James, I edit and publish a magazine and therefore get lots of press releases. I'm telling you that right now there isn't any noise about 64-bit software... as I keep saying over and over. Companies tend to publicize what they're about to come out with as soon as they can reasonably do so, for obvious reasons. Therefore I'm drawing conclusions that could well be off-base, but unfortunately are probably right.

Sonar has a 64-bit version. Yes. I already said that. But you'll notice that it doesn't have a streaming sampler with 64-bit memory access yet. Steinberg has said that they only need to "flip a couple of bits" to get 64-bit memory access for HALion, but that hasn't happened yet. That accounts for two companies. Has anyone else breathed a word about it?

The question is rhetorical, of course.

Next, Cakewalk reports getting a 10-15% processing improvement with their 64-bit version over the 32 on the same hardware. That doesn't automatically make it so across the board with every program - it's quite possible, from what I've been told, that some processes could actually slow down. We'll have to wait and see.

Also, the 64-bit OS X will be backwards-compatible with 32-birt software (obviously, or every single program would have to be rewritten yet again). However, I've never heard anyone say that 64-bit DAWs will be backwards-compatible with 32-bit plug-ins. Chances are that every one of our instrument and processing plug-ins will have to be updated to run at 64 bits, i.e. I haven't heard anything that leads me to believe we'll be able to mix the two.

The reality in this blind-leading-blind discussion is that none of us who doesn't work on a program knows how difficult it is to make it 64-bit compatible. My informed guess is that it's a huge deal for most of them. Therefore I don't assume that this is going to happen overnight, just because the operating systems will be ready.

Nick, I agree with you 100%!

J.

That makes two of us, Jerome. [:)]

@Another User said:

Ram prices will fall once people start buying more. People will start buying more when they can use it...

Miklos, I'm not an economist, but to it seems like that would create a *demand.* For the prices to fall there has to be an abundant *supply* - economy of scale notwithstanding.

cm pointed out that the chip manufacturers can't always stay in step with the type of RAM computer manufacturers are using all the time. Hopefully they'll catch up with the stuff Mac Pros are using by the time 64-bit memory access becomes available to us, and the prices will drop.

On a related note, PC3200 memory used by G5s is now about $100/GB (up from $75 not very long ago). While the general computer public doesn't need extended RAM access, it turns out that we actually can use it today.

In case you missed the thread:

http://community.vsl.co.at/viewtopic.php?t=9801

$800 to get about 7GB of RAM access is feasible. But the error-correcting RAM used by Mac Pros is $250/GB. OUCH!

TASCAM is working on a 64-bit version of GigaStudio, but we don't have a release date yet. Part of the problem is that the two mass-market 64-bit OS releases, Vista and Panther, aren't completed yet. So manufacturers probably don't want to release their 64-bit code before testing them on these operating systems. Once these two products ship (probably both in January) and the NAMM show begins (coincidentally also in January), I'd expect some more announcements about 64-bit apps and plug-ins.

i'm wondering how restrictive (= expensive & complicated) MS will handle the certification for 64-bit drivers and applications, nothing without a certificate will run in VISTA-64. so let's patiently wait if we hear something about a 64bit sequencer until january ...
christian

cm, are you talking about the dialog that comes up when you install something, complaining that the program you're about to install could be risky?

I like that analogy Guy! [:D]

Nick--

for what I've heard, Vista in 64-bit mode will not work with drivers like XP used to. In XP, you could dismiss the alert telling you the drivers were not certified, and install them anyway.

Vista 64-bit won't allow you to do so.

Jerome

Either that completely stinks or it's a fantastic feature that ensures every single piece of software will be 100% reliable and never ever break down.

[:)]

jerome mentioned it already - it's not only the dialog coming up (as we know it from various driver installations actually from big companies) but a kind of *execution prevention* for not certified software accessing the kernel - http://www.microsoft.com/whdc/system/platform/64bit/kmsigning.mspx
not sure now, but i suspect to have read something similar for leopard ...
basically such a behaviour makes sense and could be introduced also for all 64bit capable procesors even using 32bit software
christian

I haven't heard anyone other than Nemesys/TASCAM talk about accessing the kernel. Is it actually necessary to get as low-level as that to write good music/audio software?

Does the Vienna Instruments player do that?

My understanding about Macs is that you can't access the hardware directly. Is that no longer true?

nick, whereas giga's streaming engine is running in kernel mode the vienna instruments are using the user space (of memory) ... but i think there is no audio software at all not _accessing_ the kernel (of an operating system).
and no modern operating system is accessing hardware directly, they are all using hardware abstraction layers, otherwise it would be almost impossible to port something to another processor platform.
christian

Thanks cm.

@Nick Batzdorf said:

James, I edit and publish a magazine and therefore get lots of press releases. I'm telling you that right now there isn't any noise about 64-bit software... as I keep saying over and over. Companies tend to publicize what they're about to come out with as soon as they can reasonably do so, for obvious reasons. Therefore I'm drawing conclusions that could well be off-base, but unfortunately are probably right.

Also, the 64-bit OS X will be backwards-compatible with 32-birt software (obviously, or every single program would have to be rewritten yet again). However, I've never heard anyone say that 64-bit DAWs will be backwards-compatible with 32-bit plug-ins. Chances are that every one of our instrument and processing plug-ins will have to be updated to run at 64 bits, i.e. I haven't heard anything that leads me to believe we'll be able to mix the two.

Heh, I wouldn't think to challenge your ear on the industry. I was just pointing out some facts about a particular piece of software that I happen to know a bit about. Beyond that, I wouldn't presume to know anything.

Just fyi, Sonar is using some sort of 32-bit thunking layer (they call it BitBridge) to allow 32-bit plugins to work in their 64-bit environment. This is certainly going to be a requirement for 64-bit computing to take off, as many musicians have obviously invested an enormous amount of money in 32-bit plug-ins, and it's likely that, for whatever reason, many of those will never be converted. If Cakewalk could figure out how to do this, I don't see any reason other vendors can't do the same.

I do agree with your assertion that widespread adoption is likely still a ways off for the bulk of the industry. And, notwithstanding those who will make the jump immediately when it's available, there are many with a fairly significant investment in 32-bit hardware that will probably wait for a more normal obsolescences cycle before buying new systems, and that means a few years past the initial hardware/software offerings even.

Naturally, it's all pure speculation. But, that's half the fun, no?

@James Boer said:

Just fyi, Sonar is using some sort of 32-bit thunking layer (they call it BitBridge) to allow 32-bit plugins to work in their 64-bit environment. This is certainly going to be a requirement for 64-bit computing to take off, as many musicians have obviously invested an enormous amount of money in 32-bit plug-ins, and it's likely that, for whatever reason, many of those will never be converted. If Cakewalk could figure out how to do this, I don't see any reason other vendors can't do the same.

However, using this system could actually make for a worse sound that sticking with a 32bit system!!!! I would also be interested to know whether or not introducing a 32bit app into the melting pot means that the RAM advantages of a 64bit OS are no longer available. I don't know, I'm just asking.

DG

@Another User said:

I would also be interested to know whether or not introducing a 32bit app into the melting pot means that the RAM advantages of a 64bit OS are no longer available. I don't know, I'm just asking.

DG

The way it was described made it sound as though each plugin had access to it's own virtual address space of 2GB. I'd have to double-check to be certain of that though.

@DG said:

However, using this system could actually make for a worse sound that sticking with a 32bit system!!!!

I don't quite follow how you think it could be worse sounding... ???

James, at one point when this whole 32bit/64bit question came up there was a lot of discussion on the Nuendo forum about this, and a lot of the recommended reading was by established professionals not working for Steinberg. I don't pretend to understand all of it, but it boiled down to the fact that with a totally native 64bit system (everything being 64bit, not BitBridge kludges) it was theoretically possible for the sound to improve, although very unlikely, unless the engineer did some ridiculous things. However, as soon as you introduce a 32bit plug there was every chance that the sound would not be as good, depending on the coding involved. This is where I lose the plot, but when you have developers for UAD explaining things, I tend to believe that they know more than I do.

DG

And now there is so much bigger and more harddrives available. So....

I got a question regarding harddrives and what to "go" for.
Me and my buddy is going for a Mac Pro 3.0Ghz, but we are not sure if we should go for only standard 500 GB drives.

My idea was since this Mac has room for 4 HDs:

2 x Cheetah 146GB SAS 15000RPM with 8MB cache
1 x Cheetah 300GB 10000RPM 8MB chache
1 x Barracuda 750GB SATA2 16MB 7200RPM

Now, would these fast drives be in help regardign samplecounts with Vienna Instruments and other VSL samples with Logic ?

@DG said:

However, using this system could actually make for a worse sound that sticking with a 32bit system!!!!

I don't quite follow how you think it could be worse sounding... ???

James, at one point when this whole 32bit/64bit question came up there was a lot of discussion on the Nuendo forum about this, and a lot of the recommended reading was by established professionals not working for Steinberg. I don't pretend to understand all of it, but it boiled down to the fact that with a totally native 64bit system (everything being 64bit, not BitBridge kludges) it was theoretically possible for the sound to improve, although very unlikely, unless the engineer did some ridiculous things. However, as soon as you introduce a 32bit plug there was every chance that the sound would not be as good, depending on the coding involved. This is where I lose the plot, but when you have developers for UAD explaining things, I tend to believe that they know more than I do.

DG

Ah, I see... Honestly, this is not something I'm going to lose much sleep over. My guess is that a 64-bit pipeline is going to pretty much sound identical to a 32-bit pipeline (quite pristine). It gives some extra headroom to a pretty wide pipe to begin with, so it's not necessarily a bad thing, but I'd be surprised if there were an actual audible difference. My feeling is that the real value of 64-bits is in the addressable memory - everything else is a sideshow.