="{2000 + i}" """
# f"""channel="{i}" type="midi">\n""".encode())
# OUT_FH.write(f""" <pan value="64"/>\n""".encode())
# OUT_FH.write(f""" <volume value="100"/>\n""".encode())
# OUT_FH.write(f""" <reverb value="0"/>\n""".encode())
# OUT_FH.write(f""" <chorus value="0"/>\n""".encode())
# OUT_FH.write(f""" <filter value="127"/>\n""".encode())
# OUT_FH.write(f""" <resonance value="0"/>\n""".encode())
# OUT_FH.write(f""" <attack value="0"/>\n""".encode())
# OUT_FH.write(f""" <release value="0"/>\n""".encode())
# OUT_FH.write(f""" </instrument>\n\n""".encode())
for MSB in list(data.keys()):
for LSB in list(data[MSB].keys()):
# dig out a name if possible, else default to bank name as MSB-LSB
if MSB in bank_names and LSB in bank_names[MSB]:
name = bank_names[MSB][LSB]
else:
name = f"{MSB:03u}-{LSB:03u}"
OUT_FH.write(f""" <bank name="{name}" msb="{MSB}" lsb="{LSB}">\n"""
.encode())
progs = list(data[MSB][LSB].keys())
progs.sort()
for prog in progs:
OUT_FH.write(f""" <program id="{prog}" """
f"""name="{data[MSB][LSB][prog]}"/>\n"""
.encode())
OUT_FH.write(" </bank>\n\n".encode())
OUT_FH.write("</device>\n".encode())
OUT_FH.write("</studio>\n".encode())
OUT_FH.write("</rosegarden-data>\n".encode())
_______________________________________________
Rosegarden-user mailing list
Rosegarden-user@lists.sourceforge.net - use the link below to unsubscribe
https://lists.sourceforge.net/lists/listinfo/rosegarden-user
