Basic HD System with One 192 I/O and the Command|8

[ASPSA]

Here's a thought for a basic Pro Tools HD system and my reasoning for it...

1. 192 I/O with additional 8 analog outputs (8 analog ins, 16 analog outs)
2. A Command|8

Reasoning:

For my intents and purposes, I do not require a large number of inputs for tracking. However, I would like more outputs available for mixing purposes. Presently, I have a 24x8x2 console that I plan on keeping until I can afford to replace it with either a D-Command or D-Control system. As for now, I intend to use my existing console for mixdown only to feed the 192 I/O's analog outputs directly into 16 of its input channels. I'm getting sick of patch bays; while flexible, tbey tend to become a REAL hassle over time.

As for the Command|8, I'm not concerned about its analog ins/outs. Instead I intend to use it strictly as a MIDI control surface for Pro Tools and other applications. Eight fader channels will sufficiently meet my needs. Besides, using the "Bank" and "Nudge" switches, I can move across Pro Tools channels at will.

1. Does this sound like a reasonable setup for my needs?
2. Would you recommend some other non-Digidesign control surface in place of the Command|8?
3. Most of my work requires more than 16 tracks. However, the hardware configuration I just
described forces be to pre-mix my Pro Tools tracks down to 16 subgroups. I have not yet explored
the Pro Tools guide in any degree of detail. Is this the "correct" way to proceed? Here's what I
envision in Pro Tools. I'm guessing, though I may be wrong, that it is best to first ensure that
all tracks are converted to audio tracks. For example, do not mix with virtual instruments still
active. Is it then possible to playback say 32 audio tracks that are subgrouped into no more than
16 goups, which are then assigned to the respective 16 analog outputs on the 192 I/O?

Thanks for your suggestions; I hope to benefit from them.

[BIG D]

Hi ASPSA,

That system will work well. I would highly recommend a HD2 minimum. Many folk will tell you here that a HD1 PCI doesn't give you a lot of DSP power for more than the mixer etc and a few TDM plugs.

I use the Command 8 and love it a lot. It's a good combination when used with the Keyboard and Mouse to acheive pretty much whatever you like.

Cool thing bout a HD system is there is no need to actually pre mix you can Buss channels to be combined to an AUX stereo channel or two and then send it to a pair of your outputs and use the Command 8 as 8 extra channels for mixing on top of the analog desk. Obviously their eq and Compression etc would come from Plugins but sub grouping BVs or Keyboard parts etc works great then you can automate those channels to compliment your real time analog mixing.

As far as Virtual Instruments, I mix with them all the time without converting to Audio. BFD has 17 outputs realtime, as does Sampletank, RMX etc. I treat them as realtime MIDI instruments.

All the best,

Big D

[ASPSA]

Thanks for the reply.

> I would highly recommend a HD2 minimum. Many folk will tell you here that a HD1 PCI doesn't give you
> a lot of DSP power for more than the mixer etc and a few TDM plugs.
The spirit is willing, but the wallet is weak... What I neglected to mention is I intend to use the system on a well-configured Mac Pro. Personally, I'm waiting for the dual 4-core processor system, which I suspect will be available a few months from now. I imagine such a system coupled with at least 4GB of RAM is sufficiently powerful. Also, substituting an HD core PCI card wih an HD Accel PCIe adds additional power.

Perhaps you can give me some idea of what my original system configuration can handle in terms of the mix engine and plug-ins, as well as what an HD Aceel PCIe version can handle. By the way, wouldn't the Mac Pro system itself be able to co-handle the mix engine and plug-in chores?

Thanks again...

[BIG D]

Hi,

As I have a HD4, I couldn't tell you exactly what a HD1 can handle. I just know from several other posts here in the past, that anyone who has a HD1 PCIx wasn't happy. You have to remember that a HD system doesn't use any of the CPU power if you are using TDM Plugins. All the DSP for the Mixer and TDM plugins comes from the HD ACCEL and CORE cards. The CPU is only used for RTAS or wrapped VST plugs.

Now a HD1 PCIe apparently does come with extra DSP which is closer to a HD2 PCIx system so that would be more power but only supported in a MAC at this stage, and that is a totally different conversation!!

I will be upgrading my PC soon and I am leaning towards Dual Quad core system with 8 gig of ram. Once Vista is approved by Digi, which they say will be very soon, it will be a very strong contender.

If you check the LE section there is a lot of info on what they call the Quadzilla Machines.

Big D

[ASPSA]

Once again, thanks for the reply. I was unaware that TDM plug-ins required the HD cards. I should have realized that with all the talk about the TDM and TDM II bus architectures, which I suppose refers to the bussing architecture on those cards and also explains why the LE and M-Powered versions do not support TDM plug-ins. I'll check some technical documents on this.

By the way, I realized too late that I posted this question in the wrong forum. I should have placed it in the TDM/MacOSX forum, since I intend to run PT on a Mac.

Sorry 'bout that, folks.

[BIG D]

No Problem ASPSA,

Always glad to help PC or MAC based folk.

There is a great thread going on which you should read. The information given from NIKKI K is exceptional and explains very clearly the difference between TDM and LE systems.
It is in the QUESTIONS ON UPGRAING TO HD1 thread.

OK...this is where the misunderstanding of WHAT the differences between LE and TDM are, or more specificaly, and importantly, the differences between a native enviornment vs. a TDM Enviornment are. I will try to explain simply and to the point, so as to not waste pages of detailed things

The primary difference is that the TDM Enviorment is basically self-contained. The system that hosts it does graphics tasks, and then also provides for streaming audio data from disk. OK, so it also provides the slot power for the TDM cards...in this case we are talking PT HD cards. So long as the user's host system is capable of streaming the data and producing the graphical representation (GUI), then the restrictions of the session size, plug-in capabilities, etc are ONLY from the number of DSP chips on the HD cards being used. The I/O interface is connected directly to the HD cards, and therefore NO buffer is required for the live input to pass from the input to the output if the user desires. If the user places any bussing and/or plug-ins inline with that live signal, then the only extra delay added is the itme for those processes to occur. The TDM Enviornment, as I have said, is self-contained, and thus only subject to it's own laws when working with live signal(s).

OK...now we are going to hit on the answer to the original post...

Since it is a self-contained enviorment, this entire enviornment is built upon those HD cards. Think of it as building the digital mixer you desire for each session. Yes, there are restrictions, but a pretty sweet digital mixer is easily attainable. However, one of these restrictions is the point that makes an HD1 system underpowered...IMO...

The restriciton is that when the user decides to send signal out of the TDM Enviornment, it requires a portal. In TDM, this is called a voice. A voice is a one-way gate, or door. Signal may only pass through this portal in one direction, and the protal is only "wide" enough for one signal, or a mono signal. For a stereo signal, which is simply a pair of mono signals, TWO DOORS or voices are required to send the signal OUT of the TDM Enviorment. A 5.1 signal/stream requires SIX VOICES to enter or leave the TDM Enviornment!

Now, look at the TDM Enviorment...it can do lots of things. However, RTAS processing is a responsibility of the HOST system, it is a native process. To use the host system- CPU- for this, the signal must leave the TDM Enviornment. To stream audio form hard disk TO the TDM Enviornment, voices must be used. Stream in a mono signal from hard disk, one voice to get into the TDM Enviornment is used. Stereo? Once again, this is simply a set of two mono signals...thus, two voices required. 5.1? Yep, SIX voices to get into the TDM Enviornment.

RTAS processing is host based. Luckily, Pro Tools was coded with some common sense: if you are streaming a signal from hard disk, and want to apply RTAS processing FIRST, then it is simple: the signal leaves the hard drive, enters the buffer (host/native requires buffering; think of it as waiting in line...taking a number at the deli maybe...), and then can either go to the TDM Enviornment, or receive some RTAS processing and then enter. Since the signal originated outside the TDM ENviornment (remember- the 192 I/O..or 96 I/O...is simply a box attached directly to the HD card- it IS part of the TDM ENviornment; the host system is NOT), it will require a voice to get in. Since the RTAS process is being done as the first insert encountered, the software is smart enough to combine the stream, buffer, and the RTAS into one step, and thus use just the one voice (per path width number; stereo=2 voices, 5.1=6 voices, etc).

Now comes the problem: Audio tracks are the ONLY track type that can have a signal BEGIN outside of the TDM ENviornment. Every other type begins life inside the TDM ENviorment, and thus, if you want to apply some RTAS processing, you must send the signal from the TDM Enviornment OUT to the host for processing (using one voice per width), and once completed, use yet another voice per track width to re-enter the TDM Enviorment. So, a mono Aux In with an RTAS plug-in on it would need to use one voice to send out, and then one more voice to get the RTAS processed signal back in. Stereo? Two out, two back in. 5.1? Yep...a staggering TWELVE VOICES to get out and then back in!!! So, if you had three 5.1 Audio tracks streaming in (6 voices PER track), and then used 3 Aux Ins with an RTAS plug-in on each, you would then use an additional 12 voices for each of those tracks; total so far? 54 voices!!!

Place a TDM plugin, then an RTAS, then a TDM, and then an RTAS...and you would need to: voice to send for the first RTAS, voice to get back in; do TDM, then send back out AGAIN for the next RTAS, and then back in. Four voices per track width!!! If you did that same chain on an Audio track that was streaming in audio from disk, and you would have: Voice to get in; do the first TDM plugin in the TDM Enviornment; voice back out for first RTAS; voice back in to do the next TDM plug; voice back out for RTAS #2; voice back in; fnal TDM plugin, and audio goes out through the I/O interface. Total voice "cost?" 5. Per path width (twice as many for stereo, six times as many for 5.1- that would be 30 voices if it were a 5.1 track!!!)

Careful planning is required when doing this with HD1. TDM plugin power will be limited with one HD Card. Building the TDM Mixer takes HD DSP chips. The number of voices determines how many to begin with, and then how many Aux Ins and how much bussing is done further eats those DSP chips. Turn on Auto Delay Comp, and use the Long setting, and you are now eating several DSP chips for a (simple) 64 voice session. With not many chips left for TDM plug-ins, the user will need to rely on RTAS...which then brings all the crap I just wrote about into play; VOICES. Start using RTAS on those Aux Ins, and you have voices going fast. And I dont know about others, but I and most people I have worked with have a workflow that requires using send/return scenarios, and Aux Ins are a must for that type of thing.

Does that explain it? Although I have generalized a bit, the basic Laws of TDM are presented, and their direct impact upon RTAS is explained (basically). Since LE does not have TDM as a consideration, it does not need to obey any TDM laws. "Voices" are something almost entirely behind the scenes with LE, and since ALL processing is native, there is no need to utilize them for any processing needs, other than streaming the audio from hard disk into an Audio Track.

BTW- it is the extra stability that building that TDM Enviornment on those HD cards affords that makes TDM so attractive as a platform. Once the software running on the host system is also stable, the entire Pro Tools enviornment becomes relatively bullet-proof. I owned and worked on my HD4 Accel system for a few years, and I could count the crashes on a single hand...and still have fingers to spare. Crashes were almost always 3rd party plug-ins, or the result of me playing with some part of XP or doing something to test something.

So! This is why I say that I recommend an HD2 as a good starting point for a nice 64 voice session type user. Realize also that sample rate will jack up the DSP usage! No more voices required, but the DSP used by each voice, and by all the other DSP processes (ADC, mixer building, TDM Plugs, etc) will be double, or nearly double. So, working @ 24/96k will suck an HD1 dry in no time, and leave no, or nearly no, DSP for TDM plug-ins; thus, RTAS would need to be used. 5.1 @ 24/96k would be nearly impossible on an HD1, IMO. HD2 for large 5.1 sessions @ 24/96k would be unpleasant for me; HD3 would be a minimum for that IMO.

Any more quesitons, or any clarifications, please ask away! If i did not explain someting that well, or if osmething is cloudy- ask! Please! (and sorry if I was a bit long winded..lol...) (I really should write a book...lol...)

Quoted from that thread. Words by Nikki K

Big D

[ASPSA]

Thanks for your reply and for the thread reference. I will certainly read that thread fully, and I will check the technical documents on the TDM / TDM II architecture.

Here's a thought... There are numerous TDM plug-ins, and, collectivley, they handle a wide range of production needs. Based on the post you quoted, it is now clear why additional HD/HD Accel cards are required to handle a dense recording/mixing session, but this raises an interesting point. If one were to build a Pro Tools HD system, which is built around the TDM / TMD II architecture, then one may logically conclude that a top-of-the-line host computer is overkill. Wouldn't it be better to invest fewer dollars on a "less powerful" host computer and budget those additional monies for addition HD (Accel) cards? (Note: I apologize for the following question, since this is the TDM Windows forum, but I'd like to continue the discussion in this thread rather than cross-post it in the TDM Mac forum). What is a sufficiently good Mac Pro host in terms of processing power and RAM requirements?

Big D, you're a good egg. Thanks for your timme and efforts in answering my questions.

[nikki-k]

Thx for the post quote! The full thread is worth the read, IMO l)

OK...yes, it could be argued that opting for saving money in host computer choice vs HD cards might be a good idea. However, this would almost require that the host system be upported fully for the most current version of Pro Tools, and one would want to have some sort of reasonable faith in that system receiving support far enough into the future to justify the savings. IMO, a Mac Pro or Xeon based XP system is best, with eyes on support in two years or more. I would opt for PCIe, as I feel it has to be qualified very soon. It works for HD1 and HD2 (in my experience). If one were to use the system solely for audio, and not use virtual instruments, the host system could be chosen with lesser power, and thus lesser expenditure. With prices where they are, I would still go at least dual core and look at providing a good external drive solution (nice SATA 3G/s drives in an external case, with room for removables). If doing video, dual CPU's is a must. Xeons here are my choice.

What is the system to be used for? Is it a pro installation? If yes, then obviously grab a nice highpowered host, and add HD cards as needed. If this is more of a project studio enviornment, then grabbing a high powered host and an HD1 would be a great start. A 192 I/O should be a minimum when considering I/O interfaces IMO. You can always add an HD Accel process later, giving you more to work with.

I can work with an HD1 all day long. Seriously. But, I wouldnt be exceding 64 tracks, I would have ADC on Long when mixing, and I would have taken time to think through the project (not just a single session) and plan templates ahead of time, and plug-in usage. With an HD2 Accel, I simply have the extra elbow room to allow for some decently serious DSP allocation planning mistakes. It also allows me to use certain TDM only plugs that I really love, and have not found substitutes for yet. If you aren't used to TDN plugs yet, then you should have a decent number of RTAS to choose from. I love ReVibe. I have not found an RTAS sub that is as nice yet. TL Space native works, but is a different beast. With BF Reel Tape suite coming soon, that gives me a cool Delay and Flange. I only have a few more to find RTAS subs for and I would be able to go native, and possibly get by an HD1.

I have a Mac Pro. I sold my old XP rig (single P4) and went this route because I use virtual insturments, and my primary use for my rig is composing. I prefer an HD2 Accel in it, but could scrape by with an HD1. If I were doing full band recording, regular clientelle list and all...I would seriously want the HD2 Accel...but still have the high powered host, such as the Mac Pro.

Using RTAS on Audio Tracks is totally cool IMO. Works great, and uses no extra voices (if you dont use any TSM before them on those tracks). Aux Ins and the other track types will grab two voices when you use RTAS. Render virtual instruments, and then all you have left on non-Audio Track types are submix effects and "grouped" type effect send-return type Aux-Ins. Done with pre-planning, and *most* music type sessions are fine. When mixing, crank the buffer up, and the host should be capable of providing DSP power greater than an HD7 system with no glitches.

It all depends upon the user to do a little "footwork" to learn about what they will be using, doing a bit of pencil work to figure out their intended uses and such, and all should be very well!

[BIG D]

ASPSA,

you are welcome, and the new Post here by Nikki-k once again gives all the answers you need.

I am running a dual core AMD 4400+. I also run my HD cards in a Seperate chassis and by having the HD4 I never really have to think too much about the process. I just keep adding and opening tracks etc as I need them. It is great to not get in the way of the Creative process, but as Budget dictates, Nikki's pre planning on the HD2 is THE answer. I do use a fair share of Virtual RTAS instruments. Pretty much BFD, Virtual Guitar 2, Sample Tank 2.1, Lots of XPAND, Velvet, Essential Strings and Kompakt player, in pretty much all my sessions all used via the FXPansion VST Wrapper. I find I have to Track the Virtual Guitars to get more than 2 or 3 instances but everything else I keep pretty much real time.

From conversation I have had recently with a close friend in Retail here in Australia, the Version 8 Vista compatible Pro Tools should only be a couple of months away, (maybe longer who really knows till it lands) and that will really make for some powerful Dual Quad systems on both MAC and PC platforms. Exciting times really, especially for us old school guys that started out on TAPE in the 70s!!

All the best,

Big D

[BIG D]

Hi Nikki,

You are welcome on the quote. I always like to pass on the good information which yours always is. Thanks a million for your input and the time spent on research.

Big D