Concerning the load times, they are not long at all, about 30-40 seconds for large instruments. This is on a 1.5 ghz system, so you can probably get a lot faster. RAM useage seems also fairly small. I have only 512 mbs, but nothing has been strained in that area. The polyphony was up to 64 note on some rapid single line passages involving the four-layered instruments, obviously because of envelope release tails.
The problems with Gigastudio bug me, and there is no fix available. They occur with the most recent updates. They involve crashes, not really with doing any spectacular - just using the program. And also they cause, almost every time the system is booted, disc consistency checks, warnings about having to use chkdsk when attempting to defragment, and a few other very odd errors that have occurred ONLY when Gigastudio is installed. This computer is flawlessly functional at other times, including running a professional video NLE, Sonic Foundry multi-track recording on Vegas Video, and full DVD encoding and burning functions. All of those, which dwarf most applications in data handling, work perfectly. So the fault is not with this system. I've heard some buzz that Gigastudio is going to have to come up with a new version to solve these problems. However that may be I do not like the way they handle their customers and have been turned off totally by the company.
I am thinking very hard about other options. All of this is solely to use VSL. Otherwise I would kiss Tascam good bye with a few choice obscenities. I have both EXS24mkII and Gigastudio, but no matter what anyone says I do not believe hardware samplers are out of the picture, especially when one considers the 512 mb AKai Z series or a bank of maxed-out Emulators combined with multi-track hard disc layering. The amount of data that kind of system can handle is gigantic. And it is all totally reliable and controllable, unlike Gigastudio. And the programming is straightforward, and actually more flexible than Gigastudio 2.5. For example, EXS24mkII imitates EOS in its design especially with being able to route any modulator in midi to any destination. Gigastudio cannot do that and that immediately makes it inferior with my set up, because on the Miraslav library I used an approach that involved the pitch wheel for filtering and volume so the mod wheel could be used for increasing the volume of attack articulations. I used absolutely no velocity for controlling volume of the main sustains and layered detaches (or tongued slurs), so that it could be used to create varying degrees of marcato attacks with the staccato samples. This system, when applied to strings and brass, allowed for a completely intuitive, "felt" expressiveness to the initial performance, without having to think consciously about programming all the different layers, etc. One could simply play an entire violin section line, for example, from smooth portato or legato (though without VSL's actual legato samples) to varying degrees of detache and marcato in real time. And if it was a tricky one, overdubs in midi could layer in the controllers for volume or attack. EOS is an incredibly powerful piece of software (which no one today seems to want to hear about) and NEVER CRASHES. EVER. The instrument simply functions once you turn it on. Like picking up a trumpet or a guitar. It's all what you do with it, not what problems the computer decided to create this morning.
Does anybody care about that besides me? I know stage musicians do, because they do not touch with a ten foot pole a sampler like Gigastudio. Can you imagine such a thing? "Sorry ladies and gentlemen, but tonight's peformance has been cancelled due to disc consistency errors."
At any rate, I am in awe of the work of VSL in the conception and carrying out of this recording project which is beyond any ever attempted before. The actual work done by them is an art so far above all these damnable computer problems that it is an insult to discuss them in the same breath.
Though I should probably say a few Hail Marys before admitting it, I actually have one criticism of the samples. The solo oboe has no espressivo vibrato in the long notes, only in short notes. This seems odd, and perhaps the result of a problem with the oboe player (?) The other woodwinds are astoundingly beautiful. The clarinet captures the dark chalameau and pp expressiveness wonderously, and one can hear the exact same breathiness in the recordings - something I've never once heard before in samples but heard every time I was next to a live clarinet player.
The timpani rolls have perfect release samples, so that the roll is finished off precisely with an eighth note. Here, as well as in the string tremolos, the samples are not long enough to avoid loops. So these will have to be looped at some point, either by users or by programmers.
I would also add to this already overlong post that the solo trumpet and horn represent almost maniacal attempts to capture everything the instruments do. (I haven't yet heard the tuba and trombone.) The varying lengths of short notes and dynamic changes are so subtle that it is hard to imagine how one could do more with the original instruments.