Ive been a long user of string sample libraries and am fascinated with the process of how they are created. As a means of educating myself, I am endeavoring to create my own. As a start, I have chosen the violin the sample (probably jumping in the deep end here) and have all the recording equipment, location (beautiful church) and violin player lined up (a buddy of mine). The violinist plays for the Sydney Symphony Orchestra and has kindly offered his services for this little project of mine. All this is coming to no cost to myself (thank god) and the only thing this will cost me is time.
What I'm really wanting to know is what to give the violinist to play to record all the articulations with? Do I just get him to play all notes individually or play scales and what not? I am not looking for any industry secrets here, just some friendly advice. Please know that what I'm doing here is not going to be sold in any way, this is only for educational purposes. I am hoping that I might be able to link it here later for everyone to look at and use, and to possibly build upon as we go along. I will be keeping a close 'blog' here of my progress and hopefully we can make this a community effort.
good luck with your endeavour/endeavor ... 2 spellings for that word in Chambers dictionary...looking forward to the story...sounds exciting...sure someone here can recommend a Mic...positioning etc...
Having recently done something similar with a girl singer (although the idea was different, I did want her to sound a bit 'sampled') I was struck by how difficult it is to get consistency between notes, never mind layers. I found new respect for the major sample library developers!!
I think it would be good to have a really detailed plan of everything you want to achieve in the session in the way of articulations and layers of intensity and to follow that closely, and I would also present that in a written form (as written musical notation perhaps but not necessarily) so that the player doesn't eventually go mad because it is a slow and exhaustive process. I'd probably keep everything as separate as possible note wise to help when editing, but of course it would be lovely to have scales, trills, whatever as well.
The recording part of it is interesting as well - I'd try to record in as dry an environment as possible - Oh just re-read your post and see that you are using a church! Handling the reverb tail will be a bit difficult if you intend building up a section or playing more than two parts or notes at once. The ambience is effectively double tracked and starts to cancel itself out at higher frequencies, at least I've had this experience with tracking one fiddle player in a room and after a few double tracks it sounded awful.
No doubt much more experienced samplers will know how to deal with that and chime in... sounds a fun project though....
Mic positioning is also key. If you can, do a test run before the session date to find the optimal recording distance and location in the hall for the sound you want. Then, especially if you're doing stereo recording, be very precise on the angle of her playing and distance from the mic to keep the samples from "walking around."
I'm pretty envious! I can't offer much from a recording-point-of-view other than suggesting to ensure you try to have recordings that are close-miced, or as dry as possible. This will let you add your own ambience/reverb later for the situation you're re-using the samples in (and makes them easier to work with). Other than that, the following is just what I would do if I was lucky enough to be in your situation.
Covering the basics first - I'd be wanting to break it down into long and short notes. Short would cover the usual playing styles (spiccato, staccato, pizzicato, legno, 0'3s, 0'5s, etc.). For those I'd want to implement round-robin recording for bowing so I would have the player go through each note and play the same note, same style multiple times (so 8 variations of G3-spiccato-bowed-down, 8 variations of G3-spiccato-bowed-up, 8 variations of G#3-spiccato-bowed-down, etc.).
Then the long notes. Sustains in different styles (vibrato, non-vibrato, tremolo, trill half, trill full, sordino, etc) and dynamic layers (p, mf, f, ff). I'd want to have them recorded both bowed down and bowed up and have those automatically cycled. Again, this for each note - I'm not sure how long I'd want to record them for. Looking at LASS, they recorded for > 8 seconds. So I'd maybe record G3 upwards for each half note for 10 seconds.
I would imagine that I'd also want to create legato interval patches for realistic sounding legato, again in multiple dynamic layers and styles (vib/non vib, portamento, gliss, detache, spiccato). For this I'd have the player go through each half note and play the upwards and downwards interval within an octave (G3-G#3, G3-A3, G3-A#3, G3-B3 etc.) with a 10 second hold to make it easier for looping later. Older legato instruments in libraries used to just fade this recording in, but I think newer ones have dedicated legato interval samples that play. I'd say it'd be better to over record and have more to process later on, than to have things missing you need to rerecord.
Also, something with prerecorded effects including human sounds and phrases (spiccato/staccato passages, runs, etc) at multiple bpm that could be manipulated later on. I didn't cover anything solo-violin related like double stops or open strings or anything, as I've not got much experience with working with them.
Basically, I would suggest planning your articulations and instruments and working from there, Having the player go through each articulation a note at a time makes sense in my head. For things such as legato you will need to record specific samples to make it work.
Are you intending to work with Kontakt for this library? Let me know if you'd be interested in any help, programming-wise. I've got quite a bit of experience working with Kontakt and have plenty of ideas on what you could do, clever-script-wise.
I hate to contradict Blake on one thing after all his fabulous scripts, but unless you really want that sound close mic-ing a violin is not particularly flattering!
Aha, thanks - feel free to contradict. I have near zero experience when it comes to recording (other than creaking on a chair, turning pages and breathing) :D From an end-user point of view, I think I meant more 'dry' than 'close'. Now that you mentioned it, I imagine a close-mic would give a more 'scratchy' sound?