Page 1 of 1

need feedback on this conceptual idea

Posted: Sat Feb 21, 2004 1:20 pm
by laughing_bear
just thinking out loud about the future of my recording ....more HD power on one maschine seems like the logical next step, however, there is a considerable downside, the bottleneck of PCIbus serving 1.streaming samples 2.hostseqeucner recording/playback, and 3.plugins.

In other words, before a 2.x Ghz CPU comes to its limits, the PCI bus is probably overloaded much quicker, particulary if one would want to record tracks at higher rates 24bit/96khz....

Thinking about a solution.... more maschines, interconnected via 1Gbit networkcards.... also provides considerble expansion options in the furture....also allows to use cheaper HD stacks with RAID 0 or RAID5 hardwarecontroller solutions....

In the first step.... get second maschine, twisted cable, 1Gbit Card in both maschines, RAID Controller plus 4* 250Gig HD's in the second maschine.... access HD's via network, use Nuendo because of enhanced network support, and have all streaming jobs done by maschine 2.....

Doin a little math... Sample at 44.1khz / 16 Bit - means 705'600 Bits/sec need to be transfered...

A 1 GBit/s-Net has a theoretical 1'073'741'824 /sec Data Transfer.

That means theoretically I could playback 1521 Tracks at 44.1Khz/16bit, SIMULTANEOUSLY!

The 4 RAID HD's should have no problem with that.

PCI Bus with Datatransferrate of 133MBytes/s....comes to 1064Mbit/sec, just a tiny little more than that GBit net.

In comparison.... Firewire 400bit comes only up to half of what 1Gbit net can provide.

Of course all the above are theoretoical values, and from 1521 tracks you probably end up with 1200, LOL which sounds good enough to me.....

What do you think?

Posted: Sat Feb 21, 2004 4:54 pm
by Francois
If you can read French, we have a very lively 'prototype' forum on Tritonton, animated by Teuf and SpiderWeb. It's full of interesting views, ideas...

Posted: Sat Feb 21, 2004 5:17 pm
by laughing_bear
I can read a little, but all in all it is very long ago that I spoke any french...desperate...

I translated the above text with babelfish... to me this makes only little sens in french, probably quite difficult to read for a french person, right?

juste la pensée hors fort environ du futur de mon puissance de l'enregistrement....more HD sur un maschine semble comme la prochaine étape logique, cependant, il y a un du côté incliné considérable, le goulot d'étranglement des échantillons 2.hostseqeucner recording/playback de la portion 1.streaming de PCIbus, et de 3.plugins. En d'autres termes, avant qu'une unité centrale de traitement de gigahertz 2.x vienne à ses limites, l'autobus de PCI est probablement surchargé beaucoup plus vite, particulary si on voudrait enregistrer des voies à des taux plus élevés 24bit/96khz.... Penser à une solution.... plus de maschines, reliés ensemble par l'intermédiaire des networkcards de 1Gbit.... fournit également des options considerble d'expansion dans le furture....also laisse employer les piles meilleur marché de HD avec l'INCURSION 0 ou les solutions du hardwarecontroller RAID5.... Dans la première étape.... obtenez le deuxième maschine, le câble tordu, la carte de 1Gbit dans les deux maschines, contrôleur d'INCURSION plus 4 * 250Gig HD dans le deuxième accès le HD de maschine.... par l'intermédiaire du réseau, utilisation Nuendo en raison d'appui augmenté de réseau, et faites faire tous les travaux coulants par le maschine 2..... Doin des peu de maths... Échantillon 44.1khz/16 au bit - le besoin des moyens 705'600 Bits/sec d'être transféré... Un 1 GBit/s-Net a un transfert de données théorique de 1'073'741'824 /sec. Cela signifie que théoriquement je pourrais playback 1521 voies à 44.1Khz/16bit, SIMULTANÉMENT ! 4 L'INCURSION le HD ne devrait avoir aucun problème avec cela. Autobus de PCI avec Datatransferrate de 133MBytes/s....comes à 1064Mbit/sec, juste un minuscule peu plus que ce filet de GBit. Dans la comparaison.... Le câble d'incendie 400bit vient seulement jusqu'à la moitié de ce que le filet de 1Gbit peut fournir. Naturellement tous les ce qui précède sont des valeurs theoretoical, et de 1521 voies vous finissez probablement vers le haut avec 1200, LOL qui me retentit assez bon..... Que pensez-vous ?

Posted: Sun Feb 22, 2004 12:31 am
by Francois
Pure gibberish!

need feedback on this conceptual idea

Posted: Thu Nov 01, 2007 7:58 am
by stultzsweeties
Purchase a Yamaha digital mixing consule and record it directly to a 4" tascam two track recorder and let's get back to simple. Unless equipment is perfectly designed to work together, how is the signal every going to get the way you want it. Here's what you are trying to do. You're in a cave about 200 yards in. You've made 3 turns and a curve to get where you're at. Now you're flashlight batteries are dead and you can't see(hear). You have four mirrors. You (at your own thoughts) stratigically place the mirrors at various angles until the light at the end of the tunnel is finally projecting the suns light back to where you are in the tunnel. This is where you are at, musically, with your signal, weak because signal in doesn't equal signal out, and any one of those mirrors(devices) could be the cause of why the light in does not match the light out. The answer is to go to the light. Bypass all the mirrors.

Posted: Fri Dec 07, 2007 10:00 am
by teuf
Francois wrote:If you can read French, we have a very lively 'prototype' forum on Tritonton, animated by Teuf and SpiderWeb. It's full of interesting views, ideas...
I'm not dead :mrgreen:

Teuf

Posted: Fri Dec 07, 2007 10:35 am
by Lorenzo
teuf wrote:
Francois wrote:If you can read French, we have a very lively 'prototype' forum on Tritonton, animated by Teuf and SpiderWeb. It's full of interesting views, ideas...
I'm not dead :mrgreen:

Teuf
Happy to see you're still here around!