Thank you for the comments on my music, Paul.  It is appreciated.

No, I do not use a DAW.  Even though I have purchased all the Cubase versions since version 2, I do not use it for Vienna.  Ever since I was told that I needed to keep dual versions of my music (score and performance), I have disliked it.  I found it very annoying to make changes in one and have to reflect that change in the other. I gave up on it.  With the capabilities of Sibelius now, I prefer to use that for my VSL work.

I predate computers and even calculators. (For math, I used a mechancial slide ruler or manually worked out the answer.) My first computer cost $800 and for that and a week of soldering parts on to the circuit board, I got the play with patterns on the light-emitting diodes.  I seen computers come and evolve in capability. I would spend hours counting machine clock-cycles to make my programs run faster only have it replaced by high-level program writing software.

I do not mind those that spend their time playing with alternate samples and fine-tuning DAWS to make their music sound better.  They are laying the ground-work for the programs of the future.  I liken their efforts to me counting clock-cycles.  I could spend the time learning, but in five years it will be replaced by a program or plugin that will do a better job in a fraction of the time. I would rather compose.

thanks for this interesting contribution, Mark. I listened to some of these samples and at times I felt there might be a bit more richness than with the Special Edition alone but in general the sound was rather on the harsh side. I'm still rather agnostic. Also, I'm coming round to the idea that you need VelXF for a typical tring quartet, otherwise the sound is simply too bland -- the biggest probem being a real muddiness in faster passages. On the other hand, setting the VelX slider on maximum gives a real brittleness to the violins in particular so I've decided on around 90 for the violins and nearer 100 for viola and cello which is the best compromise I've found to date and guarantees that slow sections at least are genuinely expressive.

You can see  an example of my efforts here if interested. I suggest starting at 14.20 for the short scherzo which I think is probably still better with Wallander but certainly an improvement on my previous attempt or straight to the Cavatina (yes, exactly --that one!) around 17.00 for lyrical expressiveness.

https://app.box.com/s/rbqotm0dou76u9dk911p8hvfounhkfnh

David

@lunar_28664 said:

I'm coming round to the idea that you need VelXF for a typical tring quartet, otherwise the sound is simply too bland -- the biggest probem being a real muddiness in faster passages. On the other hand, setting the VelX slider on maximum gives a real brittleness to the violins in particular so I've decided on around 90 for the violins and nearer 100 for viola and cello which is the best compromise I've found to date and guarantees that slow sections at least are genuinely expressive.

Aside from keyboard and plucked instruments, you need VelXF for just about everything including strings.  You also should use it in conjuction with the Exp fader as well to really bring out the detailed nuances of the instrument.  Also you should be making subtle adjustments to Attack while the piece is playing in real time.  If it's too much it tends to cause the muddiness in faster passages that you might be hearing.  Too little makes things choppy.  The release fader I set to about 70 and leave it alone.

But the real kicker for solo strings, or any solo instrument for that matter, is tunning.  Ever so slight manipulation of the tunning fader (15 to 30 cents) can pump life into your MIDIstrations.  In addition to increases and decreases in tempo at appropriate parts of the piece; slowing at the end of a melody line that usually falls on the root note for example. 

If you have VIPro you can achieve a lot with the humanization function and time offset.  But all of these tools are useless if you don't know how to use them and learning the craft of MIDIstration takes blood sweat and tears I'm afraid.  It's just like learning an instrument.

@jasensmith said:

Aside from keyboard and plucked instruments, you need VelXF for just about everything including strings.  You also should use it in conjuction with the Exp fader as well to really bring out the detailed nuances of the instrument.  Also you should be making subtle adjustments to Attack while the piece is playing in real time.  If it's too much it tends to cause the muddiness in faster passages that you might be hearing.  Too little makes things choppy.  The release fader I set to about 70 and leave it alone.

Not everyone seems to agree on VelXF, I get the impression, though I'm with you in general. And you can't make subtle adjustments to attack when a piece on playing in real time unless you play through a DAW -- the initial discussion was partly if improvements can be made using notation software without going through the "blood, sweat and tears" of MIDIStration. It's looking like the answer is "no"!

I use Finale for notation but I concede that I don't use it with VSL.

Are you saying the Vienna Player, or VIPro won't work in real time with Sibelius?  Well that's a crying shame😢   I suppose you could map CC controller changes after the notation but with only 128 values to work with in MIDI that would be a lot of trial and error. 

Hello jasensmith!

Our instruments can be used with Sibelius and with Finale in real time.

Best regards,
Andi

I may add the to my posts to say that I do use SE and Jazz Drums to create Jazz backing tracks in Cubase to practice improvization. My use of the Vienna libraries has created an amazing learning envoronment.

May I join in on this thread, please, having read every post carefully twice over the summer and autumn, during which time I've completed a string quartet using Sibelius 8 and the Solo Strings packs?

I can see that I have the same concerns as you, David; and that my work sounds uncannily similar to your files, Paul (thanks for posting!).

So much so that - just as, say, Janet Baker's voice or the Amadeus' string sound are recognized immediately for their timbres or textures - the VSL/Sibelius sound is unmistakable.

This isn't a bad thing. And I'd say we have to be realistic about what a notation program like Sibelius (and eventually Dorico) can do.

I think I'd settle for knowing that I can, though, consistently produce the best possible.

And both Andi and the documentation (especially v. 3.2 of the 'Optimizing Sibelius Playback with Vienna Ensemble' pdf) are both helpful, of course!

But there are indeed so many combinations of so many variables in the patches and presets etc that even something simple - perhaps in binary tree form to help make the necessary either/or decisions and eleminate wrong turnings - for, say, solo strings would be a great help. Also some way of knowing which tweaks are additive and which cancel one another out, and/or changing which settings achieves nothing without some other parameter's being altered first.

Similarly a list of changes to experiment with - but one that put each parameter in order of likely effectiveness… "x will make the biggest change because it does y" and "don't expect to hear greater realism with a until you've balanced b and c" etc!

For instance, I find the violins have a thinner - and often noticeably 'whinier' - sound than the cello (which is invariably excellent); and that the viola - because it's a quieter instrument, perhaps (?) - tends to sound growly in its lower register, and insipid above C.

I do believe all this is only a good starting point; and would really welcome specific/concrete guidance… I know the samples are capable of great sound.

Another data point: I use Sibelius and VSL, without a a DAW. I have considered using a DAW, but I don't have the patience. I would much rather write music than produce it. (Some of my music is here:  https://soundcloud.com/mhcoffin. The string stuff uses VSL solo strings full library.) I am completely certain that If I spent a lot of time messing with a DAW, the simulation would be better, but it isn't bad, and life is short. In any case, I doubt it would sound as good as it did last week when I heard the Acadia Suite played by real people. :-)

I agree, Mike!

(I started with Logic and quickly realised that I was more interested in notation; and that that was where I wanted to put my time and energy: not least because working that way (in my case, so far, Sibelius) taught me so much more about serious music and music theory.)

That's why I'll happily settle once I know that I've got the best I can - even though that may not be 'real' playing.

One has to realize that using samples involves bypassing everything that performers do.  

The sample user has to become a conductor, but even more - he has to become a player to a certain extent.   Conductors actually have it much easier than the sample user.  They can stand in front of a great orchestra, wave their arms, and beautiful sound comes out.  

With samples, you don't have that luxury - you have to work harder.    Because you are doing a lot of what the players do, with expression, and what the conductor does, with controlling all the elements of the perfomance. So there is no easy solution - it is pure musical work, figuring out what sounds best on each line of your composition.  

Thanks, William - that makes good sense.

Which types of work could those of us who want to 'conduct the samples' (for I think that's really what you're saying is necessary) undertake in order to get sounds approaching those - presumably - of the time when the players' own work was itself captured in samples?

I'm confident that, if I knew which variables to start with, given a simple (= controllable) (Sibelius) score, I could make headway.

Any specific suggestions, bearing in mind uour wise observations, please?

TIA!

VSL announced a new offer for solo strings today. It has two samples for cello only, so far… violin to follow.

And, aside from a certain roundness - and absence of reverberation - my own writing for VSL cello in the Solo Strings I pack is not that different… encouraging.

This does prompt me to ask: could the degree of wetness/dryness (I use MIR Pro), overuse in either direction, be responsible (in part) for any artificiality of tone… a sensation of synthesis?

I know this can be done 😊

My first forays into realism began with an experiment to slide up and down (actually left-right, of course, acorss) the wet-dry (reverberation) spectrum.

Although I have MIR Pro inline in all my VEPro 6 channels, I think that an easy mistake is to try and accentuate 'realism' by having too much 'atmosphere' in spurious echo.

For the moment, Dry is more real-sounding.

Anyone ese?

Two other parameters which make the Solo Strings Viola sound less realistic appear to be:

1. volume/dynamic:the softer the more pleasant, and
2. pitch: above C4 it sounds harsher (more like a saxophone!) than B3 and below

I wonder whether Andi and the wonderful VSL experts could help us here, please?

thanks, Mark, and  the others who have made recent contributions -- just spotted these now. I must say I'm coming back again, one year on, to my original question as to whether it's worth investing in the solo strings vol. 1 above the SE set for use in notation software and find I'm as confused as ever!

As the number and nature of articulations are (or from the list seem to be) much the same then any improvement must be down to 1. a far larger number of velocity steps and obvioulsy samples 2. smoother and more organic programming of the articulations. Is there anything else? I have made some headway in getting better sound from the SE.  Again, it would be very useful to trial the full package as that would answer my most important questions within hours. If this really isn't possible, has anyone here used both and can tell exactly in what way Solo strings vol. 1 is better than SE or would even be able to demo an example of both in a string quartet (or similar) using best practice in both cases?

The situation in notation software has now of course changed and Dorico is likely to become the programme of choice for many, especially when at least the initial playback features are fully implemented as this should make it easier to get the best out of VSL.

David

David,

Yes - and thanks for keeping this thread going! These are important matters, aren't they.

The solo strings have more articulations than the strings in the SE do (I have both), of course.

But I believe the ways of controlling and using them are essenmtially the same.

I too would welcome any and all guidance as to how to make the most of the various arrays of combinations of which they are capable.

another update after working quite intensively at creating new scores out of earlier string quartets which originally were simply recorded into Cubase. The patch which most often makes the ghastly screeching noises so often heard is the default "normal" or sostenuto one. "legato" should be used whenever possible for expressive music.  With the Special Edition, I 've found that provided there are no long notes then substititing "rep" can work for many figurations and tends to give a cooler less vibrato sound. In "detache" the vibrato is often too strong and "short" is simply too short unless the music is fast and metronomic in which case it often works well. "spiccato" has more character than "short" but only very short notes. Looking again at the full articulation list for the Solo Strings 1, the control of vibrato, among other things would surely give more scope for improvement without much extra effort.

Vel XFade mostly doesn't work for slow music with solo strings in SE becuase there is simply too much vibrato and other oddities (though I might try yet again combining with Expression in different ways to see if there is something I might have missed -- again any suggestions welcome). What I somehow missed earlier from Andi's guide and is potentially very useful is the ability to write Vel XFade on/off commands directly into the score in Sibelius. Fast music can get a wilder, more spontaneous rendering and there are probably also individual passages in more meansured sections as well which might work. It is starting to add up! Here is an example of one of my recent endeavours -- still much to be done but I feel it's really coming along now.

https://app.box.com/s/pnopcvxml3cjg5v6a6cv34i55g47m0i5

David

Thanks so much, David; this is extremely helpful.

I have a growing 'Help' file of my own. I have added all your comments to it.

I'll try them out.

How exactly do you get 'rep' to work, please?

Are we approaching a stage when we might even venture a (collaborative) FAQ somewhere - perhaps text, explanations and sound samples?

I'd be keen to help/participate.

Would Andi jump in to support?

Hi Mark -- glad you found my observations of some interest. "rep" is just an abbreviation for repeat and like all the other articulations mentioned is just called up as a Sibelius technique (CTRL+T). "rep", which is basically short for legato repitions is supposed to adapt itself to the speed of the music being played thought the finer points elude me so far. I expect you've already seen Andi's helpful "Optimising Sibelius Playback  in Vienna Ensemble"? However, it could do with a more detailed explanation on the thorny question of velocity XFade. For instance which setting change what and how best to combine, if at all, Expression with XFade. The manual is little use as it does barely more than just list the options.

It's possible that we might together have something to contribute in terms of a help file but I suspect most of it must have been covered elsewhere though not always easy to find. I find I am too uncertain in many areas and often opinions vary considerably as to which are the best techniques with someone like Beat clearly knowing exactly what he's doing.

David