as this thread fades into history, have now deleted recordings to make space on my website[/edit]
 as this thread fades into history, have now deleted recordings to make space on my website[/edit]
Thanks much for the comparisons. Were these all out of the box for all or did you do any pan/eq/reverb apps - especially with VSL?
To bad you could not add Sonic Implants to the comparison. (No matter how hard you try, never can please the people, right? ) You could send midi to me, but undoubtedly we have different sound cards etc and not sure how relevant a test it would be.
There is some reverb, no directly applied eq (but reverb has the effect of eq ...) ; these are all as used in my current piece, so I was simply trying for a good sound in each case.
sorry, can't do sonic implants -- just can't justify the purchase of another library with no concept of string articulations, as if "staccato", "spiccato" and "legato" were the end of regular bowing techniques!
You're kind guys, love to listen to that comparison stuff. Many thanks!
Yea I totally Agree with Theodor Gold all the way!
Can't wait for that Pro upgrade.
Also you guys doing comparisons really have to say which reverb you are using. Convolution or whatever.
Reverb means everything it will make or break a recording...so I have been told by Synergy543 HA!
All I can Say is...HA!...HA!...HAAAAAAA!!!!!
I also liked EWQL, and the GOS/VSL Solo combo.
Is it just me or is there some strange artifact in all the GOS ones? Like a kind of "beating", strongest at the beginning. Could you have introduced something like this when you tuned them?
Interesting too that I noticed the same thing in yours as had been mentioned in Sam's VSL versions - the velocities don't sound quite right, natural or in control. Seems like they might be quite hard libraries to use.
That's quite possible. I used melodyne and tuned hundreds, maybe thousands, of samples over a few weeks. Melodyne does change the wave form as it modifies the sounds, and sometimes adds artifacts that I may have missed.Originally Posted by Ouch that hurts
Yes, the velocities need more work in vsl -- though when I reviewed the recordings, don't think I did justice to the vsl libraries, especially the solo -- you can get good sounds if you take care about the timbre shifts. And for more refined recordings, I usually apply a different reverb to vsl and to gos.Interesting too that I noticed the same thing in yours as had been mentioned in Sam's VSL versions - the velocities don't sound quite right, natural or in control. Seems like they might be quite hard libraries to use.
As far as reverb: for this recording, the vsl, gos, and ao went through gigastudio and a simple hall reverb ("nfx", not gigapulse). EW Gold was in kontakt2, no extra reverb. I still find EW Gold the most difficult to work with, mentally: the articulation names make no sense, and they are not consistent between string sections, so each choice takes squinting at kontakt and trying out the sound. Gold also has something of the vsl problem: different articulations have different dynamic ranges.
I seem to remember when I last looked at VSL in detail (which was some time ago, so I could be wrong) I was surprised that many of the instruments had relatively few velocity layers - like quite a few of them had only two layers, whereas most competing orchestral libraries nowadays have at least four per instrument.
Now I know from using Quantum Leap Brass that two-velocity instruments are a devil to work with, because there's always that obvious "line" to be crossed from one velocity to the other, where the timbre change is noticeable. More velocities allows the programmers to space out the timbre changes more gradually.
Could this be what's happening with these VSL demos? I do take on board that most of these problems could be solved with a little more tweaking and writing "for the samples". But it's still important to note that they're there.
One thing I wonder is this: People often say that VSL eats a lot of RAM, because of having to load all the legato transition samples. Maybe they were forced to make a compromise in velocity layers in order to accomodate this. ie, doubling the number of layers would require doubling the number of transition samples (would it? are there actually different transition samples for each layer?), which for a library that is already stretching current PC systems to their limits, would just not be practicable.
Maybe that's the price to be paid for being at the cutting edge. I've flirted with buying VSL, but I'm starting to think I'd rather save my money and wait until 64 bit Windows is established, and soft samplers written to use it, whereupon it will be much easier for state of the art libraries like this to really stretch out and fulfill their vision, without any such compromises.
But I've never used the library, so I could be talking rubbish.