Key Switches are keys on the midi keyboard which allow a composer to go from one style of play to another quickly to add realism and flexibility to your song. They are often color coded differently from available notes an instrument can perform for easy detection. For example say you have the Solo Violin KS (Key Switch) patch up. You start off with regular style of play and decide that at a certain part of the song you'd really like the violin to come in slowly, longer attack of sorts. You simply click the appropriate KS and all the notes will have this property. You can switch between these styles at will and hear your violin part add much more layers of realism.
Another example is with the flutes. Say you want long sustained notes, and then you want the flute to play trill *going up and down a semi note really fast*, there is a Key Switch you can press on the keyboard which will change the notes available to you into trill notes. You can switch back when you are done the part.
Many people refer to instrument libraries as VST's *guiltily myself included*, but in reality the two are different. VST's refer to plug ins which affect the way things sound. Compressors, exciters, filters, equalizers, etc. All these things which by themselves emit no sound, but help mold the sounds to your desire are VSTs.
VSTi's are the instruments. Synth or orchestral makes no different, they are the instruments and though many VSTi's come with decent abilities to mold their own sounds, they very rarely will have the power VST's will give you as developers don't want to spend money on things that will take away from the quality of their package.
Velocity layers are layers designated for notes on an instrument, to perform differently as volume levels are changed. Most high end libraries will have up to 4 velocity layers per note on several of their instruments for added realism.
Lets take it lke this, imagine that you pick an instrument, like the violin. And you want at a certain segment to play really loudly. What would you do in real life? You would take the bow and press down on the string, applying more pressure to generate more sound. As a result what happens to the texture of the note? It gets more course, rawer and more aggressive. VSTi producers recognize this change and will have recorded the loud note, so when you click it and say loud, the library will find the sample which will correspond to your demands by finding the one which is course and raw. If you then went into a quiet segment, and set the volume to very low, again we consider what you'd do in real life. You'd slide the bow gently over the strings, and as a result the sound is very gentle and smooth.
Recording this as well the producers let the library find the soft texture and play it. Understanding now? It's like in flash/photoshop where you have layers, but you can only see one at a time. Depending on your volume level the library will find the texture appropriate for it. This is generally why libraries can get very large. If it takes up to 40 megabytes for one instrument, with 4 velocity layers it will take up to 160 megabytes, assuming of course they havn't lowered the quality in one set which is really never the case.
DFD is a recent addition to orchestral libraries, inspired by the massive RAM consumption necessary to handle them. It means Direct From Disk, which is frankly, taking samples directly... from the disk. By doing this RAM space is saved and more instruments can be added into your efforts allowing composers much more freedom. How does it do this?
In a nutshell DFD will load say the first 10 milliseconds of every sample necessary to RAM. Leaving the remaining samples on the disk. When the note is pressed RAM will play the first 10 milliseconds it has, and the split second necessary to retrieve the data from the harddisk is covered by this. After that the sample is played through the Harddisk until the note is let go or by some other means the sample is dropped. Users are allowed to modify how much RAM is used per sample and how much relies on DFD. It is however discouraged to change the settings unless you know what you're doing, otherwise you can get very choppy and sloppy performance, or sell yourself short on RAM.
There are a few things you could do to increase performance for not only DFD, but in general.
1: Try to have a hard drive reserved especially for the samples. DO NOT get a hard drive with less then 7200 RPM SATA (see Computer Jargon section below for more details). This is generally the industry standard nowadays but cheaper and worse alternatives are out there. Getting something worse will result in less data retrieval which in the end will make you rely on your RAM to compensate.
2: Get quick RAM. DDR2 being again the industry standard. Try to get 667 mhz.
3: Running multiple instances over running multiple instruments per instance will provide you better performance. However this may consume a bit more of your RAM as some audio engines like Kompakt take up some initial RAM. This is especially true for Kontakt, which takes up to 100 mb's per instance.
4: Do NOT run your samples from external hard drives. The USB, or Firewire cable is just not going to cut it in terms of data transfer. USB 3 *technology in development) MAY fix this problem as transfer rates are up to 10x faster then USB2 (from 6mb/s to 60mb/s), but until then, say no to slow!
5: Refrain from using big libraries on Laptops. Their harddrives are generally 5400 RPM, and data transfer rate is often not good enough for the massive strain high end libraries inflict. However this does not apply if you are using electronica or techno VSTi's which are just processor intensive. Even then you'd need one hell of a laptop to run as powerful as a desktop.
Mic Positioning (s, f, c)
Mic Positioning refers to where the mic was during the recording of the sampling. Why is this important you ask? Well as good as reverb technology is, there's simply no substitute for the real thing. Different Mic Positioning allows different sounds for the composer for different reasons. For example, say you want to make a song with thumping drums. For this you would use the "s", or "f" mics. Mic positioning is explained below from closest to furthest.
c - close. The mic is right in the face of the instrument. Generally very dry samples with no or little reverb ideal for industrial or other means where the composer can add effects themselves without worry of having to cut or trim out unwanted reverb or muffled noise.
s - stage. The mic is sitting where an audience would be. The instrument in question of course, is on the stage. Ideal for creating pieces in which the composer wishes to sound as realistic as possible in terms of a performance.
f - far. The mic is a good deal further away. We're talking very back. The instruments have a natural reverb, slightly muffled due to distance, textures somewhat dulled. Ideal for ambient compositions in which you apply your own form of reverb. I find far instruments are great for molding. They do however tend to sound heavily reverbed if you put too many of them together affecting austhetic qualities.
There are also mic positionings on different levels in terms of height, however these mic positionings are not often deployed. A B C D they are often referred to. There's no meaning behind the letters, only where they are positioned. For example, say you had a bass drum, Mic A would be positioned at the bottom of the drum, absorbing all the thick thump. B would be positioned on the lower edge of the drum, C is positioned in the middle. You get the idea. The reason behind this is to capture different naunces. For perfectionists, it is sometimes difficult to hear the difference and in all honesty in the hustle and bustle of a full song these differences are drowned out, so don't start making choices on "well this library has 30 different mic positionings!!!" In fact this is a bad thing, the more mic positionings, the more space taken up for one instrument, and less substance you have to work with in total. I'd rather 5 different percussion, then 1 percussion recorded in 5 different locations.
Hard Drives: There are many types of harddrives out there, suited for many different needs. For orchestral libraries you'd need a 7200 RPM (Rotations Per Minute) SATA hard disk. There is also 5400 (mostly laptop harddrives) and then there are faster monsters (10,000 - 15,000 RPM). The faster the better for you. There are sometimes complaints about how the fast harddrives make a whining noise all the time, but with todays advancements this is rarely a problem.
For the more tech savy you could set up a RAID set up. RAID set ups increase performance by telling the computer to think of 2 or more hard drives as one unit, so when it is seeking information 2 or more seperate hard drives are both going at it at the same time to retrieve data at a much faster rate. RAID set ups are a little expensive as you have to have all the afflicted harddrives to be the same, and three 500 gigs hard disk in a RAID setup won't equate to 1,500 gigs at all. RAID thinks each 500 gig hard drive is one, so if you're worried about space RAID may not be the best way to go. You get sweet sweet performance though.
RAM: DDR2 is the standard RAM that comes with any computer. It is in fact now impossible to find ANY retailer who sells DDR because they're slower, less efficient. 400 mhz and 667 mhz is what you'll generally find in the DDR2 family. These refer to the speeds of transfer between data, so 667 mhz will transfer data between harddrive and itself at significantly faster rates, allowing you more robust control.