Hi all.

I used to use Madtracker back in the day, but I eventually fell out of it because I could not produce the sound I was looking for. I primarily relied on samples because, at least as I recall, Madtracker doesn't really provide an interface for making instruments from scratch.

The other day, I was thinking of using Madtracker again. Mainly, I was inspired by a song I heard, and I was inspired to create something equally moving (an idea that never really manifests in reality), but instead of using Madtracker, I looked at what else was out there. What I found was Sunvox. I found the interface to be simpler, while at the same time providing much deeper control of instrument sounds.

Anyways, I always find that the songs I produce are lacking in some way. For example, the image below. The waveform of the commercial song (on the bottom) is optimized in some complex way that I do not yet understand. I guess my weakest point is sound engineering. By that I [perhaps fallaciously] mean the creation of instruments that sound full and unique, and that, when used together in a song, make full use of the audible spectrum. There are a lot of sounds in music (mostly ambient-type sounds) that I can pick out but have no idea how they are produced. I am a programmer with experience in C++, OpenGL, and Matlab, so I can produce any sound whose spectrum I can think up, so maybe the problem is that I just have not adequately experimented with the resources at my disposal.

Another skill that is currently over my head is eliminating certain artifacts from instrument sounds. For example, the kick in the bass drum sounds good in Sunvox (good for a prepackaged free drum sound), but when I export the song to .wav, the drum also appears to make some other high pitched sounds that need to be eliminated. I can export the drum track separately, but do not know what way is best to isolate (I could not see the artifact in the spectral graph) and eliminate the artifactual tones. I found a very informative explanation of a solution here, but I am not sure of how to implement this and other techniques of sound engineering using only the freeware at my disposal: Sunvox and Audacity (+ Matlab). At least, I have tried to do so in eliminating the artifact in my kick and have failed to do so.

I am realizing that much of what makes a song appealing to me is heard and processed subconsciously. When I apply all of the consciously identified parts in my music, I find that it still sounds somewhat empty compared to the music that I am inspired by (or inadvertently attempting to imitate ). I guess what I am hoping for by posting this here is criticism and advice on how to improve the sound of the instruments I am using. I'm sure there are rules of thumb that I have not yet picked up.

Thank you very much.
Hixidom

My song

EDIT: An idea I had that I think would be a great learning experience for me would be to pick a song that I love and try to replicate it exactly. In doing this, I will have the comparison of the original song and my cover as a direct representation of where I need improvement.

[The images I attach to posts never seem to show up]

hey mate i like ur song u seem to have a good sense of melody which i think is most important, but indeed production wise could use work. you say ur not happy with quality of ur sounds, well these are the kind of sounds sunvox are producing. myb u should move on to something else: cubase, protools, logic, ableton.. are all great choices. Pick one and learn it as well as you can and get into some vsts- fx and instruments. You will see a new world of sound open up. Also u mentioned mixing, well i find good to try to get best source sound u can, and then setting levels and eq to get everything nice n balanced.. keep it up

i am not a musican , however most of my friends are and they speak highly of nuendo/cubase for creating some quality sound , also logic and ableton works to smooth some processes

basicly you can creaste your track in nuendo/cubase and then export it to ableton and vice versa and between other softwares to exactly get what you want

I may be too cheap for that route. Besides, if I can fully understand what makes certain constructed instruments sound better than others, then I can make my own program to produce them. I'm tempted to make a tracker from scratch, or at least some sort of fx/synth module that can work with a tracker like sunvox. It's not that I want full control. More so I want a full understanding of what's going on under the hood of these programs. If I can attain that, then I don't see why I can't produce the same sound in something like Sunvox or a program of my own, albeit with a little more time and effort. I could at least make a program that produces instruments: Some sort of instrument laboratory. I could then output instrument samples and use them in Sunvox.

It is also possible that something about the way Sunvox exports the tracker content to WAV is inferior and thus results in inferior sound quality. Though I can't comprehend what could go wrong in writing samples to a file. I'll see if there is any way to increase the bitrate of Sunvox, or if I can get the source code and do it myself.

My view is that a free program can be (and, in some cases, is) as good as a commercial product used to perform the same task (Blender comes to mind). If Sunvox is inferior, it is only because not enough time has been devoted to developing it. Perhaps I should see if I can help to do this. Then again, Sunvox is purposefully minimal, so its current simplicity may be intentional.