Well . . . . .

It looks like I may have to be a guinea pig on this one.

When we get our system up and running, I'll report back on what video card we used and what success or problems we experience with it.
Perhaps it will be of some use to others migrating to 64-bit configurations.

We deliberately chose not to go with the nForce4 with PCI-Express, as it is too new, and we are afraid that a setup that is that focused on video performance could in some way be at the expense of, or cause problems with the audio performance.

It seems to me that the best way to handle this issue would be to have completely separate, dedicated buses for audio and video - and even have it so the components can be externally plugged in/ hot swappable.
(Maybe the PCI-express is a step in this direction - I'm not techie enough to understand all that yet.)

The Infiniband technology sounds very promising as well.

Externally plugged in components would also do a great deal to alleviate the problems of heat in an enclosed system.
You could also employ state-of-the-art vibration damping on the external components that need it (audio interface, high-speed HDD, CD-R/DVD-R, etc.)
External components would also allow for greater electrical isolation between components to reduce EMI interference, etc., and the noise issues might be easier to address as well.

I'm also hoping to see 64-bit audio interfaces, and 32/64-bit AD converters in the next few years but nothing much seems to be happening on that front. :(

At any rate, it will be very interesting to see what direction this all evolves in.
I think designing and building the ideal audio workstation is the most exciting frontier in computer technology.

BTW - I love this forum.

Cheers!

64 bit audio? 24 bit already exceeds the dynamic range of human hearing. The point of greater bit depths would be ... ?

Hey Maluvia, welcome to RO, glad to see like the forum.

We deliberately chose not to go with the nForce4 with PCI-Express, as it is too new, and we are afraid that a setup that is that focused on video performance could in some way be at the expense of, or cause problems with the audio performance.

I agree with you on this. For a DAW it is best not to go with the newest designs. Staying with proven designs is best because a DAW is all about stability and stability comes from known reliable components. However you seem confused about the video/audio issue. Mobos are not really designed with one or the other in mind, it is the components connected to it that determine it's uses. The chipset is there to facillitate the transfer of data from the buses to the CPU but it cares not what passes over those buses. In other words the fastest video card will not make your system treat audio any differently than a 5 year old video card, if anything it may help improve the audio (depending on the card). This is because the VPU on the video card handles the video processing so the CPU doesn't have to. For any 3D video card built in the last 7 or 8 years a DAW just doesn't present any kind of challenge. Any good quality 2D card will do well for a DAW (get at least 16MB's of Video RAM so you can cover 2 high resolution monitors)

It seems to me that the best way to handle this issue would be to have completely separate, dedicated buses for audio and video

They are seperate buses. PCI for audio, AGP or PCIX for video.

External components would also allow for greater electrical isolation between components to reduce EMI interference, etc., and the noise issues might be easier to address as well.

By removing the components (AGP/PCI etc.) from the case you are exposing them to RFI, a far worse problem than EMI. EMI is not really transmitted from card to card but rather is induced through the cabling inside your case. By carefully and neatly routing your cabling inside your case EMI can be reduced to imperceptable or eliminated from entering your audio. Your metal case acts as a Faraday cage to stop RFI from entering the audio. It is best to have your DAC's located externally in a metal case to prevent EMI and RFI interference but once the audio is converted to 1's and 0's interference plays no role. If you have EMI issues start at the source (A/C), clean power is where it all begins.

You could also employ state-of-the-art vibration damping on the external components that need it (audio interface, high-speed HDD, CD-R/DVD-R, etc.)

David French is the king of Shhhh around here and has employed some great technologies to maintain the silence. I'm sure he would help you with any questions you may have.

Good Luck with your build and please list it in this thread when your done.

(Dead Link Removed)

The point of greater bit depths would be ... ?

The point would be to bridge the remaining gap between the fidelity of digital and analog recording.
I thought this was a goal of all digital music recording.

I have yet to hear a digital recording that matches the fidelity of high-quality analog - but I believe it is possible if the digital recordings are made at sufficient resolutions.
It will be a great day when digital recording can truly match the quality of analog for pro-audio recording, and not be just a poor-man's substitute, as it is for many of us who cannot afford 24-track reel-to-reels.

The degree of multi-tracking possible in digital recording as well the incredible fine-tuning possible with digital editing clearly surpasses analog.
But the audio fidelity of analog still surpasses digital at present resolutions - imo.
However, I think that at higher resolutions this gap can be eliminated.
As far as which of the two resolution factors - sample rate or bit-depth has the most significant effect on audio fidelity - to my ears, it is clearly bit-depth.

24 bit already exceeds the dynamic range of human hearing.

Well it does not exceed the range of my hearing. :wink:
If it did, it would be hard to see the point of using it at all.

Of course the type of music which is being recorded has a bearing on how high a resolution is beneficial or necessary.
If one is recording primarily or exclusively electronic or software instruments, then I would agree that resolution would not be that big an issue - or perhaps any issue at all.

However, when attempting to digitally record acoustic instruments, especially guitar, as we are - the resolution is critical.
As I mentioned, I have yet to hear a digital recording - even 24/96 - of my husband's guitar that can match the fidelity of his analog 8-track, and this has been a point of continuing frustration for us, as well as the reason we are building this new DAW.

My husband has a unique guitar - a 28-yr old Gibson Hummingbird with a gold Martin thin-line pickup, and a handmade deer antler saddle.
He plays it tuned down quite low - always playing in alternate tunings, never standard.
The incredible depth of sound, the rich timbre and harmonics which are the result have proven very challenging to capture.

I have become very sensitive to this sound, after listening to it and helping to carefully hone it for years.
I can detect immediately if he tunes his guitar a microtone up or down from usual for a particular piece, as it completely changes the equalization.
The guitar is naturally rich in the midrange frequencies and this is what seems to get comprised when recording digitally.
What I hear coming out of the computer is harsh and strident - the high-end (and bass) exaggerated at the expense of the warmth of the midrange.
On our first CD, we had to spend many months with the computer, laboriously tweaking the software EQ and compression (which should not have even been necessary) in order to try to restore the warmth and fullness of the original sound.
We did not completely succeed, but it came out fairly well considering what we had to work with at the time.

After many years of experimenting with various components, signal routings and different types of cabling (we finally tossed out the 31-band equalizer, and completely abandoned balanced cables in favor of Zaolla Silverline unbalanced), we have finally achieved transparency of sound on the analog recordings.
I will not be satisfied until we can achieve that kind of fidelity and transparency with digital recordings.

I do believe it is possible with sufficient resolution.
Whether or not 64-bit AD conversion would be audibly better than 32, or 24-bit I cannot say since I have never heard it.
Perhaps 192/32 would do it - perhaps even 192/24.
That is what we hope to soon find out.

Thanks for the welcome Don.
(I was beginning to wonder if I had stumbled into another men-only forum, as I have often done before.) :wink:

I appreciate your clarification of the role of the different buses and the interaction of the video and audio signals.
I still do not completely understand this.
That is one of the reasons I am here - to learn more about these kinds of things.
I only know that I have read about this being a problematic issue in digital recording.
The issues surrounding EMI/RFI are also confusing to me and I hope to learn a lot more about that subject as well.

clean power is where it all begins.

This we understand and is the reason for our choice of power supplies, though I wish PCP&C would up it to 600 Watts.
However the specs on this PS are excellent, especially with regard to it's heat-tolerance, and rock-solid stable voltages - so we hope this will do the trick.
We have learned the hard way the importance of having a clean, powerful audio signal before all else, and I would think the same would apply to having ample clean power to the computer components.

BTW,
We are planning to run dual operating systems on separate partitions of the first drive: Win XP home, and Vector Linux.
We are going to give Ardour a shot for our multitracking on this next project.
They claim to have complete transparency in the software - no filtering of the sound, or plug-ins to bypass - and can support any resolution the hardware can, so it will be very interesting to see how this turns out.
The XP partition will be for Cubase, other Steinberg software, and some other legacy Windows audio software we still like to use.

If anyone else on this forum is using Ardour I would be interested in hearing how it is working for them.