@Another User said:

I can also load several times more samples in K2 than in VI.

I can load on my XP systems ca 25.000 samples.
We decided to develop the Core engine with fixed buffer sizes to assure stable performances.

best
HerbI get stable performance from my template using several instances of Kontakt 2 with minimal buffer sizes. The VI is clearly wasting memory even on my setup which is far from being the fastest setup on the market.

I know I can load only patches, there just aren't light enough patches. I'm forced to load two repetitions for every sample which is fine for staccatos but that doubles the load of already heavy legato and portamento patches.

I know I can use memory save but I don't want to. I want to use a template and have everything available all the time.

@Another User said:

several instances of Kontakt 2 with minimal buffer sizes

would you please kindly waste the same amount of energy describing what precisely this is (*several*, *minimum*, actually which buffer) as you present us describing what you don't like?
christian

Hi, being new to the Vienna Instument i went throught a couple of threads in this forum before asking this questions :

i juste bought the special edition and i think it's great, the player i realy very powerfull and the sound really rock !

however here is my concern....CPU power !

i have a dual G5 2 X 2,3 with 3,5 Go RAM runing OSX 10.4.8 with Cubase SL 2.2 as host and as i descovered that VI is mono timbral in the sense that i have to load an instance for each instruments i whant in my template, so let's say that i load around 20 instances to make a medium orchestra that will put my CPU load to 70% !!!

no problem with the RAM or Disk flow aparently...so is that kind of performance normal ? i couldn't find Nick post about loading 7 Go of RAM but is there a workaround because with this kind of CPU load i can't play much...

i'll be soon upgrading to cubase 4 and adding 2Go of RAM but o'm not shure that will solve my concern.

thank for your help.

@Another User said:

I know I can load only patches, there just aren't light enough patches. I'm forced to load two repetitions for every sample which is fine for staccatos but that doubles the load of already heavy legato and portamento patches.

there are no two variations of legato and portamento patches??

@Another User said:

there are no two variations of legato and portamento patches??

That's my bad then. I tried some legatos and they didn't sound the same every time I press the same key. Maybe that's a feature. [:)] I use SIPS and regular sustains instead of sampled legatos in my template and don't have exact figures but the VSL script legatos for K2 don't use anywhere near the 60 megs most of the Special Edition seem to require. I have to look into this.

I want to apologise for my harsh approach on this matter but I know that you have lost some friends because of the VI and having now tried it myself I can understand why. It's not just your products. You have started something which propably would have started anyway but developers developing their own players is the worst case scenario for many users, myself included. One technically efficient and easy to use player for everything is far better than a different player for every library. I don't know if it's the copy protection, licence fees or what that made you build your own player but for me the reason doesn't seem to be technical superiority.

I have four different string libraries (one of yours) in K2 blended into one so that the best samples are prominent at different velocities and instrument ranges. I use several scripts where needed. I have compensated for the normalization in the instruments so that a low clarinet register doesn't sound as loud as the most prominent range and so on. The VI is really a leap backwards for me in both usability and technical efficiency. I know this isn't something you want to hear but it's my sincere truth and I think it's better for both sides that I tell you about it.

You can't compare a scripted legato, which is using only simple single note sustains with true interval legato recordings.

True legato means, you get real recorded interval samples, one variation (not two) for each intervall, but all intervall steps from minor second to an octave, up and down. So there are 24 different interval samples for each key and for each velocity, same goes for portamento.
Therefore these patches uses an huge amount of samples.
24 times more samples than a single note sustain patch.

If you don't need true legato quality you are simply using the wrong library.

best
Herb

Janila:

Rather than construing this as a drama with good guys and villains, it might be more accurate to say that you, personally don't like the concept of the VI software and much prefer the way of working you have become accustomed to with Konakt 2.

Purchasing the Special Edition was thus, for you, a mistake. I can understand that spending serious money on something you find unusable and defective can only be an unhappy exeprience. There's absolutely nothing worng with saying that and it requires no justification. On the other hand, there are people like myself who find the VI software a pleasure to work with. To coin a phrase: "that's why there's chocolate and vanilla." (Please be assured that I am not a spokesperson for VSL - - just a relatively satsfied user of ther products.)

And to add to Stevesong's wisdom, hardware requirements can often be problematic, coupled with workflow. There's a dedicated thread here about the infamous 3GB switch that frees up more memory in a windows scenario. That's the OS providing a limitation that would be a challenge for ANY developer. Couple that with Herb's detailed explanation of multiple samples used to create that 'sound', and the combination of the two present interesting challenges not only for developers but users too. Hardware's coming along, and hopefully, the new Win OS will allow users to exercise more freedom and 'space' when building up big orchestral templates.

Something i constantly forget, until it bites me, is just what we're asking data to do. And the volume of data that passes over the CPU and through RAM is considerable.
You've given the example of Kontakt, and how successfully it works for you. My one experience with Kontakt was completely the opposite. I thought it a bloated, extremely CPU/RAM heavy player that presented more problems than solutions.
And i found the 'tidying up' after playing in a line with Kontakt was a real pain in the backside. The interface, no matter which colour they dress it up with, is unpleasant to stare at day after day, and for me at least, looks like it was designed by a surly undertaker. (IMHO)

Steve's right. Everyone to their own, but that doesn't mean the product's bad, or has serious limitations. It simply means in an individual workflow, one design may fit more comfortably than another. The VSL player has proved extremely popular, just as there have been passionate discussions about the more general future of locking samples into a format. I for one find it rather incongrous that so much is made of Kontakt as a viable tool, yet it's just as restrictive as a locked format, especially for me, as i have no wish to use it and repeat a less than happy experience. Frankly i don't understand why other developers continue to release libraries in only this format, as, using your own frustrations as an example, there are some of us who don't want it. (And of course the same can be said for Windows player versus itunes formats, both restrictive and locked)

Whatever you do Janila, i wish you luck and success, with relatively trouble free audio 'motoring.' It's no fun to spend a deal of time trying to figure all this stuff out, and we all seek the same Holy Audio Grail, that of writing for a large percentage of time, and spending minimal time with the digital toolkit open trying to tweak something. I for one like going back after writing lines in, and tweaking not only samples used but applying filters to develop the sound in a direction i choose. But then i'm using a tiny self built library that i know well from years of use. Maybe the same will happen when i finally get back home and buy a substantial library to complement what i already use. I certainly know it will take practise, and some deal of dedicated study time to get the best out of any new sample library, and explore all the opportunities it may present, and that includes learning to drive the thing!
And i'm not a VSL owner yet, as i'm still in the middle of my study in Russia. So i get to stand back and be substantially objective about this stuff. (Whether that's good or not, i have no idea!)

Regards,

Alex.

@Another User said:

... 6 K2 instances ... 12 megs of voice memory which gives 128 voice polyphony per instance. I'm using 18k preload buffer to be safe ...

if i'm right and this is 18 kB preload buffer per stereo-sample (44.1 kHz 16 bit) a rough calculation would give me 66 MB/sec throughput while having about 4000 random accesses per second to the disc.
this also would mean you can have about 50.000 samples loaded.

not knowing now details about your machine and harddisks i'd say you should be able run a lot of vienna instruments instances (which are using 64 kB preload buffer per 24 bit stereo sample) on such a setup. i'd also assume vienna instruments would take just a fraction of CPU load in a comparable situation.
christian

@Another User said:

... 6 K2 instances ... 12 megs of voice memory which gives 128 voice polyphony per instance. I'm using 18k preload buffer to be safe ...

if i'm right and this is 18 kB preload buffer per stereo-sample (44.1 kHz 16 bit) a rough calculation would give me 66 MB/sec throughput while having about 4000 random accesses per second to the disc.
this also would mean you can have about 50.000 samples loaded.

not knowing now details about your machine and harddisks i'd say you should be able run a lot of vienna instruments instances (which are using 64 kB preload buffer per 24 bit stereo sample) on such a setup. i'd also assume vienna instruments would take just a fraction of CPU load in a comparable situation.
christianI'm not using anything extraordinary, SATA drives without RAID and NTFS with 64k cluster sizes. I have also split the libraries to different drives so that the busy instruments aren't all on the same drive. CPU load isn't a problem.

It seems that I can get 50% less preload with K2 when the load of the empty instances is taken into consideration, I can select samples and use scripts and I can split all my other libraries on the two computers and K2 instances. That leaves me with a guestimate of 70-100% more articulations loaded in K2 compared to the VI. Maybe that doesn't sound like a lot to you but it means that I would need a third computer to actually get something useful loaded with the VI. That might be good option anyway but it's not what I had in mind when ordering the VSE.

I'm not against the VI but I'm certainly against leaving out other options. I can understand that you have a business to run and you have to consider which options are going to make money instead of just spending it. I still have hard time in believing that happy GigaStudio/Kontakt users are nowadays such a rare species that you couldn't get your money's worth from releasing new Horizon libraries. Can I already preorder my Appassionatas for Kontakt 2? [;)]

Personally I don't see much point in arguing about closed vs. open format. There are both good and bad sides, but VSL has already made its decision. That said, I still think janila has a point regarding making it possible to have xfading automatically turned off for short note articulations. This is something which I'm planning to add to my crossfade script (if I only get time) and I almost assumed that this option was present in VI. Of course it's possible to achieve the same result by modulating xfade using a controller, but that seems unnecessary and in some situations a bit distracting. I'm a little surprised by the response to this since it seems like a nice feature request.

Regards,
Nils

@nliberg said:

I still think janila has a point regarding making it possible to have xfading automatically turned off for short note articulations ... Of course it's possible to achieve the same result by modulating xfade using a controller, but that seems unnecessary and in some situations a bit distracting ... it seems like a nice feature request.

I agree this would be a useful feature.

Best,
Jay