So last may i posted on EAB forum proposing myself as MOD musician. Got two gigs: one from the gut who create downfall for its dizzy-like game and little later was asked by Philippe “Meynaf” Guichardon to port the music of Heroes of Might and Magic II in MOD files for its Amiga Port of the game.
And is on the second task that i want to focus on this article and possibly some future ones.
Tough nut to crack! The soundtrack is composed of over 20 tracks, the music is done on pc either via Midi with custom instruments or via CD tracks, in some cases in the “price of destiny” expansion pack there are singed parts (in german) so i know will never be the same as the original PC one, but who am i to defuse an impossible challenge?
Looking at the context and some of the music, plus the fact that the porter do not want to mix channels (processor intensive), i realized i can build some of the mods in a way that three channels carry the main melody and the fourth is of support and can be overridden by speech or sound effects.
As reference i took this youtube playlist of the CD tracks, and that is mostly the kind of feeling I want the mod files to have.
As mentioned in other articles, i actually have no real Amiga handy where to compose the score, however my actual favourite weapon, MilkyTracker on Win XP, has been put to the test for writing .mod files and feedback has been positive.
Another thing is that since there are some songs (i.e.combat 1) that have a recurring background, i could sample it and use it on the mod file. the “Combat” scream on combat 1 could be replaced with a generic crowd cheer (whoah) .
Before to leave i started to port lurking enemy (the enemy turn song) and had some problems both for its basses and for the motion of the song; i repromised to give it a second look or to rewrite as mixed version.
The day earlier instead i tried to port the grassland theme; so all possible problems with Milky showed up – which from the other side is good so i have an idea of the workflow i could employ; first the four version sound kinda flat (midi can rely on being projected in two channels) plus the lack of faithful instrument brought me to look on my mod bank for similar sounds. Redo a tune is harder than come out with your own: especially in a limited environment as Paula might be, plus memory contraints and the fourth voice as wildcard. But that is what makes it so compelling: being able to reach good results in a limited environment :)
I tried several combinations of instruments for Grassland, mostly coming from digging through instrument free data banks; a decent cello in the background is hard to find and for clavicembalo sounds I resorted taking it from a tune of The Weasel, however the high tune clavicembalo sound kinda weird;
That made me look for better clavicembalo (harpsichord) samples; the ones i found comes from collections around the net: very good this Harpsichord sample bank from a 1720 Blanchet and this Oboe sound from freeloops.com; those two instruments helped me in give a more realistic sound;
Result is pretty impressive: might not have all the sounds of the original but has its own depth :)
The research for more realistic sounds brought me quite around on the net, looking mostly for .wav files until i started stumble in more professional sites that provided soundfonts; this format is used for most professional tools and usually provide a good instrument quality, however is not directly “consumable” by milky or other trackers: a tool to export in .wav file format is needed.
Looking more around i found an open source .sf2 to wav converter for Windows at this address: http://sanfransys.com/main_downloads.htm; what it does is to split all audio files composing the soundfont and exporting them in .wav file on a folder.
Then the sample is passed through Audacity, where is flattened to Mono and saved in .wav 8-bit (save -> other uncompressed format -> .Wav (microsoft) 8-bit) then loaded in milky and “resampled” to match the protracker range of sound going in the Sample Editor [smp ed.] and choosing ‘advanced->resample’ from the right click menu.
Once in the option requester, setr the “relative tone” and “fine tune” values to 0, select “Precise sinc” on the Interpolation dropdown option and click OK.
At this point i usually save the instrument in .xi format from the [save] button on the top of the instruments list. Range used by protracker mods is between C-3 to B-5 according to milky scale. If higher or lower octaves are needed i usually load a copy of the instrument in another slot, go in the instrument editor [Ins. Ed.], press octave up or octave down at need then go in the sample editor and again use resample putting to 0 the values to render the sample.
Plus since, unlike the good old protracker, there is no chord and no merge samples feature, sometimes i need to “render” in .wav using an empty mod the chords and composed samples and do on the rendered track the flat/8-bit/resample loop again.