KSP Problems Still With Us

Back in January of 2006 at the request of Markus, I carefully detailed and posted a list of 6 KSP problems (along with procedures to easily reproduce them). Of course all these problems had been reported numerous times before to Tech Support, etc.

Finally, V2.1 of K2 has been released and only 1.5 of the 6 problems have been corrected. It appears as though 4 of these problems have never even been touched.

Two of the problems listed relate to the change_vol() function and the most serious of these is the noise and pops that this function produces, even with tiny incremental changes issued in a while-loop. This function is VITAL to future script writing for all kinds of important performance tools such as legato, vibrato, and formant-corrected pitch bending. We cannot fully tap the potential of the KSP until this important function is made to work properly.

Rather than my repeating all the text of my January Posting, here's is a link to the thread.

http://www.native-instruments.de/site/forum_us/showthread.php?t=31385&page=8

The KSP problems were reported in the above thread as posts #115 and #116.

If you can't correct all of these problems, PLEASE at least fix the change_vol() noise problem.

Bob

 

I forgot to mention which of the six problems have changed. Problem #1 appears to have been corrected and problem #5 has at least been touched, the right-side cutoff has been corrected. However, problems #2, #3, #4, and #6 haven't been touched!

Bob

 

The script that crashed K2.0 in the thread linked to above causes Kontakt 2.1 to crash as well.

There's also another bug - the effect of using SET_CONDITION cannot be bypassed. Eg. the following script changes the way an instrument with release triggers behaves even when the script is bypassed:
``on init
````SET_CONDITION(NO_SYS_SCRIPT_RLS_TRIG)
``end on

How about if someone at NI kept email contact with some of us scripters to get a better idea of what the problems are? I think such a conversation could be very fruitful - no support questions, just discussion about KSP itself and how it could be fixed and improved.

Besides the problems pointed out by Bob (the change_vol bug is the one that needs to be prioritized) the combination of scripts and release triggers suffer from some bad design decisions. I described some of the problems in this post.

Regards,
Nils

 

**** Bump Up ****

Hey David, did you see this post? Never got any 'official' response. It would be nice to know if these issues are on the list of problems acknowledged by NI.

Bob

 

Hi Bob,

about the change_vol function:
We have looked in great detail into the problems you reported, but unfortunately, we are not able to solve this situation right now.

The problem is the following:
the change_vol function was never intended to be used inside a while loop. The intended usage of change_vol was to set one volume level per note event, but not generating a ramp.

When you change the volume while the note is running, it is obvious for clicks to occur, simply because of the discontinuity in the sample waveform. So therefore, the change_vol function is actually working as it was designed.

We're trying our best to come up with a new implementation of change_vol, but as we've seen, this is complicated and hard to implement. So it's not on our list for Kontakt 2.1.1. We are however keeping track of these issues, just like the other mentioned bigs.

I hope you understand why this is not possible at the moment,
cheers,
nicki

 

Hi Nicki,
Thanks for finally providing some feedback on this issue, but with all due respect I think NI is seriously misjudging how much Bob's scripts raises the value of your product. When more people discover that they can suddenly get realistic legato/vibrato for free on any sample library I think these particular scripts may become an important factor in people's choice to invest in your particular platform. Actually I think sympathetic resonance and Bob's legato/vibrato scipts are the two big players when it comes to scripts. If you translate the importance of Bob's scripts into money you can make on extra sales it could sum up to respectable amounts. It would be wise not to ignore that.

I hear what you're saying about change_vol not originally being intended for this use, but I cannot see why implementing this should be any harder than other things. Kontakt already does something similar with all its envelopes, modulation+smoothing, as well as the fade_in/fade_out functions. Actually it doesn't have to be the change_vol function as I see it. What we need is just some way to flexibly shape the volume of individual notes. A simple way to fix the problem would be to add a smoothing parameter to the change_vol function and let Kontakt do some interpolation. I'm sure there are other, possibly better, ways as well.

I hope you understand why this needs to be fixed.
Regards,
Nils

PS. I would very much like a response to this post. Thanks in advance. DS.

 

Hi Nicki,

Thanks for the reply (if not the substance). I couldn't say it any more forcefully than Nils did in the prior post. I think NI is really 'missing the boat' here. I have talked to many library developers and without exception they're hoping to use scripts to do some wonderful things. Then I always have to tell them that until NI fixes the change_vol noise problem (or provides a suitable alternative for individual note volume control), things like equal-power crossfades and other volume contouring will not be possible. These effects are sorely needed to produce the kind of sounds being demanded.

As an example, about 6 months ago I wrote a Pitch Control Engine script that was to provide smooth formant-corrected pitch bending. This script could have been the 'heart' of all kinds of other scripts for more realistic portamento. And, except for the noise, it works beautifully. But unfortunately, the noise produced by the change_vol function forced me to set aside the PCE script until such time as this problem is corrected. Now, here it is 6 months later and the problem is not corrected and worse, now we are being told that it probably never will be!

When I recently wrote and released SIPS, I used the change_vol function to add a bit of tremolo to the vibrato effect, thinking that soon NI would be releasing 2.1 with the problem corrected. I have responded to many complaints about the noise of the vibrato function by saying something to the effect of, "'Hang in there for a few more weeks, and hopefully NI will have this problem fixed. In the meantime, turn down the Depth control and try to live without it." Now, sad to say I have to tell everyone that it not only didn't happen but, that NI says it's too hard to fix!

All I can tell you guys is that a lot of users and developers were hoping that the change_vol noise problem was going to be corrected because we know that without it, we're going to have to continue to work with 'one hand tied behind our backs'. This one single problem is the biggest impediment to achieving the full potential of the KSP in real life. I think NI is making a serious mistake by not elevating the priority of this issue and then correcting it.

Bob

 

I am with Bob and Kotori here - Brainstorming, maybe it is possible instead of fixing the volume function, provide an alternative function using crossfading - even morphing (no strike this idea out, its too good) . Just lateral thinking

Zero

 

Hi Nicki,

I'm just echoing what Nils and Big Bob have said. I beta tested the Legato script that Bob wrote and can tell you it will/has changed the functionality of every library out there. Period. I know people who dropped Giga and went to Kontakt just because of scripting and this Legato script (SIPS).

There is even a website/forum dedicated to the open exchange of Kontakt scripting and progamming in general... and everyone knows about the change_vol() limitation. I'm not calling it a bug because I realize that it wasn't inteded to be used in a 'ramped' fashion, but now that you see how vital it is for the development and added success of Kontakt it would be foolish for NI to put this feature on the back burner.

Have you tried the SIPS legato script? Just curious.

Thanks for taking the time to read this and hopefully you'll be able to communicate the importance of this request to the Kontakt development team.

Cheers,

T

 

I was moved when I read Nicki's reply. Great relationship with end-users ! I'm proud of being a Kontakt-user.

I add that users in Japanese forum of Kontakt always consider the scripts written by users as the overwhelming advantage of Kontakt2 over the other software-sampler such as GS3.

I think the alternative function of volume control in scripts must build up the competitive power of Kontakt, Bob and Nils will be sure to do it with the new function.

 

Hi,

an alternative soluce to change the volume while note is playing is to assign the samples to a group, and change the amplifier value of the group with the KSP script.

For example : one group -> one key or one layer-> one group

but the numbers of group is limited and i don't know how the legato script is designed.

I use this soluce to mix wet/dry overtones resonance for Piano library.

Regards,

Olivier

 

Hi Olivier,

Yes indeed and, when we get desperate enough we have resorted to such tactics. But, in general, it's not practical to expect the average script user to have to do such extensive edits to their instruments just to be able to use a script. K2 does all kinds of smoothly ramped volume controlling at the group level and even at the note level via fade_out/fade_in, so we know it can be done. There seems to be no problem using things like the change_tune() function in a loop, it provides very smooth and noiseless operation.

While it may take some effort to either fix the change_vol() function or provide a some equivalent function that will allow individual event volume control, it would be effort well worth expending. I think the basic problem is that the people at NI making the decision on this issue just have no idea of its importance.

Let's hope we can change their mind set.

Bob

 

Another example of what we're all missing out on because of the change_vol limitation:
A script that can be loaded on any instrument to instantly provide MIDI-CC controlled crossfading between all velocity layers (you choose between equal-power and linear-gain) which it's possible to toggle on and off at any time just like the Vienna Instruments. I would have written a free to script to do that yesterday if only there had been a way to do individual note volume contouring in Kontakt.

Regards,
Nils

 

Hi Bob,

as much as I understand your frustation, I really cannot agree on this one.
I'm very impressed with the scripts you and your team (and all others) have created; it's a great and innovative achievement.
We are trying our best to come up with a new implemenation of the change_vol() function. But as I've said before, this function was designed for a different purpose, and it takes some tome to implement it differently.

To say, we (NI) have no idea of its importance, is just not correct.
-nicki

 

Hi Nicki,

I'm very glad to hear that NI realizes that this function is very important and much needed.

It would be more understandable if the problem were only recently reported, but, this problem has been with us (and oft reported) for a long time now. It's hard to believe that the problem wouldn't be solved by now if NI had assigned it the priority it deserves. Nevertheless, I think that the biggest frustration in all this has been the lack of acknowledgement on NI's part. So I must commend you for that at least.

But, since a lot of people are waiting for this function and since it would also be very much in NI's best interest to provide it, can you give us some idea of when we might expect something? Just saying it's not going to be in 2.1.1 (when we thought it was going to be in 2.1), doesn't give us much to hope for. I really want to believe that NI is going to fix this problem and soon. Am I just pipe dreaming?

Bob

 

I for one would be prepared to wait a little longer for this fix in 2.1.1. Its very important.

Thanks

Zero

 

One thing I forgot to mention, besides the hassle of assigning one note per group, is that there are some situations for which it won't work out the same as with individual event control. For example, if you are doing a crossfade legato script, how would you handle the situation where there is a legato phrase containing a repetition of the same pitch? Say you play two overlapping c3s (by using the sustain pedal) , now you have to fade-out the first c3 while simultaneously fading-in the 2nd c3. But, since both c3s are played from the same group, you don't have individual control.

To handle this situation, you would have to assign at least 2 groups to each note, and so it gets even messier, no? What we really need is event-level volume control (like that of the change_vol() function but without all the noise).

And Zero, I whole-heartedly concur. How much longer can it take to include this important fix in V2.1.1? If it isn't included, who knows how much longer it will be before we see this problem fixed if ever?

Bob


Bob

 

With all due respect, if you (NI) really had an idea of it's true importance, these problems would already have been fixed with the 2.1 update and people wouldn't have to write about them repeatedly in threads like this one.

From your comments, it almost sounds as if your programmers might actually have some real difficulties fixing these problems. However, I refuse to believe that your programmers aren't competent enough to fix these issues very quickly provided that they are given high enough priority.

N

 

"Facts"
1.Many people hope for this function working normally.
2.NI dosen't recognize volume_change() problem as a bug.
3.There are some technical issues.
4.KSP is the biggest competitiveness source of Kontakt.
5.TASCAM is going to up GS3 VSTi version any time now.
6.VSL has already sold new VI which needs no Kontakt.


"SWOT analysis"
Segment "Strength and Opportunity"
from 1&4, To develop an alternative function is benefit for both of users and NI. It's effective option.

Segment "Strength and Threat"
from 1&6, It seems too late to power up KSP from now.

Segment "Weakness and Opportunity"
from 2&4, NI will have promising benefits by confirming the lack of good volume function and starting an improvement.

Segment "Weakness and Thread"
from 3&5, People will qualify Kontakt2 as a similar product of GS3.
form 2&6, People will begin to save money for VI.

"My Message"
If NI will continue business in the domain of orchestral music, they had better commit the improvement of KSP ASAP.
Many users including me love Kontakt2, and if possible, we want to keep on using Kontakt with which we are familiar. Please don't let us make decision in the near future.

 

I'd like to write a word builder for Symphony of Voices and Voices of the Apolcalypse, which would also function as a general sound cross-fader for other libraries. I'd be awesome to have a volume ramp function that wouldn't have to be used in a loop. I'd like to provide a duration and curve shape and be done with it.