Scratch on the new Macs?

[Lucas Wilson]

I don't understand why this driver issue is such a big deal.

Think of the CPU as the car, and the graphics driver is the gas the car runs on. Now with that analogy in mind... have you ever seen what happens when you try to use Unleaded in a Diesel engine? ; )

They obviously make the cards... don't they want to sell more of them?

Sure. But it's a numbers game. The amount of cards they would sell if this was resolved is completely tiny when put up against NVidia's business goals as a whole. We're talking about *maybe* a couple of hundred cards over a year or two. NVidia is a multi-billion dollar company.

Lucas, if this driver issue was solved. How long would it take for assimilate to release an OSX version? Could it be done in less than a year?

Nothing that time and money can't solve!

Lucas

 

[Andrae Palmer]

Isn't that a bit like asking "if there were no foreclosures, would full economic recovery happen in less than 6 months?" It's a bit silly to use such conjecture, because there still are plenty of foreclosures, and the driver issue is not solved and shows no signs of being solved. That's what Lucas just tried to tell you.

I don't think it's silly at all to wonder if Assimilate would be committed to the Mac/OSX either in bootcamp or with a native app. Imagine if Lucas stated that an OSX code rewrite would be impossible at this time even if the driver issue was solved... then I would have a clue that Scratch will never natively exist in OSX. On the other hand he could state publicly that Assimilate has been working on a native version of Scratch for OSX. That is why I'm here to ask questions and to have a feeling of what is the mindset of the company so I can make plans accordingly regarding hardware. The original poster is basically trying to gauge this too and that's why this thread was created. Lucas response about the driver issue was enlightening. I still don't understand what he means by the drivers are not "pure" but at least now I know there is a driver issue in bootcamp. I thought that in bootcamp native drivers could be applied to the cards.

As I stated NVIDIA does sell these high end SDI cards... why would they not want to sell more to Mac users by building the darn drivers/ROM? It's like going halfway in a business... they already make the cards that only a few people in the world use... why not make the drivers so more people could use the cards? You are making it seem like it's so difficult for this to happen. I'm personally hoping that Apple/NVIDIA finds a way for this to happen and if they do... Assimilate jumps on board with a quickness and does not drag their feet.

Perhaps ATI will jump into this game or even Apple itself... I really cannot see Apple being competitive without a proper high end graphics hardware option. As far as your arguments about Mac not being a generic PC... both NVIDIA and ATI already makes cards and drivers for Macs. I understand the EFI/BIOS issues (hence why bootcamp exsit) but there is a Mac ROM work around that is currently being applied to the consumer ATI/NVIDIA cards. If your argument is that this is such a small segment of the industry and it doesn't benefit NVIDIA to make the Mac ROM. Then I will say that $7000 SDI cards are a very small segment of the industry and if they want to make it a larger segment they can make it Mac compatible.

In any case I now have the information that I was seeking and have no further questions regarding this issue.

 

[Shane Betts]

So, until either Apple supports that card, or there is a competitive solution to NVidia that supports a GPU -> SDI fast pathway without having to go back through the host bus, the chances of getting SCRATCH on OSX are unfortunately zero.

Lucas

Hi Luki.

The graphics cards in the new Macs all come with (Apple proprietary) Display Port plugs. These ports come very very close in spec to 3Gig HDSDI (2.75Gb vs 2.9Gb I seem to recall) so the new Macs should, in theory (and with an external adaptor that doesn't yet exist but somebody has to be working on one - shouldn't you Matt:innocent: ), be able to give you GPU backed 3Gb HDSDI.

What would you do then huh? Huh? Huh?:biggrin:

 

[Shane Betts]

Bettsy, when I priced this out it was just a little bit less than buying the finished kit with no assurance that it would work. What is your recipe? And did you do it?

Hi Rob

No, I haven't done it in this particular instance. I have in the past and know a lot of guys who do and it can be done if you pay attention to every little spec. But, yes, the savings over Boxx aren't huge and, like I said, compared to a seat of the software...

My company is all Mac and we've been looking at the whole Scratch-on-a-Mac question and I've concluded it's not worth it. I've decided it's either buy Speedgrade XR or a Windows box. In fact, for the exact same reason Luki gives, I think I'd buy a Windows box for Speedgrade as well. (Don't it break your heart)

However, see my previous post. There may be some hope in Mac land yet of displaying via SDI without the PCI bus having to triple-handle the data (although in theory there should be heaps of bandwidth). Hassle your local video adaptor manufacturer! I know I will be.

 

[Rob Gardner]

Thanks for the response, Shane,
Rob Gardner

 

[Lucas Wilson]

Hi Luki.

The graphics cards in the new Macs all come with (Apple proprietary) Display Port plugs. These ports come very very close in spec to 3Gig HDSDI (2.75Gb vs 2.9Gb I seem to recall) so the new Macs should, in theory (and with an external adaptor that doesn't yet exist but somebody has to be working on one - shouldn't you Matt:innocent: ), be able to give you GPU backed 3Gb HDSDI.

What would you do then huh? Huh? Huh?:biggrin:

Hey Shane, ; )

We actually already support the Display Port on the NVidia CX cards, and we support feeding true 10-bit out that port. To my knowledge, we are the only manufacturer that does this.

But, the only monitor that supports Display Port at 10-bit is the HP Dreamcolor. And when in 10-bit mode, things are a bit unstable.

So, you tell me:

1) We support HD-SDI on Windows, with the vast infrastructure that exists across multiple manufacturers supporting HD-SDI. We also support the Display Port output on the NVidia cards on Windows in anticipation of Display Port becoming more popular.

or...

2) We support an upcoming Display Port on the Mac when we already support it Windows just so we can... wave our hands around and say we're on OSX?



Lucas

 

[Lucas Wilson]

I'm personally hoping that Apple/NVIDIA finds a way for this to happen and if they do... Assimilate jumps on board with a quickness and does not drag their feet.

Andrae,

If you look at my threads in the past, I have pretty much always said that if there were supported NVidia SDI drivers for OSX, at that point, porting would be an option for us. Does not mean we would do it, but it would be a serious option to consider.

Perhaps ATI will jump into this game or even Apple itself... I really cannot see Apple being competitive without a proper high end graphics hardware option. As far as your arguments about Mac not being a generic PC... both NVIDIA and ATI already makes cards and drivers for Macs. I understand the EFI/BIOS issues (hence why bootcamp exsit) but there is a Mac ROM work around that is currently being applied to the consumer ATI/NVIDIA cards. If your argument is that this is such a small segment of the industry and it doesn't benefit NVIDIA to make the Mac ROM. Then I will say that $7000 SDI cards are a very small segment of the industry and if they want to make it a larger segment they can make it Mac compatible.

Apple already does have a proper SDI option for their customers - AJA. I mean really, if I worked for Apple, I would have zero business motivation to support the NVidia SDI cards. I hope I'm wrong. I hope they do it. But being neutral about it, I do not see any way that it makes business sense for Apple to spend valuable driver team man-hours to support the NVidia SDI cards.

ATI has had the opportunity to make SDI cards for a long time now. We've had discussions with them and I know other companies have as well. No interest there.

Lucas
-----
ASSIMILATE, inc.
LA, CA, USA

 

[BradWright]

If you go by the numbers, the growth opportunities are on the Mac. The PC market is contracting right now, while the MacIntosh platform is expanding. Now, this may not have an impact Assimilate for a while. However, it leave's the door open for Apple and other third party solutions to arrive on the platform. The amount of users on Final Pro dwarfs all other editing platforms. If the Scarlet camera becomes a huge hit, the Red platform could become very mainstream.

The combination of these all these trends could point to a sea change in the video industry in the next year. I think it would be a bad idea for Assimilate not go after the Mac market. However, if they don't, the Apple developer community is going to step up.

 

[Adam Glick]

If you go by the numbers, the growth opportunities are on the Mac. The PC market is contracting right now, while the MacIntosh platform is expanding.

*this is a little bit off-topic but...

The statement above is inaccurate. Everything I have been reading shows that Apple's Mac sales are actually declining more quickly than those of its PC cousins.

http://www.schaeffersresearch.com/c...ontrarian.aspx?single=true&newsid=91789#91789

http://www.macnn.com/articles/09/03/10/mac.sales.could.be.down/

http://www.fool.com/investing/general/2009/03/12/is-apple-doomed.aspx

http://news.cnet.com/8301-13579_3-10166252-37.html?tag=mncol;txt

http://www.dailytech.com/Apple+Sees...+Slips+to+Fourth+Place+in+US/article13951.htm


Of course, I am biased, but If I were a VP of product development @ Apple, the decision NOT to devote the resources to developing an SDI driver would be an easy one. There's likely no profit in it. Unfortunately, that's gotta be the bottom line from a "is it smart business?" perspective...

 

[BradWright]

*this is a little bit off-topic but...

Of course, I am biased, but If I were a VP of product development @ Apple, the decision NOT to devote the resources to developing an SDI driver would be an easy one. There's likely no profit in it. Unfortunately, that's gotta be the bottom line from a "is it smart business?" perspective...

I'm going by the numbers that were reported for last quarter. This quarter isn't over yet, so there's no hard data. Putting speculation aside, here's what was reported:

http://www.appleinsider.com/articles/09/01/23/apples_mac_growth_a_standout_in_ailing_pc_market.html

 

[Antoine Baumann]

But, the only monitor that supports Display Port at 10-bit is the HP Dreamcolor. And when in 10-bit mode, things are a bit unstable.

Hi Lucas, can you elaborate on that subject.

Seems there isn't a nvidia quadro option on the new mac, so I guess it kind of rule out the sdi options anyway?

Thanks,
antoine.

 

[Andrae Palmer]

Apple already does have a proper SDI option for their customers - AJA.

I was just day dreaming of AJA coming to the rescue... what the heck are they waiting on? AJA+RED... surely this company must have a vision of what is going on. I think it's about to happen. I feel it. The next 2 months should be interesting.

 

[Shane Betts]

So, you tell me:

1) We support HD-SDI on Windows, with the vast infrastructure that exists across multiple manufacturers supporting HD-SDI. We also support the Display Port output on the NVidia cards on Windows in anticipation of Display Port becoming more popular.

or...

2) We support an upcoming Display Port on the Mac when we already support it Windows just so we can... wave our hands around and say we're on OSX?



Lucas

Well, that pretty much sums your attitude to the Mac right there.

Wave your hands around??! You keep saying that you don't port to Mac because you can't get a GPU-SDI path and when one is suggested you come back with that.

Not impressed sorry.

 

[Shane Betts]

I was just day dreaming of AJA coming to the rescue... what the heck are they waiting on? AJA+RED... surely this company must have a vision of what is going on. I think it's about to happen. I feel it. The next 2 months should be interesting.

I agree that AJA are in the box seat. The problem companies like AJA and Blackmagic Design face here is that, with the GPU processing the debayer, the stream has to go to the PCI bus, out to the GPU (at 36MB/s), debayered and sent (now at uncompressed 2K at around 300MB/s) back through the PCI bus to the AJA/BMD display adaptor and out to the display.

Now add all this up and you get less than 700MB/s which, for PCI Express should be a walk in the park but the questions are how many graphics cards support two way data traffic and who would write the code to make it possible.

I believe the greatest possibility here lies in that new Mac Display Port. That port has very similar specs as 3Gb HD-SDI that AJA and BMD are currently using to display 2k. Someone will surely come up with a convertor box to sex-change Display Port to HD-SDI and then we're away.

However, given Luki's response to the question, I don't think his company views porting to Mac with any intent so perhaps one of the other apps will take advantage of it - once the Red/Assimilate strangle hold on debayering expires or is broken.

 

[Lucas Wilson]

Well, that pretty much sums your attitude to the Mac right there.

Wave your hands around??! You keep saying that you don't port to Mac because you can't get a GPU-SDI path and when one is suggested you come back with that.

Not impressed sorry.

Shane, your exact words were:

"...so the new Macs should, in theory (and with an external adaptor that doesn't yet exist..."

Key phrase there being "doesn't exist yet." So by your own words, a GPU-SDI path on OSX does not exist and you're suggesting I build a market around it? The reason that nobody has come up with a Display Port to SDI adapter yet, by the way, is that the market is not demanding it.

I agree that DisplayPort is a cool technology. But currently, the DreamColor is the only monitor that supports it natively. The drivers that run things at 10-bit are unstable at best, and the Dreamcolor is still in very early days of testing for color-critical workflow.

I said below that if a good GPU->SDI path would exist on OSX, then we would seriously look at a port. But that does not look like it's going to happen anytime soon.

There is a reason that there are no serious high-end, high-performance color applications running on OSX. Quantel, DaVinci, Filmlight, Iridas, ASSIMILATE, Autodesk, Digital Vision. None of these companies have their high-end finishing tools on OSX. Iridas and Filmlight both have OSX ports of their products... but not for supervised color, because SDI is not an option.

But other than that, *nobody* has ported to OSX. That's because Apple controls their drivers, and could give a crap about working closely with other manufacturers on driver tweaks, bug fixes, etc. The companies who need serious realtime performance at high-resolution are locked out of OSX because there is just no way to get at the drivers at a serious enough level to do optimization and performance testing. And believe me, this is not for lack of trying. As an example, we did some OSX testing, and found a fairly serious driver bug that killed performance. Isolated it, tested it, proved that it was a bug, and generated test cases for duplication. We fed it into the proper development channels, and spent *months* trying to get a response out of Apple. Since the dev channel didn't seem to work, we tried other avenues. We got *nothing.* Zero, zip, nada.

With NVidia, when we find serious bugs, they are typically fixed in the next driver roll-out, and we have great technical contacts with NVidia who are super-reponsive and care about their partners.

So before you go around saying, "gee... your attitude sucks... the Mac rules... let's wait until the stranglehold ends..." you should take a cold and neutral look at your beloved platform and then you might understand why it is that none of the high-performance tools you really want are on OSX.

I'm not impressed either.

Lucas

 

[Lucas Wilson]

Hi Lucas, can you elaborate on that subject.

Hi Antoine,

Pretty simple...

As a connection, DisplayPort supports 10-bit. (so does DVI, but there aren't any monitors that accept a 10-bit DVI signal) The Dreamcolor supports 10-bit input via the DisplayPort connector.

We did some work, and support the 10-bit output on NVidia cards with the DisplayPort connector. So, if you have SCRATCH, an NVidia card with a DisplayPort output, and a Dreamcolor display, you can have a 10-bit signal path without SDI.

I hear lots of wildly differing opinions about the Dreamcolor, though. Still waiting to see how all that shakes out.

Lucas

 

[Shane Betts]

Shane, your exact words were:

"...so the new Macs should, in theory (and with an external adaptor that doesn't yet exist..."

Key phrase there being "doesn't exist yet."

Well actually, I finished by saying "What would you do then?" In other words, and sorry if I did not make myself clear, what would you do if somebody did come up with a DP-HDSDI adaptor. That was my question - (that's why the question mark came after that sentence:whistling: )

We (clip) spent *months* trying to get a response out of Apple. Since the dev channel didn't seem to work, we tried other avenues. We got *nothing.* Zero, zip, nada.

Luki, I have worked in Mac development and I know what a pain in the arse they are. Believe me - let's talk "gamma shift" for a second. And I know Microsoft give way better support to their developers. Still, a lot of folk do successfully develop for Mac and there's a huge number of folk out there cutting on FCP so all I'm saying is, "There's a market and if the resource (SDI) became available, what would you do?"

you should take a cold and neutral look at your beloved platform

If you read back earlier in the thread, I'm the one who began by saying "why would you want to?" to the question of running Scratch under Mac.

Yeah, I use a Mac but, as I don't have to actually cut on the things any more, I'm totally agnostic - somebody else's problem. If it works better I'll install anything. My problem, like a whole lot of people in this business is that we cut on FCP. Now, if I was still cutting I'd want Avid myself but the editors we see coming through are, in the vast majority, FCP guys - so that's where we go. That locks us to Mac and it would save a whole lot of aggro if we could sew the whole shebang together on the same platform as our NLE and I'm sure that's where the original question comes from.

and then you might understand why it is that none of the high-performance tools you really want are on OSX.

Not sure if you mean high performance there or high priced. The two terms are not necessarily linked. We are, after all, on the Red forum and I'm sure Sony, Panavision et al use similar language when discussing Red.

There are different business models at play here and some very powerful, low cost tools are available now on Mac, Windows, Linux etc. because the developers involved realise they can actually spend more on R&D and make more money by selling a large number of low priced units than they ever could selling a small number of high priced ones. The products you list all fall into the latter category and I know the rationale - "How many colour graders can you sell?" Avid were asking themselves the same question when Apple released FCP, just as the camera manufacturers were when Red began. Well, the answer just might be, "A whole lot more than we currently do". Just sayin'.

 

[Antoine Baumann]

The drivers that run things at 10-bit are unstable at best
Lucas

Lucas, you are talking about nvidia driver?


thanks,
antoine.

 

[Barend Onneweer]

Oh man, this thread is very quickly descending into a platform argument.

From a user pov - who cares what platform Scratch runs on? The hardware is only a fraction of the cost of a complete system - and Scratch runs full-screen and has it's own dialog boxes so you don't even see Windows while working.

Bettsy, the argument of staying on a single platform for editing and finishing is not a serious one, is it? Getting your FCP edit into Scratch is no problem regardless of the platform. So putting in all the effort to port to OSX just because it makes you feel more comfortable is hardly a business case.

When you suggest the new business models around low-cost color grading software - are you talking about Color? Because you know as well as I do that Apple isn't really selling Color. Apple is selling boxes and uses Color as a free incentive. And I don't see Apple allocating huge amounts of R&D to Color, or Shake for that matter.

So why aren't you using Color? It seems pretty close to what you want. Low-cost, runs on a Mac, uses AJA SDI out?

Bar3nd

 

[Shane Betts]

Oh man, this thread is very quickly descending into a platform argument.

From a user pov - who cares what platform Scratch runs on? The hardware is only a fraction of the cost of a complete system - and Scratch runs full-screen and has it's own dialog boxes so you don't even see Windows while working.

Hi Barend

I totally agree but it's not coming from me! Go to page one of this thread and read my posts. The very first thing I posted was exactly what you've just said.

I was simply floating the theoretical question with Luki that, should an SDI solution come into being on the Mac, would they port Scratch. The answer is obviously no and I do understand - although seems to me Luki has been using the no SDI argument for a while and clearly it goes deeper than that. But, as I've said, I've worked in Mac development and they are an arrogant and frustrating organisation to deal with. But, you know, you buy your ticket you takes your ride.

Bettsy, the argument of staying on a single platform for editing and finishing is not a serious one, is it? Getting your FCP edit into Scratch is no problem regardless of the platform.

Absolutely right. I didn't start this thread (although I feel as though I've unwittingly hijacked it) and, as I said in my first post, why would anyone bother installing Scratch on a Mac? You want Scratch? Buy a PC. Easy. Even in the case of, say, Speedgrade, I'd still buy a PC due to the SDI option.

So putting in all the effort to port to OSX just because it makes you feel more comfortable is hardly a business case.

Well, that's confusing matters a little. No, it doesn't make a business case for me but I'm not a software developer, I'm an end user. I am suggesting, however, that it might make a business case for a software developer to develop for those out there who do feel more comfortable on their favourite platform, whatever platform that may be. I would never contemplate buying a Wii but sales seem to be doing just fine without me...

I'm just suggesting that the reason this question is asked again and again is that there is a market out there of small facilities or one man operations running FCP who would probably buy a seat if it worked on their edit machine. If they sold a seat to one in ten FCP users, how many would that add up to be?

When you suggest the new business models around low-cost color grading software - are you talking about Color?

No, I'm not talking about Color. I'm actually suggesting that the colour grading application business might just be due for the sort of low cost revolutions that have washed through other sections of the industry - just as Red are currently doing to cinematography. I was just a little amazed at Luki's "wave our arms around" quote and that got me onto my old high volume low cost soapbox and for this I apologize.

So it's not about platforms. I'm just suggesting that if a Scratch or a Speedgrade were suddenly available, platform agnostic, for a couple of grand from, say, an Adobe, they'd probably make more money than they do now staying "high end".

Once again, I didn't start the thread and I was the first one to post to say, "Just buy or build a PC". Just saying that, "Give us $20k and buy a new machine" won't move as many units as "Give us $2k and install it on whatever you have" will.