Solo Strings dynamics problem - velocity switches

[poco_a_poco]

Hello all,
I have read several posts on this problem now - Halion, Gigastudio and Kontakt - related, but found no real answer yet. I continue to have this hassle with velocity switches in my Horizon Solo Strings library using Kontakt 2 - and also with the 'Kontakt 2' version of VSL library. For some reason, this seems to be a VSL problem. At least, I haven't experienced this behavior with other libraries. What bothers me is that there are pronounced velocity and timbral switches between the individual samples of a note. Take, for example the VI_oV_0'5s-1 instrument in the Solo Strings library. There are velocity switches at 56, 89 and 109.

Here ist the mp3 with just this one instrument (VI_oV_0'5s-1), one pitch, and the velocity gradually increasing from 1 to 127.

http://gd54.s7.domainkunden.de/musik/Velocity_1-127_VI_OV_0'5S.mp3

I have previously orchestrated my piano quintet with the GPO strings and got fine results with all my dynamic variations. Now, the quality of the VSL strings is certainly a step above the GPO strings. But the dynamics are in my opinion a disaster. Whereas within one range, lets say 56 - 88, the volume change is barely noticeable, crossing the line at 89 results in a marked timbral and velocity change. Of course I could go ahead and program velocity crossfades for all instruments, but the question is whether this should be my job or VSLs?? For a product with this price tag I expect it to be playable 'out of the box'. Any suggestions or answers to this problem are greatly appreciated.
Thank you,
Andreas [:'(]

[herb]

Yes, we are recording different velocities which do have different timbres,
if there wouldn't be different timbres, it wouldn't make sense to record it.

So it's up the arranger/composer to choose the right velocity out of the available options. For better controllability of the different velocities (for example if you are peforming life on your keyboard) you could use the Keyswitch Velocity mappings which are available for the short notes.

Other libraries sometimes do not offer different velocities, or do not capture the whole dynamic range, therefore you find less velocity jumps there.

If you need continous crescendo on the same key for short note from medium to fast playing, I would also recommend to look at the performance repetitions dynamics.

best
Herb

[Martin Bayless]

Herb - i know there are some language barriers in an international forum and i especially appreciate the efforts of anyone who's native language isn't english to participate in, much less host, such a forum that allows me to participate. but it appears that you may be kind of dismissing or invalidating what andreas is saying, and i'm sensing in it a frame of mind that might be limiting vsl and its users, especially in the way the VI appears to have been developed. i guess you could call it a "take it or leave it kind of approach."

i think the answer to andreas' question is no, it probably isn't practical to expect VSL to get the programming all the way there. its consistently very good, but it seems at some point, the sensitivities of the musician and the particular musical circumstance naturally need to take over. and i think any musician working in this medium who does not consider the ability to do some reprogramming an integral part of their musicianship is just short changing their potential. on no level does it seem possible that a twenty something programmer could anticipate the real musical needs and sensitivity of a trained musician with a few years of experience under their belt. yet that is precisely the message i'm reading in this post and in the development of the VI: here are the options, now choose the one that fits.

speaking from experience, there is so much "under the hood" so to speak in the pro edition, that removing all but superficial programmability in its extension into the VI doesn't anticipate the willingness and creativity of your users to help VSL reach its potential. in spite of the obvious improvements with the VI, it seems excellence is being stripped from the equation where it was once its fundamental hallmark. i truly get how difficult it can be to run a business, and especially understand the challenges of meeting new difficulties as it hits the five year mark. things change as camelot fades, but i hope criticisms here and elsewhere are serving to bring about healthy ideas for future vsl products. and i hope my comments don't suffer from my inability to do this is your native language as i would prefer; that they be thought-provoking rather than just provoking.

of course i'm asserting a certain level of self interest here, but another part of me recognizes what a gift you are offering to a wider musical community. your self interest in selling something seems roughly equivalent to me buying it, but it has always seemed that there's something more significant about VSL. it seems excellence breeds excellence, yet as it is attained, a deeper level always presents a new challenge to meet. i hope the pressures of running VSL also leave room to keep thinking towards that excellence.

[poco_a_poco]

Herb, Martin,
thank you for your replies.

> and i think any musician working in this medium who does not consider
> the ability to do some reprogramming an integral part of their
> musicianship is just short changing their potential.

Martin, I completely agree with you.

I did not post this question because of my inability or unwillingness to do some kind of programming. Believe me I had and still have plenty of that. I am a musical applications programmer and I spent four years at UC Berkelely, the Center for New Music and Audio Technologies (CNMAT), doing just that. I can easily go ahead and create all the velocity crossfades in Kontakt 2 for all instruments I intend to use. However, in this particular scenario I am writing music and I find it cumbersome having to deal with these kind of time consuming tasks. Of course I cannot expect VSL to customize their products for each individual need or application, but I did expect the instruments to be playable 'out of the box'. And if I have a crescendo crossing a velocity threshold I expect this transition to be somewhat smooth. All common sample players support velocity crossfading and I think it is not too much to ask to load NKI files that do not have these annoying and marked changes in velocity and timbre.

Best regards,
Andreas

[Martin Bayless]

sorry andreas - my comments weren't directed at you. i understand your point clearly and certainly didn't mean to imply that you were unwilling to do the editing (in fact it sounds as if VSL would do well to have you on staff). as for not wanting to do the editing, that's another story and i agree completely; who does!

[poco_a_poco]

Martin,
thank you for the clarifying comments. When I first purchased the Horizon Solo Strings I thought I was doing something wrong as far as the pronounced velocity and timbral changes are concerned. After reading many related posts, I know now that I am not the only one having this problem. I guess for the time being I just have to go ahead and do the programming of the velocity crossfades for the instruments I am using. There is obviously room for improvements with this library.
All the best,
Andreas

[rgames]

Another option is to manually draw in large volume changes where the velocity change occurs. For example, if you cross the threshold from a quiter sample to a louder sample you can draw in a very sharp volume drop at the end of the quiter note and before the louder note.

I have manually programmed a number of velocity crossfades and they work for some instruments but not so well for others. Same with modwheel crossfades.

It's a hassle, yes, but I'm not certain that there's a better way to do it. I would have liked to have seen more modwheel crossfades as standard, though.

One way that we could fix it is to program a "volume law" into the sampler, but Halion (my sampler) can't do this. Basically, you could tell the sampler to drop the volume at the crossover point. That would be nice because you would only have to do it once.

Maybe that feature is included in the VI's?

rgames

[clarkcontrol]

I believe that these marked changes in the sample's character are a sign that the strings need many more samples per note, especially considering that with solo strings there is the issue of phase cancellation if one tries to alleviate this situation by crossfading.

If VSL considers the current situation acceptable (it seems they don't since Herb just said that they ARE recording more velocities-right?), then programming is indeed what we need to do in order for these to sound appropriate.

With the new VI I wonder how much control we will have over these crossovers/thresholds.

Clark

[poco_a_poco]

Thanks for your suggestion, 'rgames'. I will try your solution with the sharp 'volume drop' - something I have not thought about. I have read some of your posts concerning this problem and you seem to understand my concern. It looks like a similar problem with HALion and Kontakt 2 in my case.

'Clarkcontrol', yes it seems that VSL considers these sharp volume and timbral changes acceptable. For reasons I have mentioned before, I do not. I have gotten satisfactory results with crossfading - about 10 - 15%, although it is a lot of work if you are dealing with a large number of instruments. I am also wondering how this problem will be handled in the Vienna Instruments. We will hear...

Best regards,
Andreas

[stevesong]

I have to concur that, despite its many beautiful and useful characteristics, the the Solo String library has a glaring and important failing in that many articulations have "stepped" dynamics with very noticeable and sudden changes in tone and loudness when one moves from one dynamic level to the next. Editing crossfades works to some extent, but, as noted, there are phase problems. As suggested, we need more samples of different dynamic levels per note and better programming so that going from one sample to the next sounds more linear. There is no question that something like this can be done as witness, for example, the "morphing" technology employed in the soon to be released Garritan solo violin library and Gigastudio 3's Dynamic Expression Filter (DEF) technology.

It is perhaps interesting - - and somewhat historically ironic - - to note that the work surrounding the invention and development of the piano in the 18th and 19th centuries was aimed at creating an instrument that had a perceptually linear dynamic response to human input similar to that of non-keyboard instruments, whereas now we have instruments that are, in reality, perceptually linear in terms of dynamic response appearing in sample liibraries with non-linear stepped dynamics characteristic of keyboard instruments predating the piano.

[poco_a_poco]

Hello 'stevesong',
thanks for your interesting comments. As I have mentioned in a previous post, I have gotten quite good results with crossfading even though there are problems with phase artifacts.

It will be interesting to evaluate the new Garritan solo violin library with this 'Sonic Morphing technology'. I assume they make use of the 'Kontakt 2' scripting engine to achive smooth crossfades between different dynamic levels. The samples on their website sound very promising.

Merry Christmas to all of you!

Best regards,
Andreas