Home |
Search |
Today's Posts |
#2
|
|||
|
|||
adding DSP power to DAWs
Mike Rivers )
in article znr1068552302k@trad wrote: In article writes: Not to mention that with certain programs like Logic Audio, there is a latency introduced to any track that you process with the UAD card. You have to insert the UAD comp-delay on all the other tracks to compensate for this delay. 8 tracks is no problem, but it takes time when you have 70 tracks, as I have found out. That sounds like way too much work. If I had projects where I had to mix 70 tracks, I'd take them to a studio that had a suitable console. Of course latency is a problem with any digital system. It can be reduced, adjusted for, and in some cases hidden from the user, but it never disappears. I think Cubase SX 2 has made a good stab at doing this right, with global latency parameters for each plugin type/version, and it takes care of the rest. I wonder if you can also do custom/user-definable inserts of outboard gear too. http://tinyurl.com/ul2d IMO, it's at best 'disappointing' that this hasn't been the status quo among DAW softwares for years, yada yada. - Brian |
#3
|
|||
|
|||
adding DSP power to DAWs
In article ,
Ethan Winer ethanw at ethanwiner dot com wrote: The idea of a custom DSP card has always been a turn-off for me. Any proprietary system is always iffy - if they go out of business you're entire investment is lost. If other companies don't support it, as Mike Rivers pointed out, you can run only a limited number of things. There are too many things that can go wrong. Versus getting a faster computer that can run everything faster. All that we do on computers is proprietary and of limited time value. Every single thing you can think of seems to depreciate along the IRS's 5 year plan quite accurately - i.e. it's worth nothing in 5 years. There are rare exceptions, and in my experience, they _all_ involve custom hardware. My aging Sonic SSP still works just fine - nobody told it that it was obsolete - and that hardware is from '95 or so. My NuVerb cards work just fine in a Mac from 1991, and they show no signs of being obsolete. I do however have boxes full of useless floppies that came with rather expensive pieces of software that don't fit into the current software framework, or have been orphaned by the manufacturer with no current updates. Get a new computer and some portion of your software breaks with no hope of replacing it. It's really not anything you can do much about. The only approach should be to buy what you can use now, do enough work to pay for it and move along when you can't use it anymore. But hey, differing opinions is what makes the world go 'round. Yep... Regards, Monte McGuire |
#4
|
|||
|
|||
adding DSP power to DAWs
In article ,
Justin Ulysses Morse wrote: So my suggestion would be to put together a mid-line white box machine and throw a TDIF interface into it so your "real" DAW can spit a few channels of audio into it for fancy reverb and delay effects, then suck the returns back into the big mix. I'm imagining maybe a $400 2GHz box here. That's got to be less than you'd pay for an equal amount of PCI DSP power. You could even make the 2nd box a G4 and run Altiverb on it. I've never tried it, but I looked into it when there was no Altiverb for ProTools TDM and I don't think it'll work so well. Altiverb is primarily a convolutional reverb and you'd need to run a considerable buffer size just to get some useful work out of most machines. So, the resulting sound quality will not be anything like what you'd get with Altiverb inserted on a track because of delays caused by buffer latency. For example, only the first 512 samples of an Altiverb preset are actually convolved - the rest are done with IIR techniques. If your buffer size is 1024, then that 512 sample early reflection stuff will be delayed by 1024 samples - this can hardly sound like what it's supposed to sound like. Of course, you can print the reverb and slip the track, but it's hard to judge what to send to the 'verb when you can't really hear it in realtime. Yeah, you could do it, but it'd be annoying and painful. There's no cross-platform compatibility issues if you're only sending digital audio between them. Hardware reverbs with digital IO (or even analog IO) are just fine for this use. They also tend to retain their usefulness over time, unlike anything software based. Heck, a recent plan of mine is to get a real plate reverb to hook up to my DAW. So, in many ways, I completely agree with you, except that I think a general purpose computer isn't such a good outboard processor these days because of the latency. Regards, Monte McGuire |
#5
|
|||
|
|||
adding DSP power to DAWs
John )
in article wrote: From: Brian Takei I was thinking that by keeping it digital the whole time (avoiding any a/d/d/a conversions) would help minimize latency. Yes, no? Definitely. Latency is a consequence of processing (all things take time), so the less that is done, the less latency. Personally, I think it's worth doing the excercice of measuring rountrip latency out/into a specific DAW, because it's a pretty simple excercise, and then you can use the results at will to determine for yourself if it is significant in your own applications. If an application involves mixing an outboard processed signal back in with its source, then any latency is even more likely to be a problem (because of phase issues). But don't let any such issues/possibilities deter you from just plugging something in and seeing if you like the results, which is probably the first thing (and maybe the last thing) you should do if you've got the stuff to plug in. As a sort of relevant side note -- as I said, I don't have much outboard stuff, or 3rd party plugins. For various reasons, I've resisted the temptations, and right now my strategy is to keep on resisting... until my "penny saved" gear fund reaches the prevailing rate of a Kurzweil KSP-8 (or better), which I estimate will be circa 2005. At that time, if it's still important enough to me, "plunk", and I expect I'll be way more than satisfied. Regards, - Brian |
#6
|
|||
|
|||
adding DSP power to DAWs
|
#7
|
|||
|
|||
adding DSP power to DAWs
Monte P McGuire wrote:
If your buffer size is 1024, then that 512 sample early reflection stuff will be delayed by 1024 samples - this can hardly sound like what it's supposed to sound like. What's 1024 samples to a reverb? Just set your predelay to 23mS less than you otherwise would have (or 11mS if you run it at 96k). I can see the problem with an EQ or compressor, but who cares about a dozen or two millisex on a reverb? It's reverb! It's supposed to show up fashionably late! Okay, on occasions when you have to have a shorter predelay and you have to have the sound of the Altiverb and you can't get the sound you're looking for by slipping the reverb track after printing, then you're forced instead to duplicate the dry track, slip one copy back, and send it to the reverb while sending the "on time" track to the mix. You still come out way ahead on processing power in your primary DAW. ulysses |
Reply |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
System warm-up | Audio Opinions | |||
rec.audio.car FAQ (Part 2/5) | Car Audio | |||
rec.audio.car FAQ (Part 1/5) | Car Audio | |||
FS: SOUNDSTREAM CLOSEOUTS AND MORE!! | Car Audio | |||
old solid state circa 70-80's` | Audio Opinions |