[[;)]]
Hi,
In my opinion this equipment is more than enough.
Greetings from the GraceHill,
theiss2003
Hi,
In my opinion this equipment is more than enough.
Greetings from the GraceHill,
theiss2003
-
Actually it's somewhat overkill IMO - Gigastudio is'nt especially taxing
on the CPU and fast Ram does'nt improove performance enough
from a cost-benefit point of view. For the price of the Rambus and
the CPU, you could most likely get 2 computers with SD Ram and
a good Celeron - thus doubling your Giga capacity. Of cource there's
the licence matter of Gigastudio.
Anyway - this is my experience: if you don't run a sequencer on the
same computer as Giga, a Celeron 2GHz, 1 Gb SDram, 7200 rpm HD
should get you 160 voices. ( all though some chipsets - mostly Via's
for Athlon, will be hard pressed to deliver this ).
kind regards
Bjarne
-
Hi,
I have done careful research on that matter and I came to the result, that Gigastudio 2.53 does NOT PROFIT from any CPU-power above 800MHz Athlon/PentiumIII, as long as you are heading for maximum polyphony and/or maxing out your RAM capacity. Cheap CL3 PC-133 SD-RAM provides you THE SAME performance as horribly expensive Rambus RAM. Make it 1Gig (2x 512MB) for Win98SE and 1,5GB for XP (3x 512MB); note that GS currently won't allow you to load more sounds than up to 1GB. The update to GS 3.0 MIGHT solve that. For maximum polyphony you rather have to make sure, that your samples are located on a separate 7200RpM-drive (no OS and/or swapfile there!). As GS 3.0 will for sure allow the system to play more voices simultaneously, a third HD to balance your samples will be a good idea (no RAID needed, just take e.g. the violins and violas from one, the celli and basses from the other HD). The only ones, that gain from more computing-power are the DSP-mixer's real-time-effects, which I don't use much due to their instability (the chorus/delay distorts once in a while and has to get re-inserted) and their relatively poor sound (eq/reverb). Despite of that fact, it has been announced, that GS 3.0 WILL gain from more CPU-power. Therefore I recommend to go for a medium-sized CPU, such as an AMD 1800+ or a P4 around 2GHz (which you will both be able to keep cool without loosing your ears).
To your question regarding how many machines you will need for VSL:
As long as you feel comfortable with "rendering" one instrument after the other, meaning that you capture your GS-output to audio and move the take to your sequencer, a single machine of that kind will be just fine. When you prefer to control the whole orchestra at once in real time, you should plan to set up one machine for each section of the orchestra AT LEAST (3-4 machines...).
Nevertheless it was interesting to hear from you Vienna-people, how many machines you used for the demo(s?).
Best regards
Roman Beilharz
P.S.: Even the VIA-chipsets work perfectly well, if you install the latest separate VIA-Busmaster-driver AFTER having installed the latest 4-in-1-driver-bundle. E.g. the Asus A7V133 Rev. 1.05. (important!) is a good board which uses "plain old" SD-RAM and can be run with a current AMD-CPU up to 2200+ in jumper-setting. Runs stabil on three machines I have built and has some headroom for more traffic to come.
-
Hi Roman,
I´m impressed, shocked and happy about your information. It brings me down to earth...
I wonder how the the hard-disk-version will work if for the first three outputs of the vienna-library one should use three or four pc´s - but it sounds very realistic.
How do you use the librtary?
Yours Sakamoto
-
How do you use the library?
Yepp, I can't wait to find some time to finish building my second machine, since I only had one GS-system up to now. As soon as some serious movie will show up, I will set up two more machines, I guess, but for now it has to be it: 2 Gigastudio-systems, 1 sequencer and a Mackie d8b...
Currently I use a template with my favorite Garritan-strings and almost the whole Vitous-Woodwinds and Brass (which need only a tiny bit of RAM compared to THIS brother here). As I have just been starting to use VSL recently, I am doing my composition/arrangement with the template and render all instruments, that need to get finalized one after the other.
This means literally: Reset the sampler, solo the desired instrument's track, load all necessary articulations into GS; insert, edit and/or re-record what seems necessary to trigger the tricky stuff, capture the GS-instrument to audio (within GS), load this recording into the sequencer (via ethernet) and continue with the next instrument to bring towards perfection.
Note, that I am far too lazy to do this with every single line: The Garritan-Strings sound excellent for most occasions already, sometimes even the Vitous do what I intended. The true killer-cases for VSL are all solo-lines, fast attacks, swells, runs etc.. These make your arrangement rock, but you have to be prepared to really dig into it. We have to learn our tools. I consider, that I will need around half a year to really SING those VSL-samples, just as I needed about the same time to be able to play the Garritans in a way that I wouldn't have to edit all of my recordings. You know, it is great to always have the latest stuff, but it is much more important, that you are able to really kick what you've already got! Less news-picking, more training, guys!! Express yourself. Follow your heart. Trust your ears. A guitar will be just fine...
AMEN
P.S.: Dual-processors are good for Macs and PCs that need the utmost CPU-power possible. Gigastudio doesn't need a dual-CPU system. To be honest, I would never use one, because I know how to build a low-noise single-CPU-system. I have no idea how to get a dual-CPU-system thermally balanced and so silent, that I was content with running it around my place.
-
All good advice. Right now with Giga, you are better off with multiple inexpensive machine than with single huge machines. Keep in mind that this library will in some cases (like legato mode) use less MIDI channels than many other libraries because you tap out at 16 MIDI channels or less once you load a bunch of legato instruments. That means for those machines, you don't need 64 MIDI channels of MIDI input, just 16. You can even use the old joystick port for those 16 channels that is usually included on many motherboards and is a fast solution compared to many new interfaces. Also, to use more than a good gig of RAM (for now anyway) you need a machine that can handle it and Windows XP in most cases. 98 just doesn't seem to work with more or see it. XP will then usually use half of whatever you fit into the machine and leave the rest for Giga, usually leaving you where you were with 98 with only a gig of RAM. You are a good ways toward two machines at that point.
Now, you may have some instruments, (like the basic multi velocity ones) and percussion and what have you that you can cram a lot of into a machine and use more than 16 MIDI channels. You might even push the polyphony so a bit faster of a machine will help. This is all Giga 2.5 philosophy. With 3.0 I am confident that there will be some use for more horse power for some things. For instance, it goes without saying that they will have to open up the polyphony limit a ton to meet and exceed Kontakt not to mention 24 bit samples. Then, faster machines will come in handy so you can't lose either way. As some of your systems start to become seemingly obselete, keep in mind that you can always dedicate one of these to just a few legato instruments and get some good use from them.
This library is pretty flexible from what I can see. Some sounds can benifit from cheap systems while others can use faster ones. 3.0 will probably take advantage of faster systems.
I'm developing a resource managment philosophy and I will share it eventually in detail.
Cheers
David Govett
-
Hi David,
thanks for seconding my post. I hate to correct you, but regarding the RAM-issue you are not completely right.
XP will then usually use half of whatever you fit into the machine and leave the rest for Giga, usually leaving you where you were with 98 with only a gig of RAM.
A well optimized XP-system will not eat up more than ca. 220MB of RAM, so if you put in 1256MB, you will currently be able to load GS to its 1GB-limit. I was not able to reach that capacity with the maximum of ca. 1GB RAM, where Win98 will still boot properly. Although it is not a big difference, I admit. Must be around 100-150MB. The striking thing to go for XP was the stability and the GS-performance, which increased to 160 continuously streaming voices (no glitches!) using the very same hardware. Win98SE started to choke and twirl under high polyphony.
William, I can't see the misunderstanding you stated??!!
regards
Roman Beilharz
-
Hello Sakamoto,
nobody knows, but I dare the speculation, that Giga 3.0 might bring us two things related to this thread:
1. The 1GB RAM-limit might get solved, so that GS will probably be able to access all RAM given by the system (Yes Tascam, I know that you believe it already DOES that and the limit was caused by XP not giving the RAM to GS - excuse this simplyfication - but I doubt that big MS will solve the prob' for you...). Problem: Almost all boards, that can use cheap SD-RAM won't let you plug in more than 1,5GB RAM.
>>>But: Are you sure you like to wait 15 minutes each time you load around 2GB of samples??
2. GS will for sure allow more voices (up to 320 mono-voices@16 Bit??) to get streamed - but that depends on your hardware. Note that the real-world-limit of a let's say 2,5GHz-P4-system with a heavy stripe-set (RAID) will be around 380 mono-voices@16 Bit or maybe 250 voices@24 Bit.
>>>But: As a single machine won't be able to hold all articulations for a nice VSL-template anyway, I doubt we have to really go for maximum polyphony ...
That's all for now, folks. "My 2 cent" (and, and, and: Who was THAT??? [;)] Hint: hangs around here...)
Please understand, that I won't reply to this thread again. It's starting to munge too much of my time...
Good ideas, visonary thoughts - and take a deep breath when you walk outside!
Roman
-
How do you use the library?
Yepp, I can't wait to find some time to finish building my second machine, since I only had one GS-system up to now. As soon as some serious movie will show up, I will set up two more machines, I guess, but for now it has to be it: 2 Gigastudio-systems, 1 sequencer and a Mackie d8b...
Currently I use a template with my favorite Garritan-strings and almost the whole Vitous-Woodwinds and Brass (which need only a tiny bit of RAM compared to THIS brother here). As I have just been starting to use VSL recently, I am doing my composition/arrangement with the template and render all instruments, that need to get finalized one after the other.
This means literally: Reset the sampler, solo the desired instrument's track, load all necessary articulations into GS; insert, edit and/or re-record what seems necessary to trigger the tricky stuff, capture the GS-instrument to audio (within GS), load this recording into the sequencer (via ethernet) and continue with the next instrument to bring towards perfection.
Note, that I am far too lazy to do this with every single line: The Garritan-Strings sound excellent for most occasions already, sometimes even the Vitous do what I intended. The true killer-cases for VSL are all solo-lines, fast attacks, swells, runs etc.. These make your arrangement rock, but you have to be prepared to really dig into it. We have to learn our tools. I consider, that I will need around half a year to really SING those VSL-samples, just as I needed about the same time to be able to play the Garritans in a way that I wouldn't have to edit all of my recordings. You know, it is great to always have the latest stuff, but it is much more important, that you are able to really kick what you've already got! Less news-picking, more training, guys!! Express yourself. Follow your heart. Trust your ears. A guitar will be just fine...
AMEN
P.S.: Dual-processors are good for Macs and PCs that need the utmost CPU-power possible. Gigastudio doesn't need a dual-CPU system. To be honest, I would never use one, because I know how to build a low-noise single-CPU-system. I have no idea how to get a dual-CPU-system thermally balanced and so silent, that I was content with running it around my place.
Hi Roman,
would you like an excellent classical guitar?
Please look at:
http://www.v4m.net
or
http://www.jovanpesec.net
Enjoy your VSL, I fully agree your comment.
[:)] Jovan
-
P4 @ 2GHz, 1G RAM, dedicated sample 7200RPM HD with 8Mb cache, XP, GS96, Aardvark Pro LX6.
I can load the entire Gigapiano in 2-3 seconds. I'm sure the VSL libraries are far larger, so that remains to be seen...
So is there any point having GS160 on multiple machines instead of, say, GS96? Cheaper!
Emlyn