Opinions on how to build the most genre-versatile acoustic d

I would like to ask for a few opinions on how you guys would go about creating a really, really detailed acoustic drum sample library from scratch using BFD as the sound source.

I have a tremendous amount of time on my hands...so I have undertaken a tremendous project. I've figured out a way to fairly quickly (not quick enough, but I'm getting there) get very detailed drum programs out of BFD and into my MPC4k/Z8. I've got a few kits in 32 layers of velocity sensitivity through various midi channels/programs/sequences, in the preset mic position/mixing, with then, in the default mic position, a program with just the mics solod (so one for Direct, Overhead, Ambient and Room) on a drum by drum basis for most of the Andy Johns expansion kit. Even with both a Z8 and a 4k, you can only load up one of the mic channels at a time at 32 levels, I have a VS1880, so I can't even really record all the mic channels on there and then mix that way. The closest I've been able to get is to load up smaller velocities of the channels, get your levels how you want (as a mic mix, almost) in the sampler, then record on a regular drum basis (ie 1 channel kick, 2 snare, 3/4 overheads, or however you want, really) into my recorder. I've began making programs inside each sample folder for 4, 8, 12, 16 and 32 sample velocities. I started a thread this morning about how to assign stuff faster. This is why.

That's kind of what i'm looking to do. Recreate the techniques of recording real drums inside my samplers, and I'm trying to get as detailed as possible.

And straight up...I know how useless all this is. This is more just to see how deep I can take the machines for my own dorkery, nothing I think would be practicly easy to use.

So, as I get through the andy johns expansion, I'm realizing the walls of using this stuff, which is really what I wanted to know. What I'm really after is making a very, very complete, pro, pushing my gear to it's full potential drum library. I realize the amount of work I've got ahead of me, and I want to choose what I do wisely, so I have some questions I want you guy's opinions on.

Would you consider the direction I'm going? I have the BFD2 library and the Jazz expansion. Once I get those, I'm going to pick up another one, continue, etc etc. Would it be better to continue sampling these drums on a drum-by-drum basis so you can make your own kits inside the mpc? Or would my time be better spent making entire kits inside the BFD environment, and sampling on a program by program basis? Should I make kits with variation folders inside, built on templets? Make variations for a "tight kit", "roomy kit", etc etc? I think I'm going to keep doing the channel solo idea. It's a pain, but it's really n ice, and some what future proofing what I'm doing (I'm already on the hunt for another z8). Or should I focus more on moving the mics inside BFD and then sample 3 or 4 versions of that instead of the individual mics?


I look at it this way. These programs are so detailed, it's essentially like being able to make my own sample library. I'm into a lot of different genres. I juse my mpc for everything from punk demos and recordeing/making beats for my friends singer/sonwriter coffe BS to pulling samples off vinyl into my s950. I've found my perfect workflow and my perfect gear, and I'm ready to start building.

I just don't know which way is the more practicle way. Mostly my question, I guess, is would you guys prefer to have a massive library on a drum by drum, mic by mic basis, or just a TON of premade kits? Which would you find to be ultimately more versatile? I have a few friends with 4Ks who I've been sharing my work with so far, but they're super new to all of this (like...less tan a month) and have no input. I have no idea of the legality of any of this, but if possible, I would love to just share what I've been doing, so seriously, which direction would you guys go?

And seriously...does anybody know how to automate VST settings inside Extreme Sample Converter? And make multiple programs? If I could automate the sampling part of this, all I'd have to worry about is the assigning, which is, thankfully, fast enough on the 4K.

First of all I don't get why you want to do this. After a ton of work, you'll end up with kits which are more lifeless and unresponsive than just using BFD in the first place. Unless you are intending to sell your final programs or something?

I am doing something similar with Addictive Drums. I'm taking some samples from there, bouncing to tape and then into my MPC1000. I use a TD-12 e-kit to trigger the sounds, and I'm doing this to avoid the latency of vsts, as well as have a hardware-only solution in one box, ready to go should I wanna do some finger drumming when on the road.

In doing this, I've run into several issues which you haven't mentioned, which leads me to believe you are a) not using an e-kit and b) haven't gotten too far with your drumkit recreations.

Issue one, the jjos2xl doesn't support randomized samples from external midi control. This means that each snare hit is gonna come out identical. To work around this, I have to duplicate my program for each variation I want. So I end up with kit1var1, kit1var2 and kit1var3 which all sound almost identical. Then, using p.stats screen, I have a pattern which very quickly sends PC messages to the drum track to constantly switch programs. This randomizes the hits slightly, but its a pain to keep editing 3 programs for every little tweak.

Issue two: you can't recreate the modelling of the drums in the MPC. I don't exactly know the capabilities of the 4k, but my 1k has no way to take advantage of the positional sensing CC messages sent by the e-kit, to crossfade samples of a strike on the center of the head with a sample of the strike towards the edge of the head. In a similar way, there is no way to blend the rim sounds when striking both the head and the rim at the same time. Addictive Drums models all this really well, which leads to a very responsive feeling when playing.

Issue three: There is no easy way (that I can think of) to move 'drums' between kits, ie, exchange a snare from one kit to another kit, as the ambience effects are all prerendered in the samples. One of the things, for me at least, with a good drum sound is a coherent ambience to all drums, so it sounds like one kit in one room, rather than a bunch of oneshots all thrown together with a bit of reverb and compression in an attempt to glue them together.

You mentioned automating the sampling procedure. I do it like this: I set up patterns or tracks in MPC which trigger notes in the VST on quarter notes. Ie a low velocity snare hit, a medium, a high and a full. Play this sequence slow enough so the tails play out to silence. Sample it to an Audio track, which will stop recording at the end of the bar. Then use chop to make 4 even length samples. The hits will be equally spaced so you don't have to do much (I normally end up adjusting the start a tiny bit anyway). If you use a pattern, you can easily transpose it to target the next drum you wanna sample. Finally, its very useful to have midi data for snare rolls that go smoothly up and down in volume, so you can test the velocity mapping and make sure there are no nasty transitions that stick out

A final drawback is in the hihat modelling. I have not found a way to be able to choke off an open hi hat, nor can you mix between varying degrees of openness. On a real hihat you can hold the pedal tightly closed, fully open, or anything in between, and your hat sound goes from super tight to sloshy to fully open. This isn't possible to reproduce via midi cc on mpc 1k. Also, a real hi hat can be hit when fully open, and then closed. As the two cymbals meet they buzz against each other and start to mute each other and finally go fszzt and become silent. The vsts handle this to some degree but you can't recreate on the MPC either.

Finally, in summary, I think you are heading in a bit of a stupid direction to be honest. Why build a worse library on the MPC when you can just use a very dedicated vst for much better results? I'm doing it for a handful of kits only, and am in no way trying to reproduce the vst.. that is impossible due to severe limitations on the mpc (you can't do a realistic drum kit with just mixing samples! The snare is the heart of the kit, and velocity mapping is only ONE dimension. you can strike the snare anywhere on the head, and with a varying amount of the rim in there too, and the mpc can't handle multiple-dimensional mapping like this, afaik. The VSTs are doing a lot of complex mapping there to achieve this, and then they add the mic and room sounds afterwards, which the mpc can't really model either. some vsts and drum machines will alter the samples based on what you have already played.. for example my td-12 knows when I'm doing a buzz roll and adds in some extra buzz roll modelling and assistance. Cymbal washes require the program to know that the cymbal is getting repeated hits, and smoothes out the attacks or something; you can't just layer up the same cymbal hit over and over cos it'll sound like ass.

manji wrote:Would it be better to continue sampling these drums on a drum-by-drum basis so you can make your own kits inside the mpc? Or would my time be better spent making entire kits inside the BFD environment, and sampling on a program by program basis?

I set up the kit I want to sample in Addictive Drums, and tweak it till it sounds right. Now, you could sample each individual hit out as dry as possible, so If you want to mix and match kits, you could sample each drum to its own program, and so have your snare on one channel and go through programs to pick the dry snare you want. This approach lets you create new kits from parts, but pushes the problem of room ambience to later in the signal chain: have to run the signal into a reverb or something and this hasn't ever been as nice as using the vst for the job.
It might be possible to sample only the overheads and get the MPC producing these at the same time on a separate bus, and mix it in later too, but that's annoying and fiddly and prone to phasing issues etc.

manji wrote: Should I make kits with variation folders inside, built on templets? Make variations for a "tight kit", "roomy kit", etc etc? I think I'm going to keep doing the channel solo idea. It's a pain, but it's really n ice, and some what future proofing what I'm doing (I'm already on the hunt for another z8). Or should I focus more on moving the mics inside BFD and then sample 3 or 4 versions of that instead of the individual mics?

I think too much gets lost when you break the kit up into separate instruments. As mentioned, my approach is to design the full kit in the VST, then sample it with all the ambience and set up the kit in the MPC to match. This is less flexible so the original kit needs to sound good, it's difficult to tweak it afterwards. Set up a nice kit and tune it right. Then create a couple of variations, close miced, roomy, boomy, punchy, etc. Personally I'd avoid having any kind of flexibility in the mpc cos it's a royal pain and has never been as nice, to my ears. My overall approach is to leave as much as possible to the VST, which is what it is best at, and just do basic sample playback in the mpc, which is what it does best. If you have some tips on being able to handle mic positioning or dealing with overheads on separate busses in the mpc or whatever then I'm very curious.

That's actually what I'm thinking, the more I sat there and pondered the work load. It also takes care of the assigning for me if I play my cards right, meaning more time spent putting kits into my sampler than building these stupid programs.


That's a great idea about how to build programs for each drum, though. I was wondering how you would actually go about individualizing kits without reassigning everything blindly to see what would sound right. The individual drums are easy enough to sample in, especially with the assigning done, but I think my time would be better spent like you said...making my kit in the vst.


Extreme Sample Converter makes the programs into keygroups instead of drum programs, and the Convert thing in the Program mode of the 4k messes up the assignments. If the program is already assigned, am I missing anything by keeping them as keygroups instead of drum programs? Or is there another piece of software that can convert the program types without messing up the sample assignments?

I totally missed your first post.

I agree with you about pretty much everything, and am trying to find work arounds. The benefit of having the 4K/Z8 combo is that they are VERY competant samplers and the 2 of them together gives me a gig of sample memory. For instance in BFD, if a snare has say 8 articulations, I put all 8 on Bank C, with all the variations I can think of. Your open to closed and closed to open hi hat example is a good example of why I'm taking my time and trying to get everything in the mpc. Some of these hi hats in BFD have up to 18 articulations I've found, so I'm trying to get all of them. There are definitely a few things I'm not getting, like I'm not making seperate samples for the top and under mics on the snares, etc etc.

The 4K has various ways of helping keep things from being too static. That's a bit of the point of the HUGE velocity layers I'm doing (32 for final mixes), to get a little variation. I could also always make a 2rd snare bank, make minor tweaks to the VST, and have another set of the articulations, just to kind of randomize stuff and have a bigger sample pool. There's also various midi pitches and nudges and stuff that I always do anyway, just to keep things a little funky.

For your third issue...do you know how the VSTS go about that? If I make my own kit in BFD, it's not like they recorded each possible kit combination, so how would it be different inside the mpc? In BFD there are ways to sample little things, like you can record just the snare aspect of the resonance you get when the kick is played, and lay it over your kick and I can't particularly tell the difference. I'm trying to reverse engineer this as much as possible.

As for this being a stupid direction...I couldn't agree more. Like I said, I don't think any of this is particularly usefull, more just a means for me to A - understand the limits of my gear and B - understand and refine my understanding of the art of sampling. I think it's interesting to try to understand how all this stuff works to try to just broaden my knowledge base and apply it when nessicary. Plus...dorking out on all of this is way better than TV. I'm kind of like you, I want a hardware only solution for actual music making. I have a crap computer, it can maybe handle a few voices at a time without cracking. Good enough to pull samples from, not good enough to actually use. But beyond that, the musical part of me gets lost on computers. I like the sound design aspect of it (like what i'm doing now), but when I sit down to actually make songs, I just get lost in all the options. The hardware aspect is good at keeping me focused for actual song creation.

A lot of this is truly just testing the possibilities and to see what's doable. I don't understand Kontakt (for instance) very well because I barely ever use it, but honestly, I don't see a TON in there that the 4k/z8 can't handle, in theory, based on my limited knowledge. Not the same, as obviously the filters are different and each instrument comes with its own effects suite and blah blah blah, but in terms of just pure samples, I guess I just want to see what's what between different formats. Just to know, I suppose?