NEWS: NEMESYS/ReVALVER PLUG-IN DEAL SIGNED

[Jim Van Buskirk]

FOR IMMEDIATE RELEASE

ReVALVER PLUG-IN FOR GIGASTUDIO TO LEVERAGE NFX KERNEL MODE DSP ARCHITECTURE

Austin, TX USA (May 11, 2001) — NemeSys and Alien Connections announced a joint-agreement today enabling NemeSys to distribute the ground-breaking ReValver music amplifier modeling effects processor for NFX based platforms. Alien Connections shall produce NFX-compatible software plug-ins, leveraging the zero timing latency, computational efficiency and rock solid performance unique to the NemeSys kernel mode DSP architecture.

\"The power of the Alien Connections product will now finally be realized in an extremely playable package for guitarists, synthesists, and sampling enthusiasts ”says Jim Van Buskirk, President of NemeSys Music Technology, Inc.. “Boasting the quality and playability of the finest amps and effect processors, ReValver NFX edition will render the DSP hardware vs. software debate irrelevant.”

“Let’s face it, NemeSys has redefined the expectation of software sampling to an all time high. Now that same real time expertise is being applied to DSP, and Alien Connections wishes to be a major part of that,” said Michael Ljunggren, President Alien Connections Sweden.

Unlike other popular plug-in architectures for software effects and musical instruments that run at a high level on top of many layers of software, the NemeSys NFX architecture runs “close to the metal” at the low level kernel of the operating system. This fundamental difference matches the architectural design of dedicated hardware DSP processors, giving NFX products the same latency, throughput and robustness.

System requirements include an NFX-based GigaStudio 96 or 160 system, a Win9x or WinMe based system to GigaStudio specifications, and a GSIF compatible sound card with MIDI. For additional information including a list of available GSIF sound cards, visit www.NemesysMusic.com (\"http://www.NemesysMusic.com\") .



Founded in 1996 and based in Austin, Texas (New Media Systems) NEMESYS MUSIC TECHNOLOGY, INC . is a leading producer of authoring and delivery systems for the Audio, Music, and Multimedia Markets. NemeSys has core expertise in audio DSP, media streaming technology, disk engine technology, PC system programming and architecture, user interface design, and product development. For more information, visit NemeSys\'s web site at www.NemesysMusic.com. (\"http://www.NemesysMusic.com.\")



Alien Connections Sweden is a Stockholm based startup company. The aim of the company is to do research in the field of signal processing in both software and hardware. This includes ASIC/FPGA design and verification via consulting services and custom signal processing designs in both hardware and software. Some software has been made publicly available, currently covering audio signal processing. www.AlienConnections.com (\"http://www.AlienConnections.com\") ###

[JWink]

I bought the DirectX version of ReValver, months before it was included in Sonar... What I want to know is: How much MORE am I going to have to pay for an NFX version??? Time to write to Alien Connections....

[Haydn]

How does this fit in with future plans for GigaStudio working on Windows XP or 2000. These operating systems do not allow kernal mode programs to run on them. That is why Cakewalk is going to DXi instruments. DXi instruments work with DirectX 8.0 allowing low latency. Personally, I want to move to Windows XP as soon as possible. I love the stability of Windows 2000 which I use on many machines at work without any problems unless the hardware breaks (hard drive crashes mostly).

[SCARBEE]

Very good news! I have been waiting for such a move - congratulations!

Will there be a Bass-amplifier/speaker simulator too?

When will the first products ship?

Thomas

[This message has been edited by SCARBEE (edited 05-12-2001).]

[Sam]

So is this the first 3rd party NFX plug? I still wish they\'d just support VST plugins. I monitor live input through VST FX all the time and can\'t discern a latency difference between these and hardware (I ususally run VST with 3 ms latency, I\'d be surprised if NFX is substantially better).

In fact my setup does allow me to monitor gigasampler through VST; with my digital mixer I just bring the GS channel back in a group and apply FX there. Still no problem with latency, and a lot of the free VST FX are really good.

And Thomas, I also have an FX send to a Sansamp, your bass does sound pretty cool through that...

[Jim Van Buskirk]

>How does this fit in with future plans for GigaStudio working on Windows XP or 2000. These operating systems do not allow kernal mode programs to run on them. That is why Cakewalk is going to DXi instruments.>

Haydn, the NFX DSP architecture does indeed allow for kernel mode processing in Windows XP. In contrast, the big three sequencer companies all have their plug-in architectures sitting on top of the OS in the application space, including DXi. The NFX architecture provides a fundamental difference, but that doesn\'t mean that there isn\'t plenty of inaccurate information being communicated - which can smear this simple fact.

SCARBEE, I have played my wife\'s Rickenbacker through ReVALVER with great success - (OK, what\'s a drummer doing holding with a Ricky anyway?) The NFX ReVALVER is targeted for a Fall release.

Sam, there are several developers now working on NFX products. The installed base has grown tremendously in the last year, and the market is hungry for NFX. We recognize that, and will offer only the finest products.

As far as your user mode plug-in setup, remember that the 3ms setting is only the final h/w buffer size. The actual latency must consider all stages of the audio chain, from input all the way through all chained effects, add to that the MIDI latency (if it is a plug-in instrument), then there are stages of buffering to safely get you through the user mode. After all that, the 3ms can then be added times the number of stages between the read and write pointers.

To put it simply, there are NO hardware DSP processors that are implemented within a high level OS. They all have low level, hard real time, highly optimized scheduling kernels that run at the same level as the NFX architecture, albeit a different processor and OS. In strict architecture terms, they are the same.

JWink, we will look into your situation, however we expect the NFX version, due to the architecture, to be significantly better.


[This message has been edited by Jim Van Buskirk (edited 05-12-2001).]

[SCARBEE]

[QUOTE]Originally posted by Jim Van Buskirk:
SCARBEE, I have played my wife\'s Rickenbacker through ReVALVER with great success - (OK, what\'s a drummer doing holding with a Ricky anyway

Bravo - this has been a major wish for me and my bass libraries!

I use the Rickenbacker a lot on film/commercial jobs. A lovely \"swamped\" boomy sound. I would love to sample that next year...

Good work Jim!

Thomas

[Sam]

Jim, I think you either miss my point or are deliberately avoiding it. As long as overall latency is under about 10ms, I find real time monitoring to be fine, pretty much undetectable. I believe 10ms corresponds to about 7 feet at the speed of sound, ever play guitar with the amp more than 7 feet away?

My 3ms figure is the vst output buffer latency, with which the audio output is flawless. I can set this latency down to 1.5 ms, which will work fine often but perhaps not under all loads. I have no idea what the input and output buffer latencies are into VST and between FX, nor whether such delays are added to my output latency or taken into account. In practice it\'s a non-issue, they are \'playably\' fast, and I am sensitive to keyboard response or voice latencies or whatever. Anyway, the most common FX are delay based anyway (delay, reverb, chorus, flange). Most reverbs sound best with 25 ms or more of predelay.

Can kernel mode FX give you better latency? Sure, they better. But I for one am not clamoring for yet another plug-in standard. I believe GS would be better served by the quality and breadth of VST plugs, and I\'d rather purchase such plugs that can work in a wide variety of audio applications.

Anyway, this is not a huge point for me, my system has pretty excellent support for GS (beyond what most people can do I suspect, ie I can rewire any programs through vst) so I can play GS through VST anyway.

BTW, I am familiar with the advantages of kernel mode drivers, I\'ve made a living doing unix device drivers and real time communication software. And I do appreciate what gigasamplers kernel extensions bring me, which has been great responsivness and total reliability (see I\'m not down on GS, it\'s a superior product).

There is one exception though - My system can\'t play the game Descent 3 with the GS kernel extensions loaded, that will lock it up consistenly. My video card has 3D accelleration with kernel enhancements too and apparently there\'s some conflict with all this real time stuff that will always lock it up in < 10 minutes. No biggie, games and some video acceleration have gotten the boot from this system, but it just shows that too much real time kernel stuff can lead to conflicts. I\'ll also note that VST has never crashed on me, people slag on Steinberg, but I also think they do first rate work.

cheers,
-sam

[David Abraham]

<BLOCKQUOTE><font size=\"1\" face=\"Verdana, Arial\">quote:</font><HR>Originally posted by Jim Van Buskirk:
>
Haydn, the NFX DSP architecture does indeed allow for kernel mode processing in Windows XP. In contrast, the big three sequencer companies all have their plug-in architectures sitting on top of the OS in the application space, including DXi. The NFX architecture provides a fundamental difference, but that doesn\'t mean that there isn\'t plenty of inaccurate information being communicated - which can smear this simple fact.
<HR></BLOCKQUOTE>

but the fact is they are working. I don\'t give a hoot about kernel mode techno-babble. This should be about working solutions. Empirically speaking the VSTi/ASIO and DXi/WDM implimentations are just as tight from a low latency perspective as GSIF. And they are already on Win2k. Talk tough after you deliver the goods.

-david abraham

[Jim Van Buskirk]

>I don\'t give a hoot about kernel mode techno-babble. Empirically speaking the VSTi/ASIO and DXi/WDM implimentations are just as tight from a low latency perspective as GSIF. And they are already on Win2k. Talk tough after you deliver the goods.>

D.A. Fenton, most musicians that I know do certainly give a hoot about the ability to have 160 streaming audio voices with excellent playability (low-latency), all with high quality interpolation, resonant filtering, envelopes, etc. and with a half dozen or so effects running all concurrently on a 800 Mhz system (like my Dell 4100 at home, which is about an $800 system these days.

If there is any other host based software system on the planet which can deliver this level of real time streaming and DSP performance, then peace. Otherwise, it seems only fair to say that NemeSys has indeed delivered the goods (i.e. an architecture which can support this level of system stress without compromising playability)and is continually advancing the features to make additional use of this excellent architecture. The techno babble is just an attempt to provide an explanation why we feel the NFX architecture is needed for a high performance sampling system.



[This message has been edited by Jim Van Buskirk (edited 05-13-2001).]

[David Abraham]

<BLOCKQUOTE><font size=\"1\" face=\"Verdana, Arial\">quote:</font><HR>Originally posted by Jim Van Buskirk:
>I don\'t give a hoot about kernel mode techno-babble. Empirically speaking the VSTi/ASIO and DXi/WDM implimentations are just as tight from a low latency perspective as GSIF. And they are already on Win2k. Talk tough after you deliver the goods.>

D.A. Fenton, most musicians that I know do certainly give a hoot about the ability to have 160 streaming audio voices with excellent playability (low-latency), all with high quality interpolation, resonant filtering, envelopes, etc. and with a half dozen or so effects running all concurrently on a 800 Mhz system (like my Dell 4100 at home, which is about an $800 system these days.

If there is any other host based software system on the planet which can deliver this level of real time streaming and DSP performance, then peace. Otherwise, it seems only fair to say that NemeSys has indeed delivered the goods (i.e. an architecture which can support this level of system stress without compromising playability)and is continually advancing the features to make additional use of this excellent architecture. The techno babble is just an attempt to provide an explanation why we feel the NFX architecture is needed for a high performance sampling system.

[This message has been edited by Jim Van Buskirk (edited 05-13-2001).]<HR></BLOCKQUOTE>

well I\'m talking about me a person who bought Nemesys technology with a false promise about Win2k support over a year ago. Bought your software only to get hosed with your copy protection scheme, never to have many phone calls/e-mail returned, and when the product did finally get installed it did not work as advertised, but that\'s all old news because no one listened when I reported it. So yes maybe your elite cliente give a hoot. Maybe you give them favorable treatment maybe not. But the big three have delivered integrated sample accurate synthesis/sampling, 2 of them on win2k and all I hear from Nemesys is excuses about kernel mode, blaming Microsoft and worst of all rumours about lawsuits against a particular vendor who is seeking to deliver an integrated streaming solution about which many musicians that I know \"give a hoot\".

For all I know Nemesys is planning to deliver their own integrated MIDI sequencing/editing platform...maybe that\'s why they feel it necessary to belittle the working integrated solutions that many musicians have been craving for years. Maybe you don\'t know *those* musicians.

So from my perspective no you have not delivered what you promised. The good thing is now I know exactly where you\'re headed. GigaX will be a closed high performing system, like Propellerhead Reason, this has it\'s pros and cons. That\'s fine. It will work, I\'m sure it will sell well. Just not to me.

thanks for your time

-david abraham

[Kenn159]

Jim are you heading towards a closed system with Nemesys giga studio?

Ive heard a few people hear mention that they wish that Nemesys would implement a universal standard for there effects like VST or direct X instead of NFX.
I agree that the NFX reverb doesn\'t sound as good as some of the great DX reverbs like Waves Trueverb[which uses alot more DSP too].
But realize when comparing its like comparing apple and oranges .
The direct x and VST are audio effects and the NFX are midi effects .
If cubase is anything like my Cakewalk the midi effects are very limiting in what the can do , so comparing to other midi effects that I have seen the NFX , they fair pretty well.

[Chadwick]

Kenn - NFX aren\'t midi effects, they\'re definitely audio dsp stuff.

Jim: I\'m really bored with this chant of \'make Giga VST/DXi etc., compatable\'.

I would like to be able to integrate the huge world of plugins as much as the next guy, but if it can\'t be done without unacceptably compromising Giga performance - then it can\'t be done.

Perhaps you could give us a hypothetical comparison of what Gigastudio\'s performance is like now (with NFX as its \'plugin\' structure) and what Gigastudio\'s performance WOULD be like if you did what people ask - ie make Giga talk to plugins via VST or DXi.

I assume you\'re answer would be something along the lines of higher, unpredictable and unacceptable (for real time processing)latency.

I\'d just like to have a definitive explanation of the basic advantages/disadvantages in the two approaches for the end user (like me)who doesn\'t understand the difference between a system which sits on top of an OS and a system which sits at kernel level.

I think this would be a good time to do it, as I sense this vibe that there\'s a conspiracy to make the NFX a unique non-standard approach simply so that Nemesys can make money out of licensing or some such crap.

Personally, I\'m pretty sure you\'d stick VST/DXi in if it worked, I just figure it\'s too slow for a realtime sampler which is trying to act like a 160 voice E4. But I don\'t understand enough to make a credible comment.

[Haydn]

I\'d still would like to know how Nemesys is getting their software to work in kernel mode directly. I just read in a book that Windows 2000 (also XP) will not let a program directly access kernel mode and will cause the program to not run if it attempts to access kernel mode. It considers programs that go directly to kernel mode as not written properly. I have read this in a couple books. How is Nemesys getting around this issue or is this why it is taking so long for a Windows 2000/XP version?

[gigaDiga]

I don\'t know what you guys are all worrying about. GigaStudio is already compatible with realtime DirectX effects. Sorry Jim but NFX is dead in the water. I\'m sitting at home right now routing all my GigaStudio outputs through Sonar\'s new `input monitoring`. By doing this I can attach multiple DirectX plugins to the signal and, using DirectX v8, I can fully automate nearly all the parameters in realtime. End of story.

There\'s no way I\'m gonna be a fool and buy into a new plug-in format when an old standard is delivering EXACTLY the results I want. http://www.northernsounds.com/ubb/NonCGI/images/icons/smile.gif

[Chadwick]

Gigga - what\'s the latency like?

Are you playing through the input monitoring and adding effects in real time or are you adding the effects after tracks are recorded?

If what you say is right, what\'s all the hoohaa then?

[David Abraham]

<BLOCKQUOTE><font size=\"1\" face=\"Verdana, Arial\">quote:</font><HR>Originally posted by gigaDiga:
I don\'t know what you guys are all worrying about. GigaStudio is already compatible with realtime DirectX effects. Sorry Jim but NFX is dead in the water. I\'m sitting at home right now routing all my GigaStudio outputs through Sonar\'s new `input monitoring`. By doing this I can attach multiple DirectX plugins to the signal and, using DirectX v8, I can fully automate nearly all the parameters in realtime. End of story.

There\'s no way I\'m gonna be a fool and buy into a new plug-in format when an old standard is delivering EXACTLY the results I want. http://www.northernsounds.com/ubb/NonCGI/images/icons/smile.gif<HR></BLOCKQUOTE>

Yes, congratulations...that works

1) but input monitoring is more \"expensive\" from a CPU performance perspective.

2) It requires a working multi-client driver solution or multiple audio cards

3) Depending on HW it can involve double the AD/DA conversions


You could do much better with a true DXi or VSTi interface.

I appreciate the fact that the Giga architecture can\'t readily support this..but it *is* something that many musicians want. many of whom don\'t need 160 giga style voices. So yes they can quietly look for other solutions, you know keep the peace, right? then you find out about this patent situation so *nobody* can have this functionality???? how is that helping consumers? Nemesys can\'t deliever it, the patent holder wants to prevent anyone else from delivering it...then they want to come and slam the good technology that tons of people are already using?

-david abraham

[gigaDiga]

Upon reflection I\'ll step back from my statement. I personally have a Pulsar2 card. This enables me to put every output of GigaStudio through a seperate audio channel. It is this feature which enables me to put seperate effects on seperate instruments via Sonar. If I only had a 2 or 3 client multi-client card I wouldn\'t be so contented.

Chadwick... I can route the live outputs of GigaStudio directly through the DirectX effects in Sonar without recording. I can twiddle DirectX parameters in real-time without any problems or need to record audio.

Sonar will totally replace NFX for me. I say `will` because currently, without WDM drivers for Pulsar, latency is 300ms. In a couple of months time Pulsar will have WDM drivers.... THEN, not now, it will be end of story as far as I\'m concerned and latency should be 7ms.

What I meant to say by my first post was that Sonar has solved the DirectX compatibility problem with GigaStudio as long as you have a Pulsar with WDM drivers. When I look at this statement now I realise that, although I might be happy in a couple of months time, it is not really the solution that all users want to hear.

In the light of this I must say that it seems bizarre that we musicians are quite happy streaming our audio tracks through VST or DX, and yet now we\'re told that we must stream GigaStudio\'s audio through NFX. I personally will be happily sitting at home streaming my GigaStudio through DX, whereas you guys have got to buy all your plug-ins again in NFX format.

[Jim Van Buskirk]

Chadwick, it is certainly true in any real time processing system for audio that the weakest link in the chain (as far as robustness) brings down the entire system to that same level of weakness.

GigaX products have always been founded on two basic philosophies:
1.) Software creates an opportunity, not a disclaimer
2.) Innovate, never imitate.

On point one, this means that our products will at least have the same performance of hardware - period. No excuses. If another plug-in system can be used without compromises, then so be it. On the other hand, if another system introduces a weak link into the chain, bringing down the whole system, then we\'ll take a pass. Otherwise, it would be like asking musicians to render their masters that they have worked on so hard to 11kHz 8bit audio, because a few people out there only have an 11kHz 8bit playback system. What about everyone else? It\'s all art, and we don\'t want to see it butchered unnecessarily.

If some users can tolerate latency that is worse than a piece of hardware or other compromises in polyphony, etc., then there are existing solutions for them - as some of these are mentioned on this thread. NFX is simply intended for those who can\'t tolerate sub standard (i.e. below hardware-level) performance - and NFX offers a unique option in this regard. More choices, that\'s all.

On point two, today there is no plug-in standard (other than NFX)that doesn\'t bring down the concurrent real time robustness and playability of the entire system, hence the need for innovation. Sure you can have either low latency OR robustness with user mode, but not both. The practical reality varies from system to system - but no matter what, with a user mode solution you will either be leaving MIPS (CPU processing) on the table or buffering up long as needed to overcome pops in the audio - at the expense of playability.

I hope this clarifies the intent. More choices, with the option of \'no compromises\' solutions featuring a premium portfolio of unique technology.

[SCARBEE]

Why not let NemeSys experiment with a new standard like NFX? After all they did come up with GigaSampler - which is a milestone in sampling. I believe that innovation is a very big part of NemeSys and they will keep on trying to find new ways of doing things, cos they \"have\" to.

Perhaps the new NFX/ReVALVER will be a great succes - perhaps not, but at least I think they deserve the right to try this out - without getting bashed for it.

Thomas

[David Abraham]

<BLOCKQUOTE><font size=\"1\" face=\"Verdana, Arial\">quote:</font><HR>Originally posted by SCARBEE:
Why not let NemeSys experiment with a new standard like NFX? After all they did come up with GigaSampler - which is a milestone in sampling. I believe that innovation is a very big part of NemeSys and they will keep on trying to find new ways of doing things, cos they \"have\" to.

Perhaps the new NFX/ReVALVER will be a great succes - perhaps not, but at least I think they deserve the right to try this out - without getting bashed for it.

Thomas<HR></BLOCKQUOTE>

Hi Thomas:

Nemesys can and will do whatever they want, that\'s their privilege, and they have great technology. But note the comment above about those of us using \"substandard\" solutions. I have lots of hardware around here. Route audio of one of my Roland JV modules to the mixer, and route the output of the NI B4, runnning through SONAR with a Mixtreme WDM driver. Using one controller drive both the hw and the sw synth and play. The feel the same at 8.7ms or lower. What does GSIF guarantee I believe 3-10ms. I\'m not aware of many hw modules that support 160 voices, or even 128.

But sure maybe the B4 is substandard, maybe SONAR is substandard, maybe Pro52 is substandard, maybe ASIO, Nuendo, Logic, EASI are all substandard. Maybe every single solution out there is substandard *except* for Giga technloogy. But no doubt Nemesys\' attitude towards a significant proportion of paying customers is *definitely* substandard.

-david abraham

[SCARBEE]

Hi David,

I am not an expert in \"FX-technology\", but I believe, that there is a difference between having a sequencer/audio using FX and a \"sampler\" - which has to react in \"real time\". Sonar, Logic and VST can use Direct X, VST, etc. great on their audio tracks and on their \"instruments\", but it is apparently not \"close enough\" (to hardware FX latency) for Nemesys - they sets very high standards for their Giga products - that is why they \"invent\" and go the hard way. Doesn\'t Ferrari work the same way?

\"Sub-standard\" is a word to compare software versus hardware solutions - FW-wise - I don\'t think it was so negative meant.

Let\'s hope that the new NFX will be cool and lets keep a nice spirit on this topic.

Thomas

[David Abraham]

<BLOCKQUOTE><font size=\"1\" face=\"Verdana, Arial\">quote:</font><HR>Originally posted by SCARBEE:


Let\'s hope that the new NFX will be cool and lets keep a nice spirit on this topic.

Thomas

<HR></BLOCKQUOTE>

see that, you calmed me down, grrr! not easy to do http://www.northernsounds.com/ubb/NonCGI/images/icons/smile.gif

-david abraham

[Chadwick]

The nub of this argument seems to be that some users are unconcerned about a bit of latency as, once they\'ve recorded, they can offset everything in their sequencer so that it sounds synchronised, whereas there are others who really can\'t deal with real time latency over 10ms as they are playing their parts LIVE into a sequencer and need the quick response so that they can interact with the rest of the arrangement.

Personally, after hard disk streaming and sound quality, my ABSOLUTELY PRIME reuirement is TIGHT timing and hardware quality latency. That\'s why I got a Pulsar board - to take the weight off the CPU (although I\'m not sure that my timing is all that tight yet).

[This message has been edited by Chadwick (edited 05-14-2001).]

[gigaDiga]

Hmmmm.... perhaps I didn\'t make myself clear.

When WDM drivers appear for Pulsar my effects latency should be down to 7ms. This is nicely under the 10ms mark.

The ONLY advantage NFX would have over this system would be a reduced latency of say 5ms. Now if you do the homework I\'m sure you\'ll find no audible difference. They\'ll both sound as tight as each other. The only difference is that you\'ll have a piece of paper telling you one has lower latencies than the other.

NFX PRO POINTS: Lower latency by 2-3ms.
DX PRO POINTS: Tight integration into Sonar (when you automate parameters you select the parameters you want and automate them, rather than having to rig it up to a midi controller then remember which controller it was rigged to. Also when you save a song you save all the FX settings with it), A huge library of 3rd party products & free devices, proven high quality reverbs, no noticable difference from the latency of NFX.

Now I\'m not gonna `down` any new developments on the NFX front. Go ahead do it... but I know what `substandard` system I\'ll be using.
I just worry sometimes that Nemesys is a company that just wants to innovate and doesn\'t actually want to finish anything off. Several years ago all we had was GigaSampler and it was great!. A whole load of us bought into it because it was the future. Now we\'ve got a slightly advanced GigaSampler with SConvertor, QuickSound, Distributed Wave, and NFX chucked on top of it. If all of these were seperate products I wouldn\'t have bought them and I think you\'d be suprised at how many others wouldn\'t have too!

Nemesys. Stop innovating and start developing your core product:- the sampler!

[David Abraham]

<BLOCKQUOTE><font size=\"1\" face=\"Verdana, Arial\">quote:</font><HR>Originally posted by Chadwick:
The nub of this argument seems to be that some users are unconcerned about a bit of latency as, once they\'ve recorded, they can offset everything in their sequencer so that it sounds synchronised, whereas there are others who really can\'t deal with real time latency over 10ms as they are playing their parts LIVE into a sequencer and need the quick response so that they can interact with the rest of the arrangement.

<HR></BLOCKQUOTE>

I regularly play non-giga softsynths at 5.8ms or lower, LIVE into SONAR. Modern processors are definitely up to the task. The benefits of integration are enormous. The most important one is the relief from multiclient driver troubleshooting, hubi\'s, MIDI Yoke, GSFIX and related distractions.

There\'s also sample accurate timing, instant project recall, ability to perform instruments in the context of a Virtual mix, integrated bouncing (capture) etc etc


-david abraham




[This message has been edited by David Abraham Fenton (edited 05-15-2001).]

[Jim Van Buskirk]

>gigaDiga wrote:The ONLY advantage NFX would have over this system would be a reduced latency of say 5ms.>

gigaDiga, I can see that there may still be some confusion over what NFX is and isn\'t - which I will try to clear up, as we feel that your above statement is missing some of the major points of the NFX architecture.

Keeping in mind that the Windows OS is not a real time OS, nor was it ever intended as such and that everything in the user mode is not an actual real time DSP process in the traditional embedded DSP sense, where microsecond not milisecond scheduling is the norm;

In a user mode \'pseudo\' real time system you can EITHER achieve low latency OR a robust kernel which facilitates high system stress on the hardware without losing audio (or getting pops, etc.), but you cannot get BOTH low latency AND high system stress at the same time without experiencing problems in the audio.

The NFX architecture does overcome this fundamental limitation and let you get low latency note-on response AND everything in the list below happening AT THE SAME TIME and without sacrificing usability of the system (i.e. without any pops or loss of audio.)

a.)160 concurrent streaming audio voices from the hard disk

b.)while getting tight, low-latency note-on performance as the 160 voices are playing,

c.)while getting 160 channels of concurrent high quality interpolation (i.e. fine tuning, pitch shifting, etc.),

d.) while getting 160 channels of resonant filtering,

e.) while getting 160 channels of multi-envelope and LFO generation and useage, and with

f.) a half dozen or so zero latency DSP effects ALL CONCURRENTLY on an 800 Mhz system - like a Dell 4100 series.

None of the user mode DSP architectures that we are aware of are achieving anything even close to the level of real time performance stated above.

Simply stated, NFX lets you add DSP effects without compromising ANYTHING from the list above. It may be fair to say that the NFX, with it\'s kernel architecture, is the only true real time, host based plug-in architecture.

NFX simply provides more choices to customers who demand hardware level performance from their PC.

I hope this helps clear up the confusion over the intent of NFX. Low latency is just one aspect of several in a hard real time system.

[This message has been edited by Jim Van Buskirk (edited 05-15-2001).]

[David Abraham]

<BLOCKQUOTE><font size=\"1\" face=\"Verdana, Arial\">quote:</font><HR>Originally posted by Jim Van Buskirk:
>In a user mode \'pseudo\' real time system you can EITHER achieve low latency OR a robust kernel which facilitates high system stress on the hardware without losing audio (or getting pops, etc.), but you cannot get BOTH low latency AND high system stress at the same time without experiencing problems in the audio.

The NFX architecture does overcome this fundamental limitation and let you get low latency note-on response AND everything in the list below happening AT THE SAME TIME and without sacrificing usability of the system (i.e. without any pops or loss of audio.)
<HR></BLOCKQUOTE>

Well then why limit it to a MI? You guys know MIDI, you know audio streaming, and you know DSP. Put it all together, sign up Waves for NFX and create the dream Midi + Audio DAW.

-david abraham

[gigaDiga]

Thanks for the deeper info Jim,

Now I can see what you\'re getting at and I understand why the impulse is there, from your angle, to expand the NFX standard and the amount of 3rd party devlopers.

This said I must admit, unfortunately, that I don\'t think I\'ll be buying into it. There is one really strong reason which still pulls me towards DX..... creative workflow.

As a composer my main instrument is my sequencer. I want as few loose threds as possible and to streamline my workflow to reflect my use. I have a pulsar2 card which is full of synths... too many infact. I get clouded by the sheer amount of things to play with. So, in order to promote my use of the pulsar synths I have sampled them. In this way I only have to deal with GigaStudio & Sonar when making music, a simple and inspiring setup. If I really like a Pulsar sample which I\'ve used I can go through afterwards and patch the real synth through to the composition to give it extra life.

Further still, in order to promote creativity and to restrict the amount of time I have to keep jumping between different packages, routing this to that, and setting different values and saving different setups.... I merely use GigaStudio as a sample player/loader.

I never set volume levels or pan in GigaStudio, I never use NFX, I just load gigs and then flip to Sonar to start the creativity.
It seems obvious and inspiring to me that I should set volume, patch, pan, LFO, ADSR, Velocity, and FX all in the same package as I compose in and as part of that compositional process.
In calling for greater MIDI control over GIGs (manipulating portameto, lfo rate, sample start, legato layers, etc) many others are also desiring greater integration between GigaStudio and their sequencer. They too would like to forget about GigaStudio, and just have it purring along in the background, as they go crazy with GIGs in their sequencer.

You may say this is a repressive and close minded setup but I have found that the more technical and multi-platform I become.... the less music I make. I know you\'ll say that it\'s a really simple step to flip to GigaStudio and set my FX in there, but for me that\'s a step too far. Have you asked if Sonar or Logic might adopt NFX?

[David Abraham]

<BLOCKQUOTE><font size=\"1\" face=\"Verdana, Arial\">quote:</font><HR>Originally posted by gigaDiga:

This said I must admit, unfortunately, that I don\'t think I\'ll be buying into it. There is one really strong reason which still pulls me towards DX..... creative workflow.

As a composer my main instrument is my sequencer. I want as few loose threds as possible and to streamline my workflow to reflect my use. <HR></BLOCKQUOTE>

yes I\'m in a similar situation, however I want to a process that allows synths, lead vocal, and other live instruments to exist in the context of the same mix, with minimum technical assembly distractions.

FWIW Windows XP has a concept called GFX (Global FX) which will exist in the Kernel level and potential presents opportunities for the \"user mode\" vendors to get down into kernel mode, also they have removal of audio \"clicks and pops\" as one of their things to do bulletin items. (you can glean this info by reading Microsoft WinHec 2001 presentations)

so it\'s pretty clear to me that we\'ll get there sooner than later.

Meanwhile what I have with SONAR is working more than well enough to get real work done.

-david abraham