— ROGERS

A smattering dust can really ruin a good scan. Watching the local film processing guy drop your neg on his putrid floor (imbecile!) can raise your blood pressure faster than you can say C41!

So I’ve been having problems with little smatterings of dust on most of my scans. And the scanner I use doesn’t have Digital ICE (it’s an old Minolta – keepin’ it real). After using a few of the available software-based anti-dust systems out there, I remained unconvinced. On the discussion boards, people were saying to do it by hand. For bigger bits, sure, but a billion tiny grains…?

I decided it was simple to do better. I wrote (in the loosest sense) a Photoshop action which reduces dust, but doesn’t remove all the grain from the photo.

It is does all this with a difference matte made from the original (dusty) image, and a version that has been run through the Photoshop Dust & Scratches filter. Old school like.

It aint fancy.

But it is fast and reliable. Also, it doesn’t get rid of all the dust (yep, paint it). But will remove all that cocaine you chopped up on your negatives the night before… Or if you are more like me, the baby powder that they seem to sprinkle all my negs with when they leave the lab.


200805111434.jpg

See? The grain (mostly) survives. All that the matte is doing is revealing back to to D&S version of the image. Visually, it looks ok, and has probably saved me about a lifetime of manual healing brush time. Now you can get out there and shoot — post production is for the birds!

I make no guarantees this will work with your stuff. But have a go, it might.

200805111413.jpg
Read More

There is a great interview in Variety with James Cameron and his perspective on 3D films (pun intended). Of course, what he says is quite convincing, and salient. I met him once. He made some convincingly salient points then too. Via Daring Fireball, via John August.

Read More

It is always interesting when you take a diversion from concentrating on “visual effects” and go see what other people are doing with graphics technology. In many ways, the SIGGRAPH conference is a yearly fix of all things interesting, but you have to stop working to get there, and residing outside of the US doesn’t make it any easier. The alternative time-sucking solution, is web surfing.

Sometimes you come across something which really gives you a jolt. The work Robert Hodgin does is clever and beautiful:

Solar, with lyrics. from flight404 on Vimeo.

He uses Processing, an open-source programming environment for graphics and audio. See their exhibition space for more examples of what can be done. Robert Hodgin has a bit of his stuff on that page too, and great explanations of how the software was used (Birds/Flocking).

I’ve become quite interested in procedural animation of late, and have been wondering what we can do with Massive (and a bit of time) to create some visualisations. We’ve done a bit of stuff on films using Massive to create things other than crowds (like traffic, pedestrians, etc); so another logical step would be to use it with some input data to drive different aspects of it… hmmm… music clip anyone? With the right input data, I am sure you could make something quite special.

Read More

200804150740.jpg

RED announced the following new cameras, which I am sure will be dissected and analysed (and changed) over the next year or so. Release in 2009:

SCARLET – 3k resolution, supposedly under $3000; 2/3 sensor; fixed lens 8x T2.8 zoom. http://www.red.com/nab/scarlet. More information here. And a off the floor video via Videomaker:

EPIC – 5k resolution, ~$30,000; RED One owners get a 100% trade in; Full S35 sensor. http://www.red.com/nab/epic

Red Ray – a (blue ray?) based playback system up to 4k. http://www.red.com/nab/red_ray

Oh yeah, and some lenses.

The full S35 sensor in EPIC makes sense, especially with all the features that seem to have embraced the RED camera. The SCARLET will give Sony/Canon/Panasonic something to think about in that market (and a few other markets). As long as RED don’t crash and burn (my mother always warned me about things that seem too good to be true), I wouldn’t want to have an extensive future investment in telecines or film scanners. I wonder when the digital competitors will get themselves into gear?

Read More

May we all see like insects: light fields and plenoptic lenses. There could well be an interesting shift in specialised photography where extracting 3D data from digital images shot with either special lenses or sensors (and with a lot of post-processing/number crunching) become useful tools for VFX practitioners.

(Full disclosure: like most people, I’ve had a lenticular fascination since I was a kid. I just didn’t know what to call it…)

Now, of the technology, the immediate, demonstrated applications are variable focus (in post); or the ability to move the camera and change perspective within about a 10° arc. While this is something readily achievable in a layered composite, the implications of it being readily available and nicely packaged with a bow are quite interesting (see: focus stacking, helicon focus). And, if (if!) it can be applied to moving images, then there would depth information that you could use to extract layers. That is, you wouldn’t have to pull a key; you could go without green-screen; rotoscoping would be easy… oh, the possibilities… Okay, I’m getting way ahead of the technology.

But then, beyond the “3D” hype, there have been a lot of recent, significant developments in sensor technology, such as the Panasonic high dynamic range sensor… and even the CIA is hocking its image technology. Yes, the Central Intelligence Agency. Insert your conspiracy theories here. How about the Gigapan $300 (?) photo-robot…

While we may wait for a new digital camera in the coming months, we should probably bear in mind that Moore’s Law can be applied to digital photography and cinematography…

More reading: Refocus Imaging; Max Hodges comments; Stephen Shankland article; Photography 2.0/R&D

Read More