Converting OctamedSS to MP3

[Karlos]

Hello fellow Amigans,

I was wondering if anyone knows how to convert Octamed Soundstudio songs to MP3 which will include MIDI as part of the songs. I've managed to convert my protracker Mods by importing them into Madtracker. But since I sold my Yamaha PSR, I've not been able to get the MIDI running in sync when I use standard GM Midi sounds from the computer instead of my Yamaha GM sounds. It would be nice to get this all in sync and then convert it from my PC to into MP3. Unfortanatly, my real Amiga is in England in the garage! and all I have to work with is the data from my Emulated Octamed Soundstudio on the PC.
Any tips, greatly appreciated.

You'll have to do this in two stages, probably. In the first stage, use OMSS's direct to disk mixing mode as a 16-bit stereo, 44kHz (or 48kHz depending on your preferences).

For the second stage, use OMSS to control your MIDI device and use a separate tool to record the output, say on your PC. Make sure you use the same recording rate as you used for the first step, or mixing becomes complicated.

Finally, you will need to mix the two tracks together, before converting them to mp3. There are plenty of tools for this, LAME always gives decent results.

There is a knack to doing this well. What I would recommend is inserting an empty block at the start of your song. Then, on the first line, add a short sample and likewise a short MIDI sound, such as a drum or something else with a very rapid attack. These will indicate the start of your audio.

This step will help you synchronize the start of the two files, since when you capture the live MIDI performance, you'll probably already have started your recording tool for the audio and will have an otherwise difficult to estimate range of silence before it really starts. Before mixing, you can crop the second file to the start of your marker sound and then mix them both together confident that they align up properly.

Once you've mixed them properly, you can cut off the synchronising tick and silence that follows until the start of your track proper.


Another way of doing this is to just use a mixer to mix your live OMSS sample replay and the MIDI device's output and then capture the whole lot.

I prefer the former since it gives me more options for post-production during the final mixdown. Sometimes, I record several components of a track in isolation, purely for this reason.

[Karlos]

Thanks for the advice I'll probably do as you said and mix down the midi and sample based sounds separately. Or even do away with the midi by converting them into samples and then record the whole lot as MP3 or Ogg. Thanks again for the useful tips.

You might want to keep your MIDI. Converting the MIDI instruments to samples can require a lot of work. I've done it in the past and preserving the feel of your MIDI hardware is not that easy. For example, if you have a nice instrument patch, without using at least multiple-octave samples, it just isn't going to sound anywhere as good, and that's before you get into the sample replay limitations of OMSS. You'll want to use 8-bit samples unless you don't need fine control over note volume, since 16-bit sample volumes are implemented using shifts (that is to say, you can have full volume, half, quarter, eighth and so on). Very messy.

Furthermore, any effects controllers you've set up, such as reverb, chorus etc. just aren't going to be easily reproduced either.

[Karlos]

KarlOS: The last time I messed with hard drive recording, the end result sounded nothing at all like the song I had just composed. Just noise in fact, and not the noise I was going for.

You think 16-bit samples might have caused this?

That depends. Did your record in 16-bit mode? If so, 16-bit samples should replay fine, aside from the crude volume control. If you recorded in 8-bit mode, well, that might not end so well.

Also, what did you use to replay the rendered audio?

[Karlos]

Did you record with smoothing on?

All this 16 bit and 44KHz mixing makes for better quality recordings but remove the "Amiga-ness" of the sound.  If that was part of the original you may have inadvertently removed it.

Another reason I often record parts separately and mix later. If I am trying to use decent quality sounds, I will turn on the smoothing, otherwise I'll leave it off.

However, even without smoothing, you aren't getting so much the "amiga" sound as the heavily aliased-sample conversion sound. If you want actual amiga sound, use 4 channel mode and capture the audio externally.

[Karlos]

BTW I believe there was a discussion on the non-lineararities of the Amiga sound O/P at one point.  Is this on line anywhere?

I dunno where exactly, but I do remember talking about this with bloodline and others not that long ago.

Not only are the DACs quite non-linear, but they vary from machine to machine. This is why the 14-bit CyberSound calibration utility exists.