@Another User said:

Originally posted by steve@burchmusic.com
1. Should I to load a section (i.e. woodwinds) of the orchestra first, then record a section to audio, then load another section?
2. How do I manage balance and blend between various sections if I am recording to audio only one section at a time?

1. It really depends on your machine. I use Opus 1, but I can load an entire Orchestra (plus a few extras) with no problem.

2. As for recording one section at a time...good luck...[[:)]] There's no real easy way to do this. Although you shouldn't have to. If that's the case, I would recommend building a faster machine, or building a second one to load more samples.

Hope that's all you need [[:)]]...Good luck!

Steve
www.stevemazzaro.com

While I don't use VSL specifically, I've quite a number of other libraries and I'm able to get excellent results with the following (albeit lengthy) process:

I have created several templates including a sketchpad template which consists of all my regularly used instruments but has a very limited set of articulations (basically a long and a short) as well as section templates (woodwinds, brass, strings, pitched percussion, and non-pitched percussion). I compose using the sketchpad template. Once I've finished a piece and finessed it as much as possible I record each instrument onto a seperate .wav track. Then I'll load up my section patches and create more detailed versions of each part which I'll replace the sketchpad versions with as I go.

I find as long as I stay relatively true to the volumes of my sketchpad the blend works out really well. Whatever bumps there are I fix with some minor tweeks in the mix. I've only done this ENTIRE process through three times as its naturally a lengthy thing to do, but it definately works for me.

I have a similar approach....

I have a file called Origin, which basically has a track devoted to each instrument, there's a midi track and audio track for each instrument respectively, the audio is panned accordingly so once I bounce it's all set. I have them from top to bottom as follows, Strings, Woodwind, Brass, Percussion.

At the top is a master file where I keep the original piano piece. I'm also trying to figure out a way to make multi instruments of each family so I can play lines and have them doubled in octaves.

-BW

I use an orchestral template in DP. although it's complicated, i couldn't get along without it. Each instrument of the orchestra is on its own single track and i use patch changes to switch between articulations. the value is that everything is available just as it would be in a real orchestra from both a scoring and recording pov. the problem for you i think might be getting enough stuff loaded at once. if you can't it begins to obviate the strength of using a template. perhaps better to simply load instruments as you need them. at this point, i need three giga machines as well as kontakt on my g5 (to load up extra instruments like vibes or xylaphone that i dont routinely use). The problem now is i'd like to add chamber strings. not so simple though as everything is full. to use them with the template, i would have to add another lightpipe interface, expand my kvm switch as well as buy a fourth giga machine. the point is templates are great but especially with vsl, there are definately diminishing returns. getting 80% of it done is easy, but the last 20% is not only where things get really cool, it is also where things become combersome and expensive. you're on the right track to get as much advice as you can and think through your needs and resources carefully.

The method of working can be constantly evolving, as you refine your setup and experiment new ones.

My suggestion is to divide the process in several steps, and create setups for each of these steps.

First comes the composition step. I leave it to you to decide how you are more comfortable. I suppose it happens between you and your instrument.

Then comes the orchestration step. For this, you should load the sampler with all the instruments of the orchestra, one instrument per channel. The VSL programs you choose must give you access to different playing styles. In your sequencer, you can do the orchestration like you would do with any sound library. For now, there is no audio involved, only midi.

The next step is the interpretation step. This step is, for me, where the strength of VSL resides. Load the sampler with all the needed articulations of a single instrument. Set the output of the midi channel to any (create another midi track if you need to load articulation on more that one port in your sampler), and send the notes to the desired articulation. There is a tutorial by Beat Kaufmann that you should read.
For this step, you can create a template for each instrument, but you may want to modify it to fit your current needs. For example, you are working on the violin track. Are loaded all the articulations, but you need only the leg-perf instrument, the dyn-all and dyn-spec instrument. Then you can get rid of the pizz and tremolo and other articulations. Also, you may want to load the viola and cello instruments of your orchestration setup in case you want to hear the violin with them. I personally don't have templates for this step.
Once you are satisfyed with the interpretation of your melodic line, bounce the audio of this instrument only, without any effect (like reverb). Place the audio track in the sequencer and mute the corresponding midi track.
Do that for every instrument.

Now you have all the audio files. If you room is not calibrated with your sound system, I don't recommend you do the mix yourself. You would amplify frequencies that are absorbed by your studio, and diminish frequencies that are too present. This is true even though you don't touch the EQs. The simple fact that you play with the volume of the instrument is a form of EQing. Also, I recommend that you use a convolution-based reverb.

I hope this helps.

Bertrand

I am just starting out with VSL and I also use DP. In the real world I play all sorts of drums/percussion and also guitar but I am definately not a natural keyboard player!

But I find I can still play reasonably fast parts in with a keyboard so I can get the timing, tempo feel, and note velocities right - even if I get every other note wrong or hit two keys by mistake etc! .... I play just to get the performance -allowing any other mistakes to happen (it feels wierd to start with! a bit like Les Dawson!!!!) ...then I just go back and correct the mistakes and just tidy things up in the sequencer - but the *performance* is pretty much there.

Other routes I have found to achieve human/musical performances is to use DP's 'record beats' feature - it basically allows you to (re)record the tempo via a midi source (keyboard or midi drum pad). You just tap on every beat thereby setting the exact tempo between each beat. You can do this before, during or after you start entering midi data.

So you could enter everything by mouse, all notes aligned boringly to the beats of a fixed tempo, then record a new tempo map for it - an actual 'conductor's performance' if you like (well, tempo-wise at least). You could record and save as many as you like and choose the best one later on.

And, similarly, if you record a version of your piece on your preferred real-world instrument of your choice with any old microphone - you can then use it as a guide while you use 'record beats'. Now you have a midi tempo map which mirrors the *performance* you made with your real instrument (watch the BPM display - you'll be amazed at just how much it varies- even for a piece which is basically the same tempo throughout!) So now even if you mouse in all your notes so they are exactly on the beats (ie quantized) or copy and paste the same data 100x in a row at least it will follow the 'conductor track' and already be starting to sound musical. Or if you play in notes on the keyboard you will have a click track to play along with which already has all the 'pushing' and 'pulling' of tempo naturally in place.

Also I have a mic set up next to me so if i have a melody in my head I can record straight away by singing it (badly) .... I'll do a few overdubs/ drop ins until it works and then it's there ...ready to move a midi track underneath it and mouse in the notes copying the performance as it goes....

....Or even - in DP4.6 you can use the new 'pitch' feature in audiotracks where DP will analyse any (monophonic) audio and give you the pitch data (just as in midi) which you can copy and paste into a midi track- voila!

..... OK it's a bit 'off' sometimes but it's quite a nifty feature. And really useful for me as a guitar player because I can record a whole set of really... um....'exotic' chords on my guitar in odd tunings (sequentally one note at a time) and they're instantly midi-fied ready to align, then copy and paste wherever I want in my composition for strings or piano etc ..... whenever I try and play unusual guitar chords from memory via the keyboard I *always* make mistakes and only realise later on when it sounds all wrong. [[:|]]

So there's a few things I do, anyway! Hope it makes sense. Of course we are all making different kinds of music and need to do it different ways! .... but I think the key (and the joy!) to working with VSL (and samplers + sequencers) is to be imaginative and fluid in the way you work and to not tie yourself to any single set of proceedures.
[:)]

Edit: Yes you were really asking about templates!! sorry I was just reading another thread asking about about 'playing or clicking in midi notes' (anyway, hope the above is useful to you as another DP user!).....Um... yes I am already making templates -with Kontact2 I can save the sampler states (programs loaded) with the templates which helps a lot. Lots of templates I think is the way to go!!!!.... quick scratch pads, strings, horns, mini orchestra, favourite ensembles etc

Also, remember in DP the 'Clippings window' which you can use to hold midi data, settings note, composition notes, soundbites, anything useful really and have it 'to hand' in a project or transfer easily between projects (you can make a clip

see post above- ran out of space: ..... Also, remember in DP the 'Clippings window' which you can use to hold midi data, settings note, composition notes, soundbites, anything useful really and have it 'to hand' in a project or transfer easily between projects (you can make any clippings window project specific or open with every new project).

-that's all!

@Martin Bayless said:

I use an orchestral template in DP. although it's complicated, i couldn't get along without it. Each instrument of the orchestra is on its own single track and i use patch changes to switch between articulations. the value is that everything is available just as it would be in a real orchestra from both a scoring and recording pov. the problem for you i think might be getting enough stuff loaded at once. if you can't it begins to obviate the strength of using a template. perhaps better to simply load instruments as you need them. at this point, i need three giga machines as well as kontakt on my g5 (to load up extra instruments like vibes or xylaphone that i dont routinely use). The problem now is i'd like to add chamber strings. not so simple though as everything is full. to use them with the template, i would have to add another lightpipe interface, expand my kvm switch as well as buy a fourth giga machine. the point is templates are great but especially with vsl, there are definately diminishing returns. getting 80% of it done is easy, but the last 20% is not only where things get really cool, it is also where things become combersome and expensive. you're on the right track to get as much advice as you can and think through your needs and resources carefully.

Martin,

Perhaps my question to you will help in this discussion. I have my main DAW (running Kontakt and some VSL patches) as well as 2 seperate PC's running GS3.

I prefer having my template have ALL the sections and as many art files as possible. I have each instrument's BASIC patch (e.g. for VL-14 I'll have 7 midi tracks set up with Stacc, .3, .5, etc.). I do this for all instruments. Of course I don't have any perf patches (eat up too much Ram - have 2 gig machines all around).

When the piece is finished (compositionally) - I'll then go in an replace some of the parts with Perf patches.

Hopefully this is clear. How does 'patch changes' make for a better way? Wouldn't I have to load up whole gigs that would take up a lot of Ram?

Eventually doing to get another PC, but perhaps you can help me optimize what I have now.

Rob

Hi Rob -

Patch changes were the only way to accomplish what I wanted to do. If your goals aren't the same, they might not necessarily be better for you though. Of course I can tell you how I've got everything set up, but it might be helpful to know a little more about what you're trying to get done to offer many ideas.

For example, I loved Andy Blaney's Debussey mock up. When I downloaded the midi file though, I about had a heart attack. It was simply too uncomfortable to think about having to create something fragmented accross so many tracks. What I wanted was to notate and sequence within the same program with one track allocated to each instrument. no multiple tracks unless, like horns they would normally be notated on two (in which case two tracks would be fine). You said you had seven violin tracks. Are you okay with that, or do you feel a need to consolidate them?

Obviously, it would be helpful to have your performance instruments available when you need them - which suggests two things. First, Matttias' memory tweeks if you haven't already done them. And second, building your own instruments the way you need them to maximize ram useage. I didn't understand the question about loading whole gigs though. Did you mean the .gig files themselves, or whole gigabytes of data? Are you thinking about adding a third machine so you can load more data, or do you use something like gigapulse that takes up part of your resources (which could be replaced with another convolution reverb in your sequencer)?

This sounds like exactly where I was before adding the third machine, so maybe my experiences will indeed help. We are definantly on the same page about having all the instruments and articulations loaded up at the same time. The third machine did that for me with the performancee instruments, but not without careful ram planning. Even then, there were some tradeoffs. If you could clarify a little, we'll see what we can come up with.

best,

Martin

Hi Martin,

Thanks for the response. I think I have the tweaks giving me as much of my 2 gig ram as possible. WIth 3 instances of GP running on both machines I really never go beyond 50-60% CPU with notes playing.

When I load of a 'Basic patch' I essentially put that same 'basic patch' on seven different tracks within GS3 (no hit to Ram if I do this within GS3) and then cooresponding midi tracks in SX. If I have to write a string line I'll do it on, say, the sus patch and then copy the track to one or more of the other tracks - depending on the articulation needed (first note is stacc, second, third - .3's, etc.

Seems to work good and usually sounds better than one patch playing the whole line.

I was curious as to 'patch changes'. How does this help Ram usuage?

Thanks for any help you can be to further optimize the available patches I can have at my 'disposal' at any one time (the third PC will help, but I sense I will still want as much loaded as I can.

For reference, here is a quick diddy I did with the new template ONLY using cube for writing (using Basic patches). This is straight out of the board with no EQ, mastering - really just an exercise. If this was a real project I would take some of the WW lines and use a perf leg patch, etc.

It was good to have as many articulations available as possible.

http://www.robelliottmusic.com/mp3s/Misc/Playful%20Deception.mp3


Here's a second one using only basic patches. Sorry for the roughness of the mix.

http://www.robelliottmusic.com/mp3s/Misc/Memories%20of%20Her.mp3





Thanks.

Rob

Thanks to everyone who offered advise on setting up DP with VSL.
Here's something I had not known about before in DP...
You can use the split notes command in the regions drop-down menu to move notes to other tracks. As an experiment I played a short violin passage using a variety of short and long notes. Then using the split notes feature I was able to take all the 16th notes and place them on the vl 14 stac1 track. After that I used the feature to select all the 16th notes of selected pitches to be moved to the vl 14 stac2 track. In this way I was able to get up- and down-bow articulations. I did this for other notes, moving them to dim or cresc tracks, longer notes, etc. The effect was pleasing, but I'm still very much in the evolutionary stage. I will keep watching this thread for more comments.
Steve

hey Rob - terrific music!

i can't see where patch changes offer any advantage or impact ram. That said, optimizing instruments that a patch or keyswitch, etc. might change can have a significant impact by focusing on what you might need and eliminating what you dont. instruments can usually be reduced in size depending upon what's important to you. some people have reduced everything to mono which will double the amount you can load. i do this on my laptop along with whole note mapping which doubles the load yet again. These are audible changes so i dont use them on my main system, but perhaps it would be efficient to create a set of instruments like this to sketch with, then load up just the full instruments you need to render a final performance.

I take advantage of other trade offs on my main system. for example, having all legato string patches available along with the ability to cross fade between layers is important to me. But with ensemble brass and winds, the advantage of more layers seems to outweigh the benefit of the larger legato instruments. the opposite seems true of the solo instruments which can stand out of the texture and have phasing problems if layered. Similarly, I couldn't make the case for four or fiive levels of pizzicato articulations, so i reduced them and added more repetitions to play in round robin (to avoid having to use the repetition or alternation tools). Actually, i've done this with almost all the shorter articulations. I also found limited usefullness in the intermediate length articulations. either a note is short, or a longer articulation can be cut off early without much detriment in the real world. in fact anything i can do to create the articulation thru performance technique (rolls, glisses, swells and even some sfz - not strings though) can save ram and not be detrimental if done musically. Some may not agree, which raises the valid point that the trade offs each of us makes will always be very personal and adapted to our unique sensitivities.

As for load limits, my machines all run at about 62-63%. Using GigaPulse would reduce that, so I distribute that processing power instead to the G5 using Altiverb. Obviously, that cost a little more money, but perhaps in the realm of what a fourth machine might cost. This is important because the camparison betweeen our systems might be an apples and oranges one. what can be acheived with three machine plus altiverb, might take four machines using gigapulse (for about the same money, and neither being inherently better).

Obviously, not the most direct answers to your questions, but perhaps something you can extrapolate from as you think through how you want to grow your system. just let me know if i can help, and thanks again for putting up the really cool music (i actually listened to several things on your site and would encourage others to do the same!).

Thanks Martin - lots of good ideas in here. I am definitely working towards adding a 3rd PC just for VSL. That will help.

Both existing PC's have 2 GP instances loaded in the template, but really during 'writing' I can just pick up my verb through my Roland SRV-330 (through the monitoring board).

Makes me wonder if I didn't have these 2 instances how many more patches I could load up. Something I'll experiment with for sure.

Thanks for your nice comments about this short short diddys. Forgive the very rough nature of the mix.

Rob

@Rob Elliott said:

Thanks Martin - lots of good ideas in here. I am definitely working towards adding a 3rd PC just for VSL. That will help.

Both existing PC's have 2 GP instances loaded in the template, but really during 'writing' I can just pick up my verb through my Roland SRV-330 (through the monitoring board).

Makes me wonder if I didn't have these 2 instances how many more patches I could load up. Something I'll experiment with for sure.

Thanks for your nice comments about this short short diddys. Forgive the very rough nature of the mix.

Rob

I don't think thaqt unloading GigaPulse will make any difference to the number of instruments that you can load, but it will make a difference to the CPU useage.

DG

Thanks DG - just hoping for more ready access to patches / instruments. Can't wait to get my 3rd GS3 system up and running.

Rob

@Rob Elliott said:

It was good to have as many articulations available as possible.

http://www.robelliottmusic.com/mp3s/Misc/Playful%20Deception.mp3


Here's a second one using only basic patches. Sorry for the roughness of the mix.

http://www.robelliottmusic.com/mp3s/Misc/Memories%20of%20Her.mp3

I'm really impressed with the beauty of these works and the realism you achieved.

Could you tell me how do you made the background noise?

And, my main doubt is, what's the best GigaPulse configuration to create a real placement for orchestral instruments?

Could you tell me how did you create this configuration?
[8-)]

Thanks in advance.

Rogerio.

Thanks for the reminder, I (unfortunately) missed those when originally posted! Wow, Rob, really good for some "exercises"...

Thanks Polar Bear for your nice comments. This really is a wonderful library - the more 'access' you have to at one time - the faster the writing flows. Should get my third PC today - really getting excited!!!!


Hi Rogerio,

The background noise was from Garriton's GOS. I think that Belamedia has a free one somewhere on NS (do a search there.)

As far as Giga pulse - for writing and 'rough mixes' like this right out of the board, I simply set up 2 GP's - one near mics and one far. It is good enough for general placement while writing. For final tracking you would have to set up perhaps more placements - but CPU could be a hinderence here.

For all but the biggest projects I will just use the two GP's and of course do a REAL final mix (cut some of the mud out of certain instruments, etc.), then VERY light compression in the mastering stage.

What I have found for me is that the better I can get the arrangement and orchestration the less I need to work in the mix stage.

Hope this helps.

Rob

Excuse what looks like obnoxious self-promotion, but in our first issue (Virtual Instruments Magazine) composer Klaus Badelt explained how he had his pallet set up in Logic when he was using 18 GigaStudios. Christopher Caoette followed up in a letter in our second issue, which should be online for download later today.

The basic idea is to create an Environment layer with MIDI Instruments (which are named and can be assigned icons - pictures of the instruments) that point to various MIDI ports and channels (connected to external Gigas or other MIDI modules). Command/clicking on the icon assigns it to the currently-assigned track. That means you can have billions of articulations loaded without having each one permanently assigned to a track. You can also move the icons around the screen wherever you want, and also put them in a graphic box so all the violins, etc., are together.

I posted about this here a few months ago.

Oh, and to each his own, but I personally hate using program changes. I don't want my studio reconfigured on the fly every time I go back four bars.

Can we buy this mag/issue in the UK,Nick?That sounds very interesting,and I'd like to get hold of a copy.