Jump to content

Adolfo Rozenfeld

  • Content Count

  • Joined

  • Last visited

About Adolfo Rozenfeld

  • Rank
    MoGraph Demi-god
  • Birthday 08/07/1971

Contact Methods

  • AIM
  • Website URL
  • ICQ

Profile Information

  • Location
    Buenos Aires - Argentina
  1. PPro can edit AVCHD in real time without conversions. Even more, it can do that with DSLR footage as well. I am not even talking about the Mercury engine - I can edit H264 from my Canon DSLR with basic spatial transformations and (also basic) color correction in RT in my humble Macbook Pro. Conventional wisdom suggests that native editing of long GOP formats like these is a bad idea, but they do work quite smoothly in PPro CS5... on a laptop.
  2. It's not Continuously Rasterized. Instead, the transformations are Collapsed (which is what the sun switch does to Pre-Comps. Same switch, different purpose). It's not esoteric or impossible to figure out. Collapsing transformations means combining them. For example, a nested layer is scaled down to 25 per cent and then the Pre-Comp which contains it is scaled up to 400 per cent. It will pixelate, unless you collapse/combine transformations. In that case, the two scale properties are combined. A non-collapsed Pre-Comp is a lot like if you rendered a file and brought it back in - AE doesn't know what's going on inside. A collapsed pre-comp keeps the connection between external properties and internal ones live. A collasped Pre-Comp can also show content outside of its' own boundaries in the parent Comp. A layer with a blend mode assigned - if it's collapsed, the nested layer's blend mode will interact with layers outside of the Pre-Comp (this is why a Collapsed Pre-Comp can't have a blend mode assigned, unless you force it). Same for an adjustment layer - if the comp is collapsed, the adjustment will affect layers below, in the parent comp. All these things are amazing aspects to collasped pre-Comps but they can alter drastically the end result of your project if you don't understand how they work. Applying an effect or a mask to a 3D Pre-Comp with Collapse Transformations on will break the 3D compositing bin, ie that pre-Comp will loose the ability to intersect in space with other 3D layers.
  3. Manually clamping luminance and color ranges, and even the Broadcast Colors effect, are old practices based in old theory. Especially for > 8 bpc projects or projects which use wider color spaces than NTSC, PAL or REC 709. Using the limited SDTV/HDTV 16-235 output profiles (do not use them as working space, only as output in the Output Module) produces more accurate and safer color transforms.
  4. www.photozone.de: "There were a few moments when I considered not to publish the results due to "political correctness" because to date it was a quite absurd thought that such a cheap, or better "affordable", lens can perform this good and I'm sure that some will not believe the findings even though they're supported by the published field images. Anyway, the resolution capabilities of the Canon EF-S 18-55mm f/3.5-5.6 IS is nothing short of amazing." http://www.photozone.de/canon-eos/181-canon-ef-s-18-55mm-f35-56-is-test-report--review www.dpreview.com: "Canon have indeed upped the ante with this new lens; it's much better than its predecessor, especially in terms of sharpness in the corners and at wider apertures, and with reduced chromatic aberrations to boot. Indeed overall it's an extremely well-behaved little lens, with very few nasty surprises for the user, and a remarkably good image stabilisation unit; indeed Canon's main concern may ultimately become whether users have as much incentive to upgrade to more expensive optics as they did before." http://www.dpreview.com/lensreviews/canon_18-55_3p5-5p6_is_c16/ www.slrgear.com: "The 18-55mm ƒ/3.5-5.6 IS shows excellent sharpness at 18mm wide-open at ƒ/3.5, a focal length/aperture combination which typically shows off the weakness of most consumer-grade lenses. Image sharpness is essentially even across the frame, and doesn't exceed 2 units on our scale of 1-12. For a lens of this price point, this is amazing performance" (...) "diffraction starts to set in around ƒ/16, but you don't really start to see mediocre performance until the aperture is stopped down to rarely-used f-stop numbers such as ƒ/29. At that point, the image is quite soft, but at least the softness is uniform at between 4 and 5 units. All in all, excellent performance." http://www.slrgear.com/reviews/showproduct.php/product/1114/cat/11
  5. They don't like it because the previous version was much worse and because kit lenses historically have been awful. I understand that not only Canon, but also Nikon, raised the bar for kit lenses. Still, a decent zoom which will only open up to to f/5.6 in the telephoto end is not quite the sexiest lens in the world. It's just not the piece of crap everybody (myself included, until recently) think it is.
  6. Actually, the IS flavor of the 18-55 kit lens that's being shipping with Rebels for the last couple of years got amazing scores in every review. It's night and day compared to the previous version of the same lens, and it outperforms Sigmas and Tamrons zooms costing two or three times more. I was skeptical too, but the conclusions are very consistent everywhere. The resolution is incredibly high, even at the widest apertures (which are not very wide, of course). So, the main issues with the kit lens are slow apertures, build quality, lack of a distance scale, lousy focus ring, etc. In every other aspect, it's in the tradition of the 50mm f/1.8 or 28mm f/2.8 cheap primes, in the sense of giving awesome quality in a cheap enclosure. BTW, these two primes are no-brainer. The 28mm f/2.8 beats the more expensive f/1.8 version in almost everything but max aperture and focus speed. Unfortunately, they went up in price like 20-30 per cent in the last few years, not sure when. So they're now a super bargain, but not a super ultra bargain.
  7. OpenGL in AE is not really meant to be used as a primary render technology. A lof of things are accelerated for it, but they simply won't offer the same quality as the software renderer. This is why the "OpenGL - Always On" mode is kind of experimental... or hardly useful in most cases. But this is also why the "OpenGL - Interactive" mode is probably the most useful - The GPU muscle will accelerate many things as you interact with elements (drag layers around, transform elements, use camera tools, etc) to give you a draft but useful preview, and then the software renderer kicks in when you release the mouse button to render the frame at full quality. It is true, though, that it only takes a single non-accelerated effect (especially third party plug-ins) to break the whole chain. In such a case, you may want to turn off OpenGL completely. Many things that used to make OpenGL a moot point, like nested Comps, now do work. So it depends on what you're doing. FWIW, I have it set to "OpenGL - Interactive" all the time. If you use third party plug-ins in all your projects, then yes, it may not be very useful.
  8. Actually, Shape layers take a lot of clues from Text Animators. Path effects like "Wiggle Transform" and "Wiggle Paths" have parameters which are almost identical (Wiggles per second, Max.amount, Min.amount, Correlation, etc) to the Wiggly Selector. I think text animators are actually fairly intuitive once you grasp the basics. And there's very little you can't do with them. In this case, you could achieve what you want with either two animators or two selectors combined in a single animator. The latter could be more elegant, but the former is probably easier. Create two animators with a scale property each. The first animator has a range selector. In the second animator, the range selector is replaced with a wiggly selector. Use the first animator to scale up all characters, which you can do by keyframing the scale property or the amount property. Use the second animator to wiggle scale. I can make a sample project or even an animation preset if you want (but next week if that's ok, since this week I have a personal issue that keeps me away from computers).
  9. Careful with what you wish for, because you may get it I think I am sticking to well-known facts if I remind you that After Effects is a Carbon application on Mac. And that at the 2007 WWDC, Apple ditched its' own plan and killed Carbon 64 (with several applications caught somewhere in the 64-bit transition). Also, you surely know that Quicktime 64 doesn't exist on Windows.
  10. It's not possible, and as long as AI keeps using the same technology, it won't be possible. The extrude generator in AI doesn't generate actual 3D geometry. So there's nothing there to export/exchange.
  11. I think it has a small mic plug, but you could use an adapter. The problem is that it only has AGC, no manual control over audio levels and the typical problems associated with that. It has HDMI out (limited to SD, I think).
  12. That's not weird. The whole trick is based on a pattern of duplicates. If the sample number is too low, the trick becomes evident. 10 may still be a small amount. From CS3 onwards, MB is adaptive. This means AE dynamically increases or decreases the number of samples on the fly, to achieve smooth MB with no evident stair stepping. The "classic" fixed sample number (pre CS3) was 16. It can now be adaptive up to 255. For some features (anything 3D, for example), it's still fixed - but you can still set the amount, up to 128 samples (which could be really slow to render).
  13. Do you have any keyboard around you could plug to check if it goes away when you do that? At least to rule it out.
  14. You're on Windows, right? Is it possible that the Num lock key is active? Try switching it (disable if enabled or the other way around). I am thinking the shortcut for "Save RAM preview" is Control + Num Pad 0, so that *may* explain why the Output movie dialog appears.
  • Create New...