It is true in the sense that this distributed processing is the basis for 3.x and the new hardware. At least this is how it's been explained to me and is in all the demos I've seen. I also don't get as many RT DNxHD streams without the Nitris DX as I do with it.

Now, as to how much more it's worth, I can't speak for you  but my sessions are much faster and less trouble than they've ever been. My clients are blown away at how fast their sessions are now (Especially changes they can see in real time). I haven't run the numbers yet but I do feel a greater level of effeiciency than in my FCP bays. That might be subjective though.

Personally, I feel that my HD sessions are smoother and less troublesome since the DX hardware. My shop can pay for a DX box pretty quickly so I might be biased but, for me, yes it was worth the extra cash. However, in general principle, I would like Avid to start taking a hard look at what they charge for this stuff. If nothing else than for just market perception. How about a little more realistic bottom line....?

Isn't the Nitris and Mojo DX 'Old Avid thinking'?

The latest Avid DS 10 uses an AJA card and high end Nvidia graphics card that is the way to go for performance and price.

Maybe the Nitris box costs so much because that is what it costs Avid to make as they do not benefit from mass production like AJA and Blackmagic smart management will realise that it is an issue they need to address to stay competitive.

Well, in a sense, you are absolutely right. The DX box development was a product of the Old regime at Avid. Who knows what we'll see from Mr. Greenfield.

 Going back to the car analogy.... 

I drive a Nissan Sentra - good gas mileage, power windows etc all around good for getting me where I need to go.

A colleague of mine drives a BMW 325i - good gas mileage, power windows etc and all around good for getting him where he needs to go. 

What justifies the extra cost for the BMW?  Don't ask me, I drive a Sentra.

In other words, you can not make comparisons based on "included" features or functionality as it's all subject to what each of us value based on our own needs.

Kenton.VanNatten:

 Going back to the car analogy.... 

I drive a Nissan Sentra - good gas mileage, power windows etc all around good for getting me where I need to go.

A colleague of mine drives a BMW 325i - good gas mileage, power windows etc and all around good for getting him where he needs to go. 

What justifies the extra cost for the BMW?  Don't ask me, I drive a Sentra.

In other words, you can not make comparisons based on "included" features or functionality as it's all subject to what each of us value based on our own needs.

Forgive me, as I think I went about the car analogy all wrong (which I deserve for even trying to do a stupid car analogy).  I guess I should not in essence be saying, is Avid with DX worth an extra $10,000 over FCP with AJA....I'm not extremely familiar with the experience of working with FCP, so maybe it is.  What I should be asking is, is the Avid with DX worth an extra $13,000 (or whatever it is) over just Avid?  Does it give you $13,000 worth of extra performance....I mean really?  The bad part about the situation is that even if you don't think it's worth the extra performance, you HAVE to pay for it if you want any type of I/O (or their way overpriced Mojo line).

BLKDOG:

That's not true. The new DX boxes share FX processing with the GPU and CPU. It's the power of the distributed processing that allows for the increased RT performance.

BD,

That's not what Greg Staten or Frank Capria have told me. In my own testing, I actually got MORE streams of real-time when the hardware was disconnected. All effects processing is done in software. It is multi-threaded for better CPU/GPU performance. Disributed processing by custom hardware went away with classic Nitris and is probably no longer a feature of most companies' configurations.

A single stream of combined data is sent to the DX hardware. Since that's on a PCIe data lane, the bandwidth can be greater than Adrenaline's FW400 conduit. The information going to DX is either full raster or thin raster and in the case of thin raster, the DX provides scaling back out to full width. In the case of Nitris DX, that info can be DNxHD and Nitris decompresses it through embedded chips. The advantage of the thin raster toggle, is that all computation is done in thin raster in software. You get more streams with thin raster, but only if it's all thin raster, not mixed.

Distributed processing would imply that video and effects data is going back and forth between the CPU/GPU and the DX hardware via the PCIe bus and that is definitely not happening.

- Oliver

BLKDOG:

 Y'know, that was a big one for us. Losing all of your projects just because you upgrade the software really blows. 

BD,

I'm not trying to be a pain here with my posts, but I don't know what you mean. I presume you mean FCP with this comment. FCP projects, like Avid and AE, are generally always forward compatible but not necessarily backwards compatible. I have so far migrated various projects from FCP 3 all the way up to current (FCP 6.0.4). Along the way, there have been some compatibility issues with QT versions. That's troubling, but Avid has done that, as well, at various times in years past.

On the other hand, right now I have a project done on an OS9 Xpress Meridien that can't be opened on any current Avid software, unless I open individual bins from a new project. I have a client who has quarantined MXF media on their Adrenaline system that cannot be accessed any longer and had to be re-ingested.

So all systems have these issues.

- Oliver

OliverPeters:

FCP projects, like Avid and AE, are generally always forward compatible but not necessarily backwards compatible. I have so far migrated various projects from FCP 3 all the way up to current (FCP 6.0.4)

I think what he was referring to was a scenario such as this:

I had FCP 5.1.4, a client I was collaborating with upgraded to FCP 6 mid-project.  I could no longer open his project modifications on my system.  I had to upgrade a perfectly working system just to continue on the project.  Then he upgraded to 6.0.2 and couldn't open my work I had done in FCP 6.  Compatibility is a nightmare with FCP. 

With Avid, I have old bins that I've been able to open from at least XDV 3.5

"... I have a client who has quarantined MXF media on their Adrenaline system that cannot be accessed any longer and had to be re-ingested. ..."

Did anybody try moving the Quarantined files into a separate folder to scan & index?  We frequently get Quarantined files, only because the file count is nearing the MXF limit of 5,000/folder.  With fewer files/folder, the scan & index completes normally.

SandmanX82:

Yes, but Avid offers that level of true, real-time effects through use of the software, not from it's boxes.

Even if that is the case, they can only provide that sort of performance because they control all aspects of the I/O. Creating software that can deliver RT performance of a very specific piece of hardware, with very specific firmware is very different to creating software that can deliver that performance with some sort of API-level I/O interface relying on third-party hardware and drivers. This is clearly evidenced by the fact that FCP can't offer anything like that performace with any of their hardware options.

I've edited HD on a FCP 6, 8-core MacPro with BMD HD Extreme, and with Avid systems (even the much maligned Adrenaline on a 3-year old XW8200) and the performance was night and day. FCP costs less, has a lot of great functionality and presents plenty of options, but it can't match the performance in my opinion, and I've not yet seen anyone suggest otherwise. If it were my own business and I were making a purchasing decision, I'd choose the system that I thought would allow me to work faster, better and more efficiently and right now that'd be MC3 on an HP machine with Mojo or Nitris DX.

OliverPeters:

On the other hand, right now I have a project done on an OS9 Xpress Meridien that can't be opened on any current Avid software, unless I open individual bins from a new project.

That can be the case, create a new project with the same name, copy all the bins from old project into new project and it's back. All the important data is contained within the bin files anyway.

The MXF media should definitely work. There can be weird permission errors sometimes, but the media itself should still be fine.

Randall L Rike:

 Did anybody try moving the Quarantined files into a separate folder to scan & index?  We frequently get Quarantined files, only because the file count is nearing the MXF limit of 5,000/folder.  With fewer files/folder, the scan & index completes normally. 

Yes, this was tried. Unfortunately this turned out to be a huge comedy of errors on the part of the client. This wasn't the only issue. I don't want to embarrass anyone, so I won't go any further into it. But thanks.

- Oliver

Sycophant:

 That can be the case, create a new project with the same name, copy all the bins from old project into new project and it's back. 

Dylan,

Yes, I understand how to do this. I was merely pointing out that things aren't always perfect in the Avid world either. Thanks.

- Oliver

Kenton.VanNatten:

 I had FCP 5.1.4, a client I was collaborating with upgraded to FCP 6 mid-project.  I could no longer open his project modifications on my system.  I had to upgrade a perfectly working system just to continue on the project.  Then he upgraded to 6.0.2 and couldn't open my work I had done in FCP 6.  Compatibility is a nightmare with FCP.  

Kenton,

Correct, you were trying to open a project that had been migrated forward in an older version. As I said, backwards compatibility is unreliable and often impossible. You can often do this with XML. I haven't tried it, but has anyone taken an MC 3.0 project and opened it on an old MC 7.2? That's an extreme example, but it's a version of what you are asking to do. He should have been able to open your 6.0 project in 6.0.2, though. Obviously, both folks want to be in sync when collaborating on the same project.

I have generally had very good luck with these migrations, but yes, there do seem to be a few things under the hood that cause problems when you bounce around with too many difference systems and different versions. Obviously not perfect, but for someone to say you lose all your old projects is hyperbole.

- Oliver

Sycophant:

Creating software that can deliver RT performance of a very specific piece of hardware, with very specific firmware is very different to creating software that can deliver that performance with some sort of API-level I/O interface relying on third-party hardware and drivers. This is clearly evidenced by the fact that FCP can't offer anything like that performace with any of their hardware options.

Dylan,

Although I basically agree with you, the real test will be with DS V10. Since it's basically using an AJA XENA card (same as KONA), that contradicts the API argument. IMHO FCP's performance limitation has nothing at all to do with hardware. It's because FCP has a lot of very old code in it. As goofy as this might sound, play with some ProRes 1920x1080 files in iMovie '08 on an iMac and compare this to the same media in FCP. It is the most fluid interaction of any media that I've seen on any NLE. This is possible because of the new code and optimization possible because iMovie is a rewrite. That's also the reason MC3.0 is so good. Mainly, key parts of the code has been rewritten.

- Oliver