Acoustic drum sample library in the works, updates!

[Cyrosis]

Hello KVR,

Me and a few friends are starting an effort to create the most realistic acoustic drums we can, we have various drum kits lined up for sampling and it will be set up to mix together the kits however you want. The plans as of now, are as follows..

-45 velocities where possible
-multiple samples per velocity
-sustain, mute, and, crescendo cymbals
-sticks, brushes, and mallets
-assorted secondary percussion
-hundreds of midi beats to get you started
-multiple mic positions
-single stroke, flam, drag, ruff, double hit, and multiple bounce stroke on cymbals, snares, and toms
-recorded at 24/96
-simple and functional vsti


right now, since this is in very early stages, I would love to take suggestions on what would make this as real a simulation of acoustic drums as possible. If you have questions feel free to ask, but beyond the above list, nothing much is 100% which is why this is mostly here to take suggestions.

thank you for all of your input

[Cyrosis]

bump-

edited features list and finalized afew things.

we now have a "shure-sm57" and the "AKG D112" is still planned for use.

[Angus_FX]

You do know that making a "simple and functional" VST that will handle all that is a *far* from trivial undertaking, right..? Take it from someone who knows - this is not something you can just bolt together in SynthEdit! The diskstreaming and memory management issues are *major*; doubly so at 96K. You might well do better to create mappings for existing engines - BFD, Kontakt, Halion, EXS, Gigasampler.

The recording, sample selection and editing is a pretty major time investment too (think $thousands in studio time followed by man-months of mind-numbing editing), and with the amount of data you'll have, you can forget shipping it as a download.. meaning you'd best be prepared to get involved with DVD duplication, shipping physical packages, and all the associated complications.

PM me if you want to discuss this further... I'm not pooh-poohing your idea, but there is major, major work involved in producing something like BFD, DFHS, Scarbee SID, and there is a reason why they all retail for around $300.

[Zoomby]

Will it be free?

[waiting man]

this is not something you can just bolt together in SynthEdit! The diskstreaming and memory management issues are *major*; doubly so at 96K. You might well do better to create mappings for existing engines - BFD, Kontakt, Halion, EXS, Gigasampler.

agree. If you don't have the chance to make a good vsti, just make patches. multiformat and with no damn player in it.

[tonymead]

...and following the development of the nskit, you might want to look at their forums to see the progression of this as it seems similar to what you want to do.

http://www.naturalstudio.co.uk

All the best though!

tonymead

[Cyrosis]

Will it be free?

fraid not buddy..I wont say any numbers right now, but our main focus is keeping this much cheaper then the other large drum libraries

[Radjo]

bump-

edited features list and finalized afew things.

we now have a "shure-sm57" and the "AKG D112" is still planned for use.

Will it be free?

fraid not buddy..I wont say any numbers right now, but our main focus is keeping this much cheaper then the other large drum libraries

You will need alot more than those two mics to make it the best sounding acoustic drum Samples.
Good Luck to ya though fella.

SH

[Cyrosis]

dont worry, these are just 2 we are for sure useing heh, we are still debating the rest

[Douglas Whates]

...and following the development of the nskit, you might want to look at their forums to see the progression of this as it seems similar to what you want to do.

http://www.naturalstudio.co.uk

Thanks for pointing that out

I would second what Angus has already mentioned. The creation of an extensive drum sample library is a massive undertaking and requires a lot of preparation - ns_kit7 has been literally *years* in the making (the first samples were recorded in April 2003).

We made several attempts at getting a custom VSTi plug-in designed, but this proved extremely frustrating. A multiple-mic setup like you propose, especially at 96KHz, with the sort of detailed velocity layering you suggest is, I would suggest, impossible on todays computers. To give you some sort of idea...

ns_kit7, although recorded in the conventional manner (i.e. multiple mic's) only features stereo mixdowns (48KHz, 24bit) of the single hits. Despite that, out kits are still multi-gigabytes each. If we were to use separate multiple mic samples, you'd be talking about 20GB/kit. This is obviously not viable.

I am very excited by the release of ns_kit7 and am looking forward to seeing what sort of reception it gets. In terms of velocity layers and articulations I am fairly certain ns_kit7 will be easily the most extensively sampled drum-kit available. I am yet to work out the exact statistics (or research all sample libraries), but I wouldn't be the least bit surprised if one of our kits turns out to be the most extensively sampled instrument to date.

Douglas.

[Cyrosis]

ns_kit7 has been literally *years* in the making (the first samples were recorded in April 2003).

We made several attempts at getting a custom VSTi plug-in designed, but this proved extremely frustrating. A multiple-mic setup like you propose, especially at 96KHz, with the sort of detailed velocity layering you suggest is, I would suggest, impossible on todays computers.

We are not expecting to complete this overnight, we have already expected years before we release anything. As far as the vsti is concerned, we are still on the look out for the right person for the job, the listed 24/96 is still touchy ground if it puts up too much of a fuss on the memory. The list is what we have planned, and it may very well change, but as of now, everything (but possibly a removal of the 24/96) is staying put.

[Douglas Whates]

In that case, I reckon that 96KHz will probably be viable; in two or three years, computers/software will probably be advanced enough to cope.

I'm not trying to put you off embarking on this project, just letting you know that it has (and continues to be) a much more time consuming endeavour than I ever expected.

One thing I notice from your list - "crescendo samples". What do you mean by this?

Also - you list articulations such as flams, drags, etc. I have found that these articulations are only useful if you have limited velocity layering. Otherwise, it's a piece of cake to play them in with single strokes.

[Cyrosis]

One thing I notice from your list - "crescendo samples". What do you mean by this?

Also - you list articulations such as flams, drags, etc. I have found that these articulations are only useful if you have limited velocity layering. Otherwise, it's a piece of cake to play them in with single strokes.

I believe you misread the text, it says crescendo cymbals, its used more in orchestral music, but to make explaining simple, its the method of playing a cymbal in the fashion that a reverse cymbal tries to mimic and about the various playing styles, you are correct they are fairly easy to create, and with other features being more important (the recent talk of useing wood and nylon tip sticks) they may be dropped as to not just make the file size vastly larger on something barely anyone would use. Thank you for your thoughts

[Douglas Whates]

Sorry - my mistake.

So - you're talking about crescendos on the cymbals with mallets...? We recorded these too during the ns_kit7 sampling sessions but found that they are easy enough to recreate with single strokes.

[Cyrosis]

i guess we just havent had the chance to find that out for ourselves yet i will experiment and see what happens, but considering a crescendo doesnt require velocity layers, the disc space would be a worthy trade off imho.