N-Track Recordings

Some early attempts with N-Track.



Created using a combination of live and midi instruments. I plan on going back and replacing the midi drums with the real thing at some point.

Have you tried NSKit? Until you get real drums recorded, NSKit is about the next best thing in the MIDI realm.

Thanks - I’ll check into that. I was planning on using Fruity Loops for the drum tracks (it uses wav samples of drums rather than midi generated). I do have a Yamaha electronic drumki, it’s just a matter of sitting down and getting it down.

Quote (robg @ Sep. 15 2004,08:54)
Thanks - I'll check into that. I was planning on using Fruity Loops for the drum tracks (it uses wav samples of drums rather than midi generated). I do have a Yamaha electronic drumki, it's just a matter of sitting down and getting it down.

This tells me you don't know how MIDI works. All MIDI does is trigger sounds. If MIDI is triggering crappy sounds, it will sound crappy. The NSKit comes in several formats and all of them are real samples (AKA, real recorded drums). Man I wish the board wasn't hacked, I have only gone through this 100 times before. That is no fault of yours though, I'm just thinking out loud. Anywho, MIDI is only as good as the sounds it triggers. What you are hearing ar the sounds that probably come with Windows and they aren't great at all. Check out this link to read a lot of posts I have made on the ins and outs of all of this. You will need to sign up as a member of Audiominds, but you can be assured you won't get any spam or other nasties. (I think I am tired and cranky this morning. Sorry if I sound grouchy. :) )

EDIT: Never mind following the link, here is my post from AudioMinds:

OKay, here we go. Take notes because when someone asks this exact same question in a week (like they always do), they're getting referred to this thread.

SFZ vs. the world: sfz is an excellent basic SF player. It takes what ever SoundFont you throw at it and plays it back accurately and without any weirdness. It allows you to choose how memory on your machine is mananged too. So if you load a 250MB SF, but only have 256MB of RAM, there are options to make the SF play better with your machine, be it loading all samples from disk on the fly or a few other options.

The rest of the world in SF players offer things like effects, LFOs, cut off and reasonance; all things to take your SF of a trumpet and make it sound totally different. So when choosing an SF player, you need to decide if you want these extra sound design options. If all you want is to play back the SF exactly as the author intended, SFZ should be adequate.

So we come to, why SoundFonts and not just a pile of wavs?

Example 1: The NSKit drum SoundFont consists of 321 different samples. Do you really want to have to manage that many samples on their own? Or wouldn't it be more convenient if they were all prepackaged for your use? Sure it would.

Then you ask, well, I just need 4 samples... a kick, a high hat, a snare, and a ride, what in the world am I going to do with 321 samples? The question I have, does a snare sound the same when you hit it lightly vs. when you hit it as hard as you can? How about a kick drum? A ride? Doesn't a ride get "trashier" the harder you hit it? Well, if you only have one sample of a ride hit as hard as it can be hit, how is that going to ever sound like a ride being hit softly? There is more to the sound then being quieter when hit softer, its basic character is different. So SoundFonts and many other sample formats use what is called velocity layering. Depending on the velocity of the MIDI note that triggers the sample, a different sample will be played back. Sequence a snare hit with a velocity of 127 (the highest value you can have in General MIDI) NSKit will play a sample of a snare being hit very hard. Sequence a snare hit with a velocity of 25, NSKit will play a completely different sample that is just a tap. With out SoundFonts you would have to manually sit down with all 321 samples, program in what notes trigger what samples, as well as what velocities trigger each sample. PLus, what if the mic was just slightly closer to the snare when recording the tap vs. the hard hit? They may have the same volume. You obviously want the tap to be softer in volume in comparison. You can program these relationships into the SF. The SF format allows folks to package this for you, so they do all the hard work, you just play your music.

Example 2: SoundFonts are much more than just drums sounds. They can be anything from sound effects of cars starting to a symphony string section. The question is, if I have a wav file of a symphony, how in the world is that going to translate into me playing my MIDI sequence back to sound like a real orchestra? SoundFonts also do pitch shifting. If you have a SF of a trumpet, there may only be 5 samples in there, but somehow you are able to play a trumpet over 5 octaves. This is good from a space efficiency and recording time standpoint. In addition, you have a real trumpet you can play with. With a single wav file or wav files, you can't easily play the trumpet. You would spend weeks moving wavs around to make realistic sounding track. Imagine moving 20 trumpet samples around to play a simple melody in N "Okay, now I put in the sample playing A. Oops, have to cut it short, it's only a quarter note. Now the sample playing C#. Crap, the sample isn't long enough to hold the note..." Hold the note? SoundFonts use a clever technique called looping that loops the sample over and over so your trumpet can hold a note for 20 minutes and never run out of breath. A well programmed SF will play the initial attack of a sound and then when it reaches a certain point within the sound loops back to an earlier point after the initial attack and continues looping until the MIDI note is released. You can't easily do that with a wav file on its own.

I should mention, the SoundFont format is juts one of many sample formats out there. There are many others such as GigaSample, Halion, Kontakt, EXS24, Akai, Reason Refills, Emu Emulator, SampleTank, Mach 5, and a bunch more. Some of these formats allow gigantic files. This is what pros use to do all those orchestral pieces for TVs and movies among other things. Those multi gigabyte files have thousands of samples in them. You could never expect anyone to manage that many samples and still be a productive musician.

So hopefully this gives you a better understanding of SoundFonts and sampling in general and how some folks can have such realistic MIDI pieces and why some sound so cheesy. Half the quality of the MIDI is in the sequencing itself, the other is in the quality of the sample set (number or samples, velocity layers, etc.) MIDI through your 2MB SoundFont on a Live! card will almost always sound like cheese. Through quality, well programmed SFs and sample libraries, they can sound real.
MIDI is only as good as the sounds it

I understand that - that's why in more recent tunes I've imported the midi file into FL and used it to trigger wav samples, rather than use the crappy midi in my laptop.