Discuss the various methods you use in music production, from compressor settings to equipment type.
By manji Tue May 28, 2013 4:46 pm
I traded a console to a local guy who, amongst a few other things, gave me a computer with a bunch of software on it. I've been sitting here poking around, thinking about how im going to get all these gorgeous steinway pianos into my mpc/z8, when I stumbled upon like 300GB of FXpansion BFD packs/programs or whatever. I loaded it up, and I'm pretty impressed. But the thing is...I can't stand making music on a computer. It's a good sound source, but if I can't get it into a sampler pretty quick, I inevitably lose interest in it.

So, what I'm going for is how to get the most realistic drums, with velocity and all that, out of my mpc2500 (soon with jjosxl3, but for now just jjos 3.15 or whatever the free one is at) and/or my z8. I've got a lot of experience with this from ripping Drum Kit From Hell and a couple other things, so I know how to construct the kits once they're inside, but the thought of spending the next month sitting here getting these kits sampled makes me want to vomit.

There are two features of this program I really want to keep - being able to customize your kits (i.e. mix the birch and tama kit) and I really like being able to control the mics. Especially that overhead.

What my plan is at this point is to automate (I wrote a post on how to really easily automate multisampling, but I can't find it, it was on the Tips and Tricks post in the 5k forum) 8 velocity layers from each drum (ie 8 layers of Tama Kit Snare, then the same for Floor Tom, etc etc) and that way I can still easily mix and match my kits. I was then considering maybe doing each drum twice, one with the overhead mic as close as possible and the other with it as far as possible.

Would I be able to say have Kit 01 with the mic really close in one program, have Kit 01 with the mic really far in another program, and then mix between the two programs to be able to get something at least some what close to the same effect as being able to "move the mic"?


And really, here's the question - is there a better way to do this that I don't know about? I tried just copying the wav files, they're protected (of course, but it never hurts to try). In looking at other fxpansion stuff on their site, Geist actually looks like a way to maybe do the entire process inside the computer and then just drag and drop wav files and build the kits inside of the mpc (which would definitely be prefered). Does anyone have any experience with Geist?

Would I be able to load up the drums from say XFL into geist, or is it something that can only use the geist expanders (I have zero expreience with software, sorry for the super retarded questions)
User avatar
By Metatron72 Tue May 28, 2013 5:33 pm
No the other FXpansion libraries won't load into Geist. Although the one expander set available is a Best of BFD. On that one they have the full 8 layer slots used in the kits.

The proprietary "container" or however the BFD libraries are stored, I wouldn't know at all how to extract that content.

Full disclosure I use software a lot, fair to say primarily really. But I have MPC's so I can visualize what you're wanting do. (And getting dizzy thinking about the work involved as you mentioned).

If it were me I would just get the BFD software and trigger and sequence it with the MPC. Then you avoid having to rebuild the whole library meticulously. You'll need a host app to run the VST or AU of BFD, but you could have all the sequences in the MPC, take advantage of the MPC's polyphonic aftertouch. (I would guess BFD has a way to use the after touch).

Maybe something like Sample Robot could do the automated multi sampling. It's generally used on synths and keys VST's but I would think it could do the same with a multi sample drum program. Although I'm not sure if there are any proprietary bits of scripting "under the hood" of BFD that an automated app might miss or not be able to translate.

http://www.samplerobot.com/
By manji Tue May 28, 2013 5:55 pm
here's the process I've got for multisampling...much of anything, custom tailored to whatever I am specifically sampling at the time.

This is the importance of template sequences and programs and all that

I have a template sequence saved on my mpc that, every quarter note, triggers the ntoes I want how I want them (so say it plays the A, C and F# notes), on time, as they're sequenced, up the keyboard (for this, im setting up one that plays each pad, in order instead of doing it by note/scale).

Then, as there's 4 layers available on the mpc, if I want them, I make each note play at a set velocity. For example, for my hardest hits, I set each note to play at 127. Then for my next layer, each note is set to 90, 60, 30 (if I do 8 layers I'm going to have to rework this, but you see what I mean).

The end result being tha tI have four sequences, each one playing the A, C and F# notes, in order, up each octave, at a set velocity. Sequence 1 is 127, sequence 2 is 90, etc etc.

Then I add a junk note at the end, one that plays for a while. I'll explain that in a second.

Then I set it up to sample (make sure it's on the sequence you want, with the right note spreads and velocity), hit play, and let it do its thing. Then on the Sample page, I cut out that last note. The point of that is that you would never get the sample to stop directly on time. This way, by adding a last note thats garbage, you know where your new endpoint needs to be.

Now, since everything played in time, and you cut out that last note....you have (for example) 16 notes you need to chop. But since everything was done on time, you just set the mpc to chop it into 16 equal parts and viola, you have your multisample. Naming is somewhat tricky, but if you absolutely automate your workflow, it doesnt matter too much, with the MPCs auto-renaming.

ie Piano 06 127 is the name of the original sample, Piano 06 being the name of the program, 127 being the name of the velocity). Chop it into 16 parts and youll get piano 06 1271, 1272, 1273, etc etc.

From there it's just a matter of assignment. I also have template programs for my keygroups. If I sample the A, C and F# notes, my template keygroup program is already set this way. With the sample renaming, I can get a 4 layer, 61 key multisample program assigned in under 5 min, usually closer to 3-ish.

All in all, my typical "Hey, I like this and want to put it into my sampler" to "well, now that's done" time is between 10-30 min, usually closer to 10, depending on how long the samples need to be.

Autoloop on my z8 usually works, but if you want infinite sustain you usually have to do it by hand, which sucks.



Also, this basic structure works with any sequencer and anything with an autochop function that can be in equal parts. I've been doing this for years on various equipment and have built up a pretty extensive library of my own multisamples. They aren't the most pro sampled things on earth, but they're mine, know what I mean?


I tell you all that to tell you this - I just got a 4 layer kit into my mpc in about an hour, which is usual for customizing thew orkflow, but now everything is saved. The sampling in and of itself isn't too much work (plus, Im going to pick and choose from what I like. You're right, doing the whole library would suuuuuuuuck but not be impossible).

I just can't decide if I'm losing versatility by sampling whole kits into programs into my mpc, whic means I'd have to sit here and work on each kit for a while, then do my whole sampling process or if Iw ant to just straight sample each individual drum, save that and then create custom kits inside the mpc. I know I lose/gain each way I do it, but I was mostly just wondering which one y'all thought was best.
By Abscrete Tue Jun 25, 2013 4:39 am
hey Manji how is it going with that work on the multi sample stuff?

I also have been looking into it but man it seems like a lot of work the way you do it.

the reason I`ve been looking into this is because I like to make my own drum sounds I like to process and make new kits all the time so i was just looking a way to record my kits and be able to play them on a sampler, you know i love to create on soft and hardware but i dont like to play software on stage you never know when the **** computer is going to crash so a sampler would be better for live performance but I dont know a sampler capable of loading my kits I would like to have a sampler where I could load a multi with 127 velocity layers .

but reading over the internet one of the best from akai its the z8 but i dont think i can load 127 velocity layer can I?

you know the point of software like bfd its that you have way more control over dynamics and formants so the sound its more natural and its way more expressive when you play over controllers like zendrum its way better for me i like the feel and sound but i dont want to play it from a **** soft lol

so far the only option i found to do that (sort of)
its to sample my kits with something like sample robot or even better samplit that way i can automate the whole process capture multis with the 127 velocity layers then export that shit to a kontakt or camel sampler which is soft :( and then use it into something like the muse receptor or open labs gear

which are computers but those work only audio so they wont crash so they say so im like mmmm not really happy its very expensive that shit so any other ideas

does any one know a hardware sampler that could support all those velocity layers?


thanks cheers.
By Clint Tue Jun 25, 2013 2:06 pm
Sometimes, you just have to go with the original VST and keep it moving. You can still sample a finished rhythm track into a sampler, and after all that hard work trying to transfer samples from the VST, 9 times out of 10 the finished track ends up in the computer anyway.

Other options include multisample Sound libraries on CD. Or check MPC- Samples.com, pretty sure they'll have some good multisamples ready to go.
By Clint Tue Jun 25, 2013 10:04 pm
That's fair enough, we all have our ways. Maybe Translator or Awave can convert those BFD files? Certainly worth looking into. Let us know how you get on with this task, and good luck!
User avatar
By Metatron72 Tue Jun 25, 2013 10:33 pm
None of the VST's mentioned require to use loops and are not loop based products. They include MIDI files you can use with quality playing and realistic articulations and techniques, but many just use those as an example to look at to do your own thing.

These products are just computer based sound modules intended to be played by keyboards, pad controllers, MPC's on up to full drum kits with MIDI out capabilities.

The reason these products are computer based is the humongous amount of velocity layers that must be recorded so the instruments play/react/and sound like the real thing. They also offer access to vintage drums that may not be easily available like older Gretsch drums and classic Pearl drums for example.

There's not really anty hardware sampler I know of that is going to let you have a drum kit that is 2-10GB's with 128 velocity layers available to trigger live with your MIDI method of choice. Loop libraries are a different thing even the ones that allow some live control of the loops.

These products are full simulations like a modeled synth VST, just sampling is the easiest way to get the real sounds because it was recorded off the real thing.

But if it's your kits and sounds you're sampling you're going to hit limits fast as far as hardware sampler RAM and velocity layer limits, not to mention you'd be recording samples forever if you want to come near what a BFD/DFH type product does.

If you have the patience for all that recording and editing, you might be better of using something like NI Kontakt and treat it like a hardware sampler with 127 velocity layers possible and RAM only limited by how much you have.

I understand the two related things people are trying to accomplish here, but after a point software connected to external MIDI is really the only way to have it all either with a pre rolled kits situation like BFD, or a sampler VST like Kontakt id you need to go extra deep with it.

If you don't need all those velocity layer maybe hardware still will work for you as far as housing the sounds themselves. But it's a kind of thing where you'll have to choose based on how much each option fits your wants.
By Abscrete Wed Jun 26, 2013 1:10 am
thanks all you guys.

yeah I`ve been trying to find the solution so far the only option I have is the following as we all know a hardware sampler is not going to do the job we have some prety interesting stuff like the elektron sampler but is not really for the kind of job i need so

the set up will have to be as follows

i dont just use bfd i use actually a combination of other process together with stuff like bfd so is just not converting the bfd library to another format is to record the whole process so the best option would be to record the kit the final sound with samplit software i can clone the whole kit automatically so i dont need to edit anything that would be crazy with 128 velocity layers and 16 kit pieces !!!!!

then export that library to a format native instruments battery would read and its ok battery can handle 128 velocity layers and then i would have to play it from a computer thats what i didnt want but i was just thinking i can probably freeze a couple of tracks on ableton so the kit doesnt use that much ram and since all the work is done i dont need to tweak the kit live so for just playing the final result should be ok

i heard the muse receptor its doing prety good job and its stable handling vst live so maybe if i have some money later on i will try one its very expensive will see
how that works :P
By Tobintech Tue Sep 24, 2013 7:55 pm
Look at this site for hints on how to build drum sample sets from VSTs with layers. you can stop once you have the WAV file you need, the rest is how to build sample sets for the drum module the forum is for.

http://www.2box-forum.com/index.php?board=9.0

These are instructions specific to building samples for the 2-box DrumIT-5 drum module (aka Ddrum5) from same developers of the original Ddrum products before Clavia sold off the name and killed the Ddrum line. Benj went and founded his own company and finally released 2-box a few years ago.

And BTW there is no real limit on layers in 2-box, aside from memory, but most find 8 to 24 sufficient depending on their application and needs for realism and dynamics in playing. Other wise 4 is more than enough with creative cross fading and filter application