How many XITE-1s?

[rafafredd]

Hi. How many XITE-1s can anyone use in parallel for more I/O?

[garyb]

right now they cannot be stacked like that.

[rafafredd]

That is a real shame, as in I/O counting the new system is actually much worse than the old PCI cards... As I could get up to 108 channels of I/O with the old cards.

[garyb]

it's possible that there could be a madi expansion in the future, but it's not confirmed yet...

[HUROLURA]

Being able to "stack" several Xite-1 by using XTDM connectors as we former use STDM connectors to link up to 3 Scope PCI card would also be nice as it would provide a solution to increase DSP power and IO.

In case of designing a MADI rack, I would suggest that it provide severall XTDM connectors at the same time to allow the use as a bridge between Xite-1 units and A16 MK II units. And you could just call it Scopebridge ...

[dante]

Hurolura, was anything like that hinted at during MusikMesse discussions ?

[HUROLURA]

No, just my own opinion ...

[wouterz]

Yes, more IO would be great

[Bud Weiser]

it's possible that there could be a madi expansion in the future, but it's not confirmed yet...

You mean, IF,- it would be for the XTDM bus ?

Would make search for and purchase of old and used A16 Ultra obsolete !

So, yes,- we should get it for a Ferrofish A16mkII stack,- w/ MIDI over MADI please !!!

Bud

[siriusbliss]

A XTDM extension for linking between Xite and the old PCI cards was discussed and requested way back when Xite was released, but alas I think there are technical limitations.

I would vote instead to support SC in development of MADI or some sort of PCIe bridge idea.
Or even Thunderbolt.

Greg

[garyb]

yes, it could be over the XTDM bus.

[rafafredd]

Well, being a practically no-latency audio system, one of the very nice uses for the old PCI cards with many inputs and outputs was to use as theater live sound. One integrated system for both main and monitors. A simple system with three LUNA + three ADAT/Zlink expansions could do that nicely and cheap. But one XITE1 can't do it because the very limited number of I/Os. That is why I think I/O expansion is a must. If more XITE1s could be bridged together, also for more DSP, well, it would be even better.

[HUROLURA]

An XTDM to MADI converter would do the trick as it would allow to stack A16 Mk II for I/O expansion.
But allowing DSP expansion through XTDM would also be nice ...

[Sounddesigner]

I know that in the FAQ's on Sonic Core's website it states that the current software can't support multiple XITE-1's connected together; So there will be needed a update to handle more dsp's in the future if this is to occur, here's link - http://sonic-core.net/joomla.soniccore/ ... 56&lang=us

I do believe we will NEED to be able to chain atleast 2 XITE-1's together soon for more power as time passes and hungry plugins are brought to the platform. Plugins development has been slow wich is natural since Sonic Core had to first worry about the underlying platform stuff (SDK, XITE-1 Hardware, Windows 7 32bit/64bit support, Modular 4 building blocks, Open-SCOPE, etc) but once a few new heavy-hitting plugins come many users will need more power (more users will also want to work at 96khz as time passes - natural progression). I know the I/O need is already here but due to the next generation of plugins comming very slowly cause of 'bedrock functionality' needing to be addressed first there has'nt been a huge need for more dsp power imo. We were blessed to have many great plugins already available for SCOPE to hold us off until the futuristic stuff comes and they are'nt 'extremely hungry'. But more dsp power will be needed soon and S|C will have to address this, and i'm sure they will allow for multiple XITE-1's to be chained together to keep up with the times since switching to the latest generation of DSP's is'nt a smart option (too much extra development for them and there is no significant power gain unless you wait for many generations to pass).

I do believe the era of needing 3 or more dsp cards in a single system should be over, that's pre-historic operation (too much of inconvenience, can become more and more difficult to find compatible motherboards, and too many single-points-of-failure). IMHO 2 XITE-1's should be enough for most users who need more than 1 (Or S|C could build a bigger XITE-1 wich holds like 30 dsp's, but still would need to address I/O problems). I'm confident something will be done by S|C to address both I/O and dsp power problems when necessary.

[HUROLURA]

And what about a 2U Xite-1 device (similar to the A16U XLR) including 2 Xite-1 board with one internal additionnal board which allow Xite to Xite communication ?
Maybe easier to achieve while still mainly using standard Xite-1 boards ...
That would already mean twice DSP power and twice I/O.

[Bud Weiser]

An XTDM to MADI converter would do the trick as it would allow to stack A16 Mk II for I/O expansion.
But allowing DSP expansion through XTDM would also be nice ...

It all depends on manpower and marketing strategies IMO.

Before thinking about interconnecting 2 XITE-1/1D devices via XTDM, which means DSP assignment across 2 hardware devices, this should be optimized 1st for the usage of only ONE XITE unit.

Following the evolution of JB Solaris, to me it seems there´s some kind of parallel developement.
There was a discussion about making Solars a 4 part MIDI multimode instrument and adding higher polyphony.
As far As I understood, this is dependent on optimized DSP assignment and some kind of "DSP offload" functionality.

If this will happen,- not only Solaris polyphony and functionality will be improved but also XITE-1 /D´s.

We´re not at the end w/ XITE-1/D DSP power,- in fact it can be optimized to get more out of it and maybe we´ll see that w/ SCOPE 6 already.
I don´t think it makes sense interconnecting 2 not fully optimized machines, but it makes sense to extend physical I/O connections by adding a XTDM/MADI bridge for the usage of several A16mkII.

For more DSP power, I´d wish we´ll see some kind of a 1HU / 19" blackbox XITE booster, maybe just only w/ 2 XTDM connectors and a MIDI trio (In/Out/Thru) to keep it reatively cheap and not before all the other developement and optinization is done.
With a fully functional XTDM bus and MADI over XTDM,- it might be possible to run XITE-1/ 1D, a booster and additional A16mkIIs in a chain in (near) future, for Audio and MIDI.

For the time being, buying 2 A16mkII for only 1 XITE-1 or XITE-1D would be a ideal solution and only a fraction of the investment compared to buying a 2nd XITE-1.

So, YES to MADI for XTDM and don´t forget the "MIDI over MADI" because each A16mkII already adds MIDI connectors itself.

Bud

[Bud Weiser]

And what about a 2U Xite-1 device (similar to the A16U XLR) including 2 Xite-1 board with one internal additionnal board which allow Xite to Xite communication ?
Maybe easier to achieve while still mainly using standard Xite-1 boards ...
That would already mean twice DSP power and twice I/O.

I don´t think owners of a XITE-1 will buy a new and much more expensive hardware unit soon.
Personally, I don´t want to see a new machine leaving XITE-1 / XITE-1 D units in the dust.
I more wish for improvements, expansions and connectivity of actually existing gear !

How many XITE-1 have been sold worldwide up to now ?
How many customers will buy a double powered XITE-1 mk II at which price tag ?
A all new box will reduce the price of a XITE-1 and XITE-1 D and it will reduce the re-sale price of these too.
I don´t think that´s in the interest of S|C as also not in the interest of owners of a XITE-1 or XITE-1D.

Bud

[HUROLURA]

I guess the dual Xite-1 I imagined would cost slightly more than 2 Xite-1 units (2 Xite-1 boards + addtionnal communication interface between them) and shouldn't bother much the Xite-1/1D sales (similar to A16U XLR vs A16U).
SonicCore would "only" have to use Xite-1 board (already available) to offer such thing.
But I wouldn't expect SC to design a new dual Xite-1 board.
This would only have sense if it is possible from the original board design to insert an XTDM internal bridge but keep on using a 99% standard Xite-1 board (similar to Xite-1D vs Xite-1: same board just underpopulated).

Regarding Solaris and Xite-1 optimization, my assumption is that Xite-1 is already more "optimized" than the Solaris is because Scope allows to unload DSP with unused modules.

On the Solaris side we have 6 big DSP (12 on the Xite-1, 4 on the Xite-1D). But the design is using a similar philosophy as the one used previously for the NOAH (where 2 old DSP were used for each slot, 1 for "host emulation" and the 2 remaining for I/O mixer and FX): you allocate the DSP to handling a certain task. On the Solaris, each DSP is handling 2 notes. 5 DSP are achieving this task while the last one is in charge of FX and I/O. The first thing that could be done on the Solaris would be to make it multimode capable (4 part as specified). I wouldn't expect dynamic voices allocation of the DSP but a fixed setting like what was available with the Noah Ex 4 slots.
Then, if your patch is only using less than half of what is available in term of oscillator/mixer/filter on the Solaris then it should be possible to raise the polyphony accordingly (easy to write, probably much more complicated to achieve). This is just similar to what is done in Access Virus (max 80 notes polyphony decreases as soon as you increase patch complexity) or other VA synth (my Waldorf µQ is capable of up to 25 notes polyphony but if you use some special high speed modulation (oscillator or filter) the polyphony decrease). The current SOLARIS software is just set to worse case use, thus limiting the polyphony to 2 notes per DSP.

[siriusbliss]

An optimized Scope 6 system to further expand the DSP flow would be a good next step, and then some simple RME-type expansion interface hub would be more practical at this point IMO.

Greg

[HUROLURA]

Further scope Xite optimization would probably imply loosing old scope PCI compatibility or managing a different code depending on the target used (scope PCI or xite).

This would allow to use Xite-1 on board external RAM (good for delay and reverb FX as this would lower the PCIe bandwidth used) and use the specific features of the new DSP to further enhance the computation power. But this is a lot of work to achieve such thing as the DSP atom modules are quite numerous !!! And it would also allow the design of new DSP atom using the additionnal features.

I think there is room for enhancement in the Xite but SC shouldn't forget the old Scope PCI user base which is probably still more important than the Xite user base. Unless they offer a new lower cost hardware to provide a more entry level access to Scope to allow a cheaper crossgrade from 21065 to 21369 DSP ...