[Andrew Richards]

[Dave LaRonde] “On the other hand, the questions could be like this imaginary example:
How much do you love the magnetic timeline?
A) Can’t live without it
B) Really think it’s great
C) It’s pretty nice
D) It’s okay”

Colbert-style survey:

Final Cut Pro X: great NLE or greatest NLE?

Best,
Andy

[Andrew Richards]

[Andy Field] “Had Apple only asked the folks on this forum before FCP X – they’d still be the leading, gushed over NLE and this forum probably wouldn’t exist”

You’ve created a paradox!

Best,
Andy

[Andrew Richards]

[Lance Bachelder] “Think you may want to go back and re-read the Anandtech review – they say the compute test they did does NOT work with dual GPU cards.”

The first one, yes, but the one immediately after that test does optimize for multiple GPUs. And Tahiti dual blew the Kepler dual out of the water.

[Lance Bachelder] “I’m thinking that nVidia wouldn’t spend 4 years developing Kepler as a upgrade to Fermi if they didn’t think it was a good thing? “

There must also be a reason there are no Fermi Quardos or Teslas (yet). Kepler (so far) is lousy on GPGPU tasks compared to its predecessor and its peers from AMD. AnandTech closed with a reference to a forthcoming Big Kepler, which would presumably be a different story and be the future of the Quadro and Tesla products.

[Lance Bachelder] “There is no evidence that an app like Premiere that can take advantage of CUDA technology will be better or worse using Kepler vs. Fermi – we won’t know until CS6 is released, Adobe certifies the new cards and folks actually put them to the test in real-world situations.”

NVIDIA is pushing their Fermi-powered Quadro and Tesla products for CS6. Does Adobe rank the cards’ performance when they certify them?

Best,
Andy

[Andrew Richards]

[David Cherniack] “We may be talking about two different things, Andrew. I’m talking about the difficulties in authoring, not in learning/using. As such I’ll stand by my claim. NLE’s must do a wide variety of tasks for a wide variety of uses and users. They must interface with the analogue world of a wide spectrum of hardware. I can’t imagine that any 3D software, as complex as they are to learn and use, is as complex to write”

I’m certainly not qualified to adjudicate that either.

[David Cherniack] “But point out a ‘major’ app (other than OSX) that Apple developed from the ground up. “

There is Xcode, which is arguably the most important app that Apple develops (at least from Apple’s perspective). There is also Safari, or more importantly the Webkit engine it runs on. I intentionally left out Motion and Aperture, as they might not qualify as “major”.

[David Cherniack] “It seems Adobe is fully committed to competing at the high end. I think we’ll see some good uptake this year and when their media management rivals Avid’s they’ll be well positioned.”

I agree- they seem to be at that momentum tipping point FCP was at around FCP5. I doubt they will unseat Avid as the king of studio features and episodic TV, but it will fill most of the vacuum left by FCP7.

Best,
Andy

[Andrew Richards]

[David Cherniack] “NLEs are the most complex applications in computerdom.”

I taught myself FCP, but stuff like AutoCAD and Maya is waaaaay more complex. Full-featured NLEs aren’t trivial, but they aren’t the most complex software out there by a long shot.

[David Cherniack] “What great applications have they ever developed in house from scratch? Even the non-application OSX is derived from elsewhere. “

I guess it depends on your definition of “great”, but for OS X’s part, the engineers that built it at NeXT went to work at Apple and turned it into OS X, so it isn’t like Apple didn’t have or doesn’t now have top-teir engineering talent. Autodesk didn’t develop Smoke, they bought Discreet. After Effects originated outside Adobe. Avid acquired DS. What’s the difference?

I agree Apple’s priorities are clearly different than those of the other Big A’s, but let’s keep things in perspective. I’m assuming by “high-end” you are referring to studio feature motion pictures and broadcast episodic TV. FCP was dominant below that, and arguably still punching above its weight class when it was used to win the last two Oscars for editing.

FCP had only recently broken into the high-end by the time it was scuttled, Premiere Pro has yet to achieve the same cachet (and it looks like it will soon), and the rest of editing’s high-end has been dominated by Avid since the NLE replaced the flatbed.

Best,
Andy

[Andrew Richards]

[Herb Sevush] “So the GTX 580 would be better than the new 680?”

For Adobe or DaVinci duty, yeah, that’s what it looks like. Maybe if you are doing Maya or something like that the 600s could be better, though I’m not sure how much that kind of work overlaps with the things high-end gaming taps in those cards.

Compute performance is what you are looking for to accelerate NLEs (as I understand it).

Best,
Andy

[Andrew Richards]

Man, great interview! Love his take on the pyramid. Keep on disrupting, Mr. Petty!

Best,
Andy

[Andrew Richards]

Maybe Evertz buys them to consolidate their market? Or Miranda? This kind of thing seems a bit far afield for Mr. Petty, but you never know.

Best,
Andy

[Andrew Richards]

I was always struck by what looked like bloat at their NAB booth every year. A veritable sea of sales suits.

Best,
Andy

[Andrew Richards]

Interesting. So GPGPU duty is clearly viable over Thunderbolt’s bandwidth. Wouldn’t it be cool if you could get a dedicated GPGPU coprocessor like the guts of a Tesla in a small Thunderbolt case that eschewed the video out? If the compute functions of the GPU can be decoupled from the display (as that video demonstrates), that changes things. Let the GPU baked into the Ivy Bridge chips drive the displays and attach a compute-only GPGPU engine to accelerate enabled apps. Very interesting.

Best,
Andy