Mercury Engine/Supported cards... straight from the horses mouth.

Well, yeah, but they should make it fly too. $25k is a ripoff, the Tata Nano only costs $2000 in India. There's nothing magical about a ferrari, it's made of metal, the same iron and aluminum as the Tata is. And metal's free, it's found in the ground, so $25k for some metal and plastic and rubber? Do they think we're stupid? Ferraris shouldn't cost more than five hundred bucks MAX. Any more than that is a RIPOFF!!!11!!1111@!!!2

[/sarcasm off]
 
Xian Messerschmidt said:
The build I tested is now several months old and was running without the CUDA boost. From what I could tell, there isn't much, if any, performance benefit without a supported nVidia card. It seems to be more of a requirement than a recommendation.
Click to expand...

I've used a very recent build (benefits of being PPro User Group president) and I can tell you that you do see some benefits from Mercury Playback even if you don't have one of the approved cards. I was able key XDCAM HD 1080 footage (with the new Ultra keyer) in PPro and play it back in realtime without rendering. I also noticed smooth plack back when laying in animated stills with effects. I did find a point where it would start to stutter, but it was much better then what I've seen in CS4.
 
they are selling their highest available technology with no hardware compromises, it's a good thing!!!

the difference in adding more ram, is not the cost of the ram!!!!!!!.

the most important part about the price is the date it starts shipping, you would not expect something a year or two old to be the same price of something just hitting the shelf.

if you buy the same pc you bought when it was released 2 years ago, it would not cost the same!!!!

i think it's a good move to include the gtx 285 in the line up(marketing wise), but i think it would be crazy not to wait the newer cards to be approved, and just going for the best money to performance ratio possible.

if your paying that much for the software, why not upgrade your system at the same time as the software, and it was not long ago when the newer software required double the performance to run, adobe started to walk in the right path.
 
just to be striaght to the point here.

..not really sure what all the fuss is about. so what if they are basically the same physical card and only different by firmware. bottom line is thay ARE different. personally i cant see why that difference being either made of hardware or software changes makes anything change. ..since we cant do anything about it either way. ..just buy the best card you can afford and be happy with the results. at the end of the day were getting VASTLY improoved performance over what we had almost even possible with a monster CPU and GPU setup running Cs4. now you can get it with a decent CPU and GPU setup.

..either way, we are the ones winning here.

I mean, weve all seen kids get a treat and eat it up and ask for more. ..it makes sense. we always want more, or better things. its in our nature. but when and if i saw a kid get some cotton candy and eat it up and then tell off the cotton candy guy stating that

"he could have given him the 3x larger one for the same price because it only costs him pennies anyway , and for that matter its all overpriced @ super high markup from what the materials to make it costs, and its all basically just a ripoff anyway"

..all i would be thinking is

"shut up and eat. ...you just got candy and instead of enjoying it all you can do is complain?"

..sorry, but i think weve been given some pretty tasty candy here, and i dont see what difference it makes where it gets its varying levels of performance from. ..bottom line is the FX Quadro cards can greatly outperform the geforce lineup for MPE acceleration purposes, so the cost falls in line with that performance structure. ..the financial and structural logic behind it all is mute. as much as the HVX or RED images produced render mute tha fact that they use spatial offset, or a bayer aquisition method. ..in the end its all about the images. and in this case it comes down to the performance produced b the card. regardless of everything else. stressing over it makes as much sense as stressing over my lottery ticket only being one number off, and instead of the jackpot, I only got 250, 000. ..poor me. the dang lottery ripped me off.
 
neil, you're every company's perfect customer. If only every company like Canon didn't have so many people complaining about things, they wouldn't have had to bother with the whole firmware update thing. They could have just put that effort into thinking about how to make more money.
My point being is to make things fairer for more people or to see change, people have to say something about it. If everyone had always sat down and shut up throughout history, we'd probably be speaking german.
 
..nah. it would be some form of arabic or egyptian. german would only be a short glance backwards. :)

anyway . my mentality isnt that people shouldnt suggest that companies improove upon thier existing products. if you knew me you would know that innovation is my middle name. but to telling companies to shift the market structure of the products they deliver is another thing. ..just speaking from a reality based viewpoint here.

companies exist to make profit. as professionals we can either grab onto the new tech in an ever faster developing market, or we can get left behind. at the end of the day working pros arent going to bat an eye at the 1.5-3k needed to put a 4800 or 5800 in thier system and enjoy whats looking to be like one of the the best things thats happend to post facilities since being able to shoot and edit digital. a single project will easily more than cover that cost and the time savings from that one project alone are likely going to be worth it. so my pont is that if the industry isnt blinking at the tag on this awsome new opportunity, then the company selling it isnt going to either. ..so while wed all appreciate cheaper gear in ALL areas of what we do. it doesnt make financial sense for anyone to do so unless someone like RED comes along and forces thier hand. but in this case REDs card costs 4.7k and as far as i know it only helps out all things red, vs the 4800 would do EVERTHING @ only 1.5k .. and the 5800 at 3k almost seems like overkill considering the demos out there are pretty much overkill already with what they are doing and they are almost all on 4800s. ...so whether its firmware or not between a 500 geforce and a 1500 quadro doesnt matter. what matters is that most working pros will very happily and easily set down 1500 for pure editing bliss, and anyone left gawking at the price tag.. is unfortunately just left gawking at the price tag.

..am i simply rolling over to the man?

..no.

..im taking whats in his hand at the fair price he asked for it considering what great potential value it has for me, and turning around and using it to make a heck of alot more than what i paid for it.

..thats a win-win. and the reason why very very few people are tripping over the technical differences between what simply delivers and what knocks it out of the park. nobody wants to see the wizard behind the curtain, we just want to see his kick arse pyro show. ..and its well worth the tag.

theres a time for everything. yes, theres times for challenging comapnies to deliver more with what they are offering, and believe it or not, theres times to actually jump up and buy something.
 
If the MPE actually used the features unique to the quaddro drivers then I wouldn't have a problem. But when it's used as a marketing ploy to sell more quaddro cards via software limits, then that's just dodgy to me.
It's going backwards in the whole scheme of things where everything is becoming more available to more people.
I know this is just the beginning, companies are always going to try and cash grab at their new technologies and it will be a year or two before ppro, FCP and Vegas are accelerated with $80 gamer cards.
 
bimdas said:
I know this is just the beginning, companies are always going to try and cash grab at their new technologies and it will be a year or two before ppro, FCP and Vegas are accelerated with $80 gamer cards.
Click to expand...


..my thoughts exactly. they exist to make money, and no better place then with flagship technology. but the reason we should all see this as a positive is because flagships in the technology game are only new for a couple blinks of the eye, and then its old hat and a fraction of the price. they already have newer and presumably better cards planned to come out which may already be slated to be listed at a lower price point. either way, we can benifit from it now, or later, and early adopters always have to pay the tag that comes with it, but i cant begin to say how much more money/time would be saved over those couple years until these cards are a dime a dozen. its hard to hesitate have buyers remorse over a price tag (even if artificially inflated) when its earned you 40x the tag before it dropped down to below where it should have been to nothing. thats why i and most others will only priase the existing innovation, and are not pausing to think about the moral foundation for its current price.
 
Regarding the difference between the performance of the Quaddro and the GTX in terms of Premiere and supported streams, the deciding factor here was memory. The frames are pushed through the memory, so the additional RAM is MANDATORY to get additional streams, this is not Adobe trying to 'force' you to buy a more expensive card, they dont make any money from you buying nvidia cards! Basically you want more streams you need more memory, you want more memory you need a Quaddro, simple! As to the other differences and limitations of the GTX vs the Quaddro, best vent at NVIDIA not Adobe as it is their doing.

As to the list of supported cards, after being hammered time and time again about Premiere being buggy and unstable the tech at Adobe got the message. They set a list of cards for us to test with and it was a static list, no changes. That way they can guarantee that it will work right first time everytime. Yes newer cards will probably work if they added them to the list, and yes they will add cards to the list but when we started testing (over a year ago) they needed to make a call and they did, I applaud them for it.

As to ATI support, CUDA was the only technology actually launched at the time of coding start, so that is the technology they chose, could other technologies work, maybe, but when they started writing the software for the engine CUDA was it.

As to Mercury not giving you any real benefit without massive hardware. I suggest you have a look at some of my short video runthroughs on a 13' MacBook Pro with 4 gig of RAM and a 2 gig processor running about 4 or 5 streams of video with effects! www.tdepost.com The speed of the hard drive was the limitation of not adding more at this stage, not the Mercury Engine. Must have been a VERY old copy of Prem that someone was using (not sure how they got it but never upgraded) if they think that Mercury is not at the very least a massive benefit to RT.
 
evanbutson said:
Regarding the difference between the performance of the Quaddro and the GTX in terms of Premiere and supported streams, the deciding factor here was memory. The frames are pushed through the memory, so the additional RAM is MANDATORY to get additional streams, this is not Adobe trying to 'force' you to buy a more expensive card, they dont make any money from you buying nvidia cards! Basically you want more streams you need more memory, you want more memory you need a Quaddro, simple! .
Click to expand...

What about the lower end quadro cards with 1 and 1.5 gigs of ram? Do they only support 3 three streams? I've got a GTX 285 with 2 gigs of ram, shouldn't that be enough ram?
Adobe and Nvidia seem to me to have had some sort of partnership since cs4 but that's just what I get from their ads and info on their websites.
 
How many streams can you get WITHOUT a supported CUDA card. I am able to edit up to four streams of native Panasonic HMC-150 AVCHD (1080p 24p) in EDIUS on my i7 920 with 12gb RAM.
 
DJDecay said:
The cards are functionally identical, they use the same GPU chip, pin-compatible, however here are the differences.

SLI supoort (tm)
PhysX (tm)
Amount of GDDR RAM (Flex 1GB vs. 4GB) etc..

First cards target market are Game Players, second are professional workstations.
But in the computer market the car analogy does not apply.

Yes, the GeForce cards are crippled in their firmware, on purpose such that users wanting more, would have to pay. That is it. I don't care if FX 5800 comes in a gold plated case, the processor GPU/RAM/BUS/Bandwith and clockspeed are identical to GTX 285, maybe from a cleaner batch, maybe more "well tested" etc... The FX also tends to be a tad behind on the PCIe spec support, and driver releases/updates for all OS'es.

This is the same level of panache SGI had for pro users before the filed Subchapter 11, unfortunately NVIDIA is providing both the cheap and the expensive.

And their only competitor on the chip and reference board markets, sold to AMD.

So you're paying extra for some firmware, and some extra ram and a bigger cooler to chill it.
Supply/Demand dictates the price, if the GTX 285 could be functionally equivalent, then they'd need to be doing something completely "outstanding" on the Pro 3D card market.

But until PCIe 3.0 is widely adopted not much can be done.

Or here is what the NVidia's PR department would say about it.

In both cards the multi-core GPU (identical in both) has a 512bit data path
which is achieved by splitting the DAC path into 8 - 64 bit lanes (or 2 - 32bit wide words), and attaching a GDDR3 RAM chip to each one,

The original prototype board GTX 285 was designed for cheaper GDDR3 chips @ 128MB each. When the 256MB chips became cost-effective for the gamers market,
vendors like PNY, MSI, Sapphire etc.. adopted those from their reference design to create GTX 285 2048MB GDDR.

While the FX 5800 was already designed with 512MB GDDR3 chips in mind. (no one expected GDDR3 ram to fall in price so quickly) and NVDIA and their partners did not expect to have gamers pay $1k for such a card so 512MB arrangement had less fabrication.

While a vendor ASUS spent it's time changing the RAM-DAC around and using a dual board design of the GTX 285 to create a 4GB version of the card, but it's setup is "unknown/unsupported" from the initial reference design - therefore not recognized as a FX 5800.

So yes, it evades all logic, but it is absolutely true... firmware is what limits the card from "theoretically" supporting the FX 5800's features, and Z-Buffer/Frame Buffer RAM limit/per-chip-data limit is what prevents it from supporting the additional CUDA features practically - so for safety it's disabled in firmware. As NVIDIA does not envision that vendors will succeed in making a 4GB GTX 285 straight through - before it's obsolescence and replacement by an even faster GPU/RAM combo. I wish they would just start shipping video cards with GDDR RAM sockets on them.

GPU development is expensive business, so is GPU fabrication, so it actually makes sense for NVIDIA to make as many identical chips as possible, and then just change the firmware and add-ons around them. completely un-like the Intel desktop/server/mobile model which uses grossly similar in design but somewhat different chips.
Click to expand...

If it's true, it's true, but as you say, it makes no sense. Which is why it's difficult to believe.
 
Xian Messerschmidt said:
How many streams can you get WITHOUT a supported CUDA card. I am able to edit up to four streams of native Panasonic HMC-150 AVCHD (1080p 24p) in EDIUS on my i7 920 with 12gb RAM.
Click to expand...

Can you edit those streams in RT when color corrections and grades are added? That and the, not having to transcode my 7d footage is what I'm most looking forward to. Hopefully by the time they release, Adobe would have added more MPE features to the GTX cards.
 
bimdas said:
Can you edit those streams in RT when color corrections and grades are added? That and the, not having to transcode my 7d footage is what I'm most looking forward to. Hopefully by the time they release, Adobe would have added more MPE features to the GTX cards.
Click to expand...

So long as I'm using the EDIUS native filters, yes I can.
 
You must log in or register to reply here.
Back
Top