neobustatunez ([info]neobustatunez) wrote,
@ 2007-08-09 13:36:00
Previous Entry  Add to memories!  Tell a Friend  Next Entry
Current mood: relieved
Current music:Sevendust - "Waffle"

Delayed Reaction
So, as previously mentioned, I made a sample library. What I haven't mentioned yet was that, eventually it grew and grew. And grew.

And now it's been commercially released - www.impactsteelsounds.com



This is especially exciting for me b/c commercial release = $$$$, and while I won't divulge any figures I will say that it's been very much worth the while. So far it's been getting some very positive reviews, and some pretty big heavy hitters in the game and television music industries have been purchasing it lately.

Course being that this is now a commercial release, I can't go into detail about how I did it. Which I guess is a bit of a shame, but for those somewhat interested in building your own Kontakt libraries, I'll outline the steps. It's actually way easier than you think it is:

1) Identify and secure your sound source. What is your library for? If you're literally recording something, you'll need mics, a room, and perhaps a mixer. If you're doing something abstract with computer-generated sounds (like synth pads, textures, etc.), skip to step 3 and just make sure that, if selling commercially, you're legally permitted to use the source material this way. It's a good idea, if performing an instrument of some kind, to spend a lot of time getting to know the ins and outs of it, and figuring out exactly the types of things that'll be actually useful for you to record (step 3). And obviously if you're recording other people, you'll need to schedule them; recording an orchestral session is possible but you'll need to find a contractor who's willing to subject his musicians to that kind of torture.

2) Equipment and performance space. You'll need to either have some kind of a recording space (walk-in closets can work well if there's space for some kind of padding, like clothes or even bookshelves) or rent out a studio. Renting is obviously the more professional approach, but you'd better do some prototype recording beforehand and have a very good idea of what you want, how many takes, and how long it's gonna take. For reference, my ~350 recorded samples from Impact: Steel took about 3 days to record, and I didn't dawdle.

As for mic choice, that's a can of worms I'd rather not open. But you can find some great deals on zZounds.com - for low cost, near-professional quality gear, some names to look out for are MXL and Behringer. MXL makes a great condenser mic pair (990+991) for only $100, and Behringer has a nice mixer for about $60 last I checked. You'll need stands and cables - there's a cheap $19 package that includes a stand, an XLR (microphone) cable, and a crappy dynamic mic. And finally, you'll probably need a mess of cables and adapters to connect the mixer to your computer. If you're getting a higher end mixer, it'll probably connect via firewire or USB; the $60 Behringer model will need a Y cable (dual mono to stereo), a stereo 1/4" cable, and a 1/4" to 1/8" adapter to connect to most sound cards' line-in ports. Even if you're gonna be using a studio for the final recording, I'd recommend having some kind of semi-professional stereo home recording setup to make prototype samples, see what works and what doesn't, and iron out kinks in your workflow.

3) Cue sheet. Invaluable! Even for small projects! Make one! My sheet for Impact: Steel was in simple outline format, listing the instrument, the articulation (and a brief description of the best way to play it), the number of velocity levels (ie. loudness levels) and the number of variations / alternate takes. Lots of velocity levels will give you a very natural response to dynamics, and alternate "round robin" variations will give a natural and "organic" feel, without the repetitive sound you'd normally associate with MIDI. Make sure to print out the cue sheet and take it with you to the recording booth/studio.

4) Recording. You're on your own here, there's too much to talk about and the bottom line is to test and test and test again until you get a mic position, recording levels, and distance that you think is right. All I'll say is that with instruments with large dynamic range, you'll probably need to adjust gain levels to avoid clipping. Remember to record a noise print (~10 seconds of "silence") so that you can do a noise reduction in the editing phase.

Try to record as much as you can (within reason) in one shot so that you have a consistent sound throughout. If at all possible, get an assistant to help monitor your levels and prevent clipping. I recommend starting quiet and getting louder, and doing all the variations at that particular velocity at once (in other words, 1 1 1 2 2 2 3 3 3 4 4 4 etc.)

5) Editing. Fun fun fun. Oh sorry, I meant "misery". But you have to do it, so, whatev. Here's my suggestion - Start off with the noise reduction tool of your choice - Goldwave (Free-ish!) has a good noise reduction algorithm that works with noise prints. After that, you'll be cutting the files out into separate articulations (assuming you recorded a lot in one shot). Then separate out the round robin variations, deleting unwanted takes. The result would be a file with 7-9 (or however many velocity layers you planned for) hits in increasing loudness. Run this through an audio splitter like WaveKnife, trim and normalize the resulting files, and you'll have something you're ready to throw into Kontakt for prototyping (the fun part!) ((Seriously this time!!)).

6) Kontakt, part 1. Throw your samples into a new instrument by finding them on the left browser window and dragging them all into a new instrument's mapping editor. You can right click the editor first and play around with some of the mapping features (like snapping to white keys only). For unpitched percussion banks, you might wanna consider going to the group options and unchecking pitch-"Tracking".

In my experience, this was actually the most important step - finally getting a chance to test out the patch just as the end user would. The flaws and shortcomings you might not have noticed are now painfully obvious, whereas some of the problems you might've been worried about may turn out not to be an issue at all. Play around with ADHSR envelopes, EQ / filtering, different reverb settings (within and outside of kontakt), even dynamic filtering. When you come up with a tone and ergonomic organization that you like, you basically have two options: Either call it a day and save it as is, or do some offline postprocessing.

7) Postprocessing. Some may call it cheating, but personally I find it to be the smartest step I took when developing IS. Basically offline post just means looking at those edits (EQ, filter, reverb, etc.) you made in Kontakt and applying them directly to the samples themselves offline. You can do this with programs like Goldwave or Wavelab that allow you to batch process .wav samples using VST's. The process itself is very easy, you just go to the program's batch processor, drag in all the samples, and set up your processing chain - just remember to turn down the volume a couple dB as the *first* step in the chain in order to prevent clipping later on (you can normalize it again later).

There's both advantages and disadvantages to processing the .wav samples themselves, though. Lots of people will complain that this can result in an unnatural and unrealistically "hyped" sound, and that it prevents the end user from making their own edits in realtime. For example, if you put a reverb right on the samples, there's no way to take it off; whereas if you did it in realtime in Kontakt, a user can just bypass it. In my case though, I wasn't recording pristine orchestral samples and so the surgically-precise "natural" sound wasn't really sought after anyways. The advantages include not only conserved CPU, but also it allows users to drag the .wav samples right into their sequencers by themselves and do whatever they want to them (believe it or not, some people still work that way).

8) Kontakt, part 2. This is just the final step really, where you're renaming things, playing with the now-postprocessed samples, making final tweaks and saving final versions. If you're doing stuff you'd like to send to other people, it's a good idea to figure out a logical place to store the samples here, and save the NKI's to reflect that. For commercial work, you'll wanna send it around and get lots of feedback from other users; one of my best experiences was literally watching a friend of mine use IS for the first time, and observing his tendencies and expectations. That helped me to adjust the way the instruments were mapped, so that I could have the most ergonomic possible experience for keyboard users... which perhaps is a concept that hasn't really been explored very much, yet.

9) Conversions. If releasing commercially, it's not a bad idea to try and convert to several other formats, such as Giga, EXS, Halion, etc. Programs like Chicken Translator and Extreme Sample Converter can do the trick, but you may need to convert down to Kontakt 1 for those to work - I ended up having to build the library in KT1 simultaneously while working on the KT2 versions since none of those programs do very well with K2 patches.



Well there you have it. Developing IS was a great experience for me (lucrative too!), and it didn't really require very much in the way of equipment. All you need is a microphone, an idea, and a heck of a lotta spare time. And the payoff is that you now have an original sound that you can absolutely abuse, and get away with it because it's your own.


Advertisement


(No comments)

Post a comment in response:

From:
Help
Identity URL: 
Username:
Password:
Don't have an account? Create one now.
Subject:
No HTML allowed in subject
   Help
Message:

 
Create an Account
Forgot your login or password?
Login w/ OpenID
English • Español • Deutsch • Русский…