Rob,

Since we have the exact same conclusion I don't know what I can offer except that the 64 thousand dollar question is when will the 64 bit answer come. The issue will go from ram limitations to polyphony and system/drive/midi performance issues. It will certainly break the back of the current sense of futility having such a capable library as VSL and not the resources to effectively or efficiently use it. I'm with you and the waiting game.

Paul suggests another wonderful solution as to the next generation of G5's ending the limitations of different programs running full force on the same machine. So we are awaiting a new technology. As my Dad said many years ago when I asked, "When are we going to get a color TV?" "When they come out with the square tube."

DC

@Another User said:

Maybe the next G5 will allow us to run either giga3, kontackt2 etc all at the same time. Or, maybe not. I don't know.

I don't think the next G5 (Mactel) will change anything for Gigastudio, it's always MacosX, a completely different OS than Windows, especially for a software written at a kernel level.

@Another User said:

Maybe the next G5 will allow us to run either giga3, kontackt2 etc all at the same time. Or, maybe not. I don't know.

I don't think the next G5 (Mactel) will change anything for Gigastudio, it's always MacosX, a completely different OS than Windows, especially for a software written at a kernel level.


Great info to have. In the world of PC's 18 mos is a lifetime. Looks like I better give a 3rd PC for giga some serious consideration.

Thanks again.

Rob

This is a very important question because it is based on the current limitations of hardware. Obviously the more computers available the better. This is probably no help since you are trying to create an ideal template, but I gave that up and instead am trying to be practical in a dirty way. I have been rewriting templates for each individual piece, rather than creating a general orchestral template. Also, I use the mod wheel string patches a lot since they are very expressive, and relatively low in RAM useage. I stopped trying to be "strict" in writing legato, because it actually doesn't sound as good as being very loose and doing a lot of "fake" overlap legato combined with a few instances of the real thing. And you find you don't need all the legato instruments depending on the piece in question. For example - basses. Honestly, do you always need legato basses? I rarely do unless it is a soli. Also, I always use mono solo instruments which I know are frowned upon by some people, yet I think sound correct in relation to ensemble imaging and are of course half the size.

Anyway, it is interesting to hear various approaches to this very basic problem - a very useful thread!

@Rob Elliott said:

Hi there,

I love the perf patches - across the board - but with only two giga machines dedicated to my Pro Ed I find I am always for want of more articulations on my Orchestral Template (but run out of room on the 2 GB Ram machines). Recently I have dumped all the perf patches and have for all instrument groups their 'Basic' keyswitched one - which usually has stac/.3/.5/sus/sfz/pfp/trem/pizz (strgs). Seems I have more 'at my disposal' during the writing stage.

I then, when needed, go back and tweaked various lines / sections with perf patches (in the endless midi-tweak stage of the production.)

What are others ideas for the most workable 2 giga PC's template of VSL Pro (have Oboe 2 and a couple other Horizon products as well.)

When I am in the writing stage I really don't like loading and unloading tons of patches (kills the workflow for me [:(] ).

I would appreciate your ideas.

Rob

Rob - a quik fix to effectively double the number of machines you already own, assuming they are upto it and assuming you have GS3, is to install VSTack along-side GS3 and use GS3's rewire mode. I have done this on 3 of my machines so far, each has got 3gb ram onboard, 1gb used for GS3 and 1.5gb for VSTack (running a couple of instances of Kontakt2).

On one of the PCs I get slight crackling if I do anything too heavy with both samples simultaneously but that one's got slower drives and is an older motherboard/cpu. I can fix it by increasing the latency to 12ms. The other 2 are Athlon64s with Raptor drives, and can handle both samplers playing heavy orchestration at very low latency.

Only thing to watch is that you load up your GS3 tempalte first, then launch VSTack and load up your kontakts, otherwise GS3 can't use the full 1gb. GS3's audio outs appear in the VSTack mixer (a little louder than standalone but the latency stays lightening fast) and as long as you are carefull assigning your midi outs so that each sampler has it's own set everything runs smoothly. I load up both samplers templates each morning and never have to touch them.

Ian

Thanks Ian and William for your input. Improving my workflow is a constant effort. I am almost getting anal about it. I realized the more efficient the workflow (less tinkering with PC's and adding/deleting patches) - not only more music is written but it is of better quality (that whole deal of 'inspiration' seems freer and in better supply [:D] )


Thanks again - and to all - keep the ideas a comin'. This is something that all of us can use.

Rob

William...regarding the mono instruments... I find it quite appealing to be able to load twice the amount of samples, even if it means a slight compromise in sound quality.

Can it be done with Legato instruments? Are there any pitfalls? Any instrumemnts not to attempt this with?

A couple questions.

How to make, say the WW's (FLT, OB, FA, etc.) into mono instuments? I would love to do this. I have the 'stereo image' on these tracks quite narrow anyways. I am sure, I wouldn't hear a degradation.

With the size of this library and others I have, I don't have much more room to make up new instruments and store them on my HD.

Can I make these mono instruments without adding to the HD?

Rob

The mono instruments are not a compromise at all in quality, just mono. I know some people don't ever want to go mono, but I think it sounds better to use a mono solo if it occurs in an orchestral ensemble. It is closer to what happens in live, with the sound emanating from one point from a single instrument, while surrounded by the stereo ensemble instruments. Also, there are no image shifts due to the player moving around. It is very straightforward to convert to mono in the GS editor, just by clicking on convert wave pool format.

So, William, does that mean the answer to Rob's question is, "yes, you have to add to your HD -- you have to convert and save each instrument?"

Not having played around enough with GS editor, is there a way to just reference the samples instead of duplicating them in this process? That is, can it act more like a .art file, which doesn't actually add anything to the HD, just how GS handles the file?

When I choose "edit" and "convert sample pool" I figure that I shouldtick off 'Convert stereo to mono'... what about the 16bit, 24bit option... which of the 16bit should I choose? Integer? Accelerated?

No, what I was talking about is simply creating a new mono instrument that does add to the hd. I deleted the stereo instruments after this process, and saved a lot of hd space as well as ram. Except for ones that I might also want in stereo such as harp in a chamber setting, or an instrument that might be featured in a concerto-like context.

Hi William.. did you see my question on the previous page?

Also if you put these mono instruments into a gigapulse envioroment.. wont it minimize the need for ensembles to be stereo?

On those options, what it defaults to should be left alone.

No, the ensembles must be in stereo or they sound extremely artificial. As if they are inside a phone booth. Even the small ensembles like Chamber Strings, which might be collapsed but not mono. With the mono instruments mixed in with the ensembles, you set up a contrast of imaging with the stereo that sounds very good, especially when put into gigapulse. The multi placement works best on the ensembles. On the solo instruments the single placement works best.

@William said:

On those options, what it defaults to should be left alone.

No, the ensembles must be in stereo or they sound extremely artificial. As if they are inside a phone booth. Even the small ensembles like Chamber Strings, which might be collapsed but not mono. With the mono instruments mixed in with the ensembles, you set up a contrast of imaging with the stereo that sounds very good, especially when put into gigapulse. The multi placement works best on the ensembles. On the solo instruments the single placement works best.

William, sorry for yet one more question (you have been very helpful). You say mono instruments work better in single placements. The challenge for me is CPU power. For each one of the these giga machines I have three seperate GP instances (near mid far) to put whatever instruments/section into. Has worked great on my template. I don't think, that if I make the WW's mono, that I'll have enough CPU to add a 'fourth, fifth, etc GP instance for the single placement GP's.

See my problem. Can I just route these new mono patches to one of the three existing stereo GP's?

Thanks in advance.

Rob

@William said:

No, what I was talking about is simply creating a new mono instrument that does add to the hd. I deleted the stereo instruments after this process, and saved a lot of hd space as well as ram. Except for ones that I might also want in stereo such as harp in a chamber setting, or an instrument that might be featured in a concerto-like context.

hmmm.. default its set to 24bit... which I know isen't right. How bout percussion.. can that be mono? And is there a faster way of mass-monoing(tm) rather than open each instrument and then convert it individualy.

And thanks for your assistance... much appreciated

My default is 16 bit integer which definitely works. No, there is no way to convert whole groups of instruments - I also wanted that, and remember Translator does it. So you could use that instead, but I ran into some problems with it so just used GS editor.

Yes, that is a problem with the CPU being taxed by all these instances and you're right, trying to place each one with single instrument placement is huge in its demands. But it the only way to get the pure convolution imaging. You don't want to mix dry signal at all. So I am not even trying to get enough instances - during MIDI composing I am just using a simpler monitoring reverb, either one instance of gigapulse or an external reverb, and then freezing tracks each with a separate gigapulse instance. Not ideal for instant real-time work, but there are no real problems once you get used to how the gigapulse effects the instruments. Besides, I always want to have that final level of control of individual tracks in a dedicated, separate, purely audio mix outside of MIDI. Maybe that is redundant but I seem to love redundancy.

William,

How do you get a satisfactory placement if you can't reference each instruments position? Just by making sure each one is on a seperate mic with instruments on the same mic in different front to back relationships?

I think MIR may solve a lot of these issues but I would love to be able to place the orchestra exactly right with Gigapulse.

Dave

You can solo then reference each one separately with gigapulse to test if it sounds o.k. The overall monitoring reverb gives you the level in relationship to the others.

William, I must say this is a very interesting and 'practicle' way to manage this huge Pro ed library (most efficient use of the HD and especially RAM). I assume that all the 'solo WW's, solo vln, solo cello, and everything but the 'ensemble' type percusssion instruments can be converted to mono?

Thanks in advance.

Rob