Sujet Les meilleurs "NAPPES" de synthé?
- 199 réponses
- 28 participants
- 27 073 vues
- 33 followers
Afficher le sujet de la discussion
On a beau dire, il y a encore un monde entre le virtuel et le réel.
PS: les meilleures nappes, en digital S&S, c'est le Wavestation ou le JD-990. Na!
Anonyme
521410
Anonyme
521410
Anonyme
521410
bemol1
2517
Squatteur·euse d’AF
Membre depuis 21 ans
Anonyme
521410
.: Odon Quelconque :.
11070
Drogué·e à l’AFéine
Membre depuis 22 ans
Citation : Ceci dit, même si ce n'est qu'une histoire de calcul, j'ai pas entendu de plug à ce jour sonner comme un Virus C, un Clavia G2 etc avec, comment dire, la patate, le ooomph etc. Alors?
L'investissement en R&D nécessaire à la production d'un hardware onéreux ne serait-il pas de la même magnitude que celui mis en oeuvre pour un soft qui sera de toute manière piraté dans les heures qui suivront, et obsolète quelques mois plus tard?On a beau dire, il y a encore un monde entre le virtuel et le réel.
PS: les meilleures nappes, en digital S&S, c'est le Wavestation ou le JD-990. Na!
« What is full of redundancy or formula is predictably boring. What is free of all structure or discipline is randomly boring. In between lies art. » (Wendy Carlos)
bemol1
2517
Squatteur·euse d’AF
Membre depuis 21 ans
bemol1
2517
Squatteur·euse d’AF
Membre depuis 21 ans
Anonyme
521410
bemol1
2517
Squatteur·euse d’AF
Membre depuis 21 ans
Native Vs DSP: Dan Phillips (Korg R&D) On Processing Power* Korg R&D Product Manager Dan Phillips, who was involved in the development of the OASYS PCI from the outset, discusses the relative merits of native (CPU) and DSP processing, and makes the case for going the DSP route: "From an audio fidelity perspective, there is no basic difference between CPUs and DSPs. It's all just math, after all. There are, however, a number of practical issues. First, there is simply not enough processing power in current CPUs to do all the tasks that we are asking them to do. Working in /Digital Performer/ at 44.1kHz, I can't do a full mix using my favorite MAS EQs and compressors without bouncing to disk. Forget about doing special effects and reverbs. And then we have 96kHz on the horizon, which -- for those who think it makes enough of a difference -- will double the processing requirements yet again. Into this already too small box, we are now supposed to fit synths and samplers as well? "Clearly, as CPU speeds increase, the box will get bigger. When multiprocessing becomes widely available and supported, that will make the box bigger as well. Perhaps in a year or two I will be able to run enough compressors and EQs to do a full mix in real time. Perhaps a year after that I will be able to do most special effects as well, and perhaps a year or two after that I will be able to do all of that at 96kHz. For now, however, the box clearly isn't big enough -- and PCI DSP cards offer the extra room we need to get the job done today. "A second issue is latency. Getting audio out of the CPU takes time. Even the lowest-latency audio cards offer a minimum of 3mS of latency -- and that at a noticeable cost in CPU processing power. You can add to that the amount of time required to get MIDI into the computer, and the time for the synth program to do its work (including potentially passing buffers back and forth). Now, compare that to, for instance, a Wavestation, which fires off a note after only 2mS. There are compensating factors, though. Some soft synths have little or no additional latency per sounding note, whereas most actual 'keyboard' instruments, including the Wavestation, take a certain amount of time per individual note. "Latency is even more of an issue for processing of external audio, since the audio must pass in through the PCI buss (minimum 3mS), through the processing, and then back out again (minimum 3mS again). In comparison to this, OASYS PCI has basically no latency for audio processing or synth output. Does the extra 3mS (minimum) make a difference with synths? My guess is that it will depend on the player; some will mind, others may not. Latency certainly does seem to make more of a difference with effects processing, however -- not too surprising, since the latency is twice as high as with synths for any given system. "Finally, there's the issue of the intellectual property -- audio algorithms and sound design -- that the various companies bring to their products. Currently, hardware and software synths are very different in this regard -- not due to the fundamental technologies involved, but because of the relative amount of experience that the companies bring to the table. The current soft-synth companies are young and fairly inexperienced, which has both positive and negative effects. On the good side, it means that they are willing to take chances, and to produce unusual and exciting products. On the down side, they do not have the background of Korg or Roland or Emu. In my opinion, this is reflected in the quality of the basic algorithms (oscillator, filter, envelope, and others), in the advanced algorithms (effects, physical modelling if any, and so on), and in the quality of the sound design which uses those algorithms. "To sum up, then, DSP and host processing both use basic math to create and process sound -- no real difference there. Host systems suffer more from latency issues, and do not currently have enough processing power to simultaneously do all the tasks required of them. Finally, through no fault of the basic technology, host-based synths currently don't sound as good as the best hardware and DSP synths." 
- < Liste des sujets
- Charte