How to refurbish Amiga tracker and chip tunes

In the following article I describe a process for refurbishing old Amiga tunes. This is a follow-up to http://blog.airmann.de/uade-multichannel-audio-support.

First, what’s the motivation behind ? Why refurbish Amiga tunes at all ?

Answer: original tunes often are poorly mixed and are not optimized for playback on modern audio equipment. So the goal is to polish those gems soundwise without altering the tunes itself, and finally transfer them into proper formats for playback on nowadays audio equipment.

So why should the original mixes be not optimal ? What’s the matter ?

  • technical limitations: most ancient trackers didn‘t provide EQ‘s, filters, stereo balancing, compressors, delay, reverb, etc.. Thus, the mixing results were pretty limited
  • missing original author‘s audio mixing skills / awareness: E.g. no proper stereo balancing, frequency balancing and volume leveling. E.g. bass and drums tracks were panned hard to left/right, frequency ranges of instruments overlap, etc..
  • A lot of ancient samples and waveforms contained DC offset which negatively affects bass frequency range

Ok, so how does the refurbishing process look in detail ? Answer: for more information read the following script in PDF format:

How_To_Refurbish_Amiga_Tracker_Chiptunes

For the lazy visual guys:

Overall Refurbishing Process
Overall Refurbishing Process
Quad Mode Rendering Part
Quad Mode Rendering Part
Postprocessing Overview
Postprocessing Overview

This entry was posted in Amiga, Production, Tech Stuff. Bookmark the permalink.

10 Responses to How to refurbish Amiga tracker and chip tunes

  1. John says:

    Hello Matthias

    Thank you for making your quad version of uade available and for your guide on refurbishing Amiga chip music. It has been very useful indeed.

    I am using the quad version of uade and Audacity to mix and refurbish my favourite Amiga game music. I have found a problem with the quad channel output and the Alien Breed music (http://www.exotica.org.uk/wiki/Alien_Breed). When I convert the module to a four channel wave the music is cut off at 4:16, when it is converted to a two channel wave it is not cut and the music is output in its entirity, this effects all the modules over 4:16 in length. Example commands I used:

    uade123 –quadmode –filter=NONE –force-led=0 –frequency=48000 –resampler=SINC -f ALIENMENU.wav bp.ALIENMENU
    uade123 –filter=NONE –force-led=0 –frequency=48000 –resampler=SINC –panning=0.0 -f ALIENMENU.wav bp.ALIENMENU

    I’m not sure if the problem is with uade or with the modules. I know a bit of programming so I could look at the source code, but I imagine the problem will be quite obscure.

    Best wishes

    John

    • 4irmann says:

      Hey John,
      thanks for feedback. Glad to hear it’s usefull. Indeed a strange behaviour. Have you tried to disable resampler and panning switches ? Resampling could theoretically be done later on using e.g. SoX which has a world class resampling algorithm. Anyway, I’ll have a look at it ASAP.

      Regards
      Matthias

      • John says:

        Hi Matthias

        Just tried it without the resampler:

        uade123 –quadmode -f ALIENMENU.wav bp.ALIENMENU

        Still chops off the end.

        John

        • 4irmann says:

          Hi John,
          I’ll have a look at it tomorrow. Nonetheless, another idea: how big are the resulting files in bytes and is there enough disk space ?

  2. John says:

    Hi Matthias

    There’s no hurry 🙂

    File sizes:

    uade123 –quadmode -f ALIENMENU.wav bp.ALIENMENU : 86.1 MiB
    uade123 -f ALIENMENU.wav bp.ALIENMENU : 47.5 MiB

    I have 87.0 GiB free space and am running Linux Mint 17.3 KDE 64bit. Uade is compiled from source.

    John

    • John says:

      The med.intro track from the TFX theme also cuts out at 4:16 in quad channel mode.

      It then ocurred to me that 4:16 mins is 256 seconds. I think this could be a clue, since 2^8 is 256.

      • John says:

        Had the idea whilst I was trying to sleep, hence my bad spelling!

        • 4irmann says:

          ok, I’ve analysed the problem. It’s a bug that should be fixed, now.
          I’ll release an updated tar.gz file as soon as possible.
          Meanwhile you can use this additional command line switch as a workaround: –subsong-timeout=-1

          • John says:

            Brilliant! Thanks for fixing the problem 🙂

            I spent some time yesterday looking at the code, but I couldn’t see much difference between the two channel and four channel code paths.

      • 4irmann says:

        That’s a worthwhile hint. I’ll check it out. Unfortunately I didn’t find the time today. I’ll do it ASAP.

Leave a Reply

Your email address will not be published.