So I just saw Zodiac and I didn’t really like it mostly cause it was too long. They could and should have cut about an hour out of it. My post here however isn’t about the movie itself its actually about the look. I saw the trailer and I was very apprehensive about the amount of gain they had used in low light scenes. I was thinking this would look like another Miami Vice (I hated that look) but when I saw the feature itself there was absolutely NO NOISE in the picture. NONE AT ALL (other then the grain on the print itself). There was also no lose in detail in the scene that before – in the trailer – had almost nothing but noise.

How did they remove it without loosing detail?

I hope this is the right place to post this. I wasn’t sure if it should be in the lighting thread, visual-effects thread or this one.

I don't really know, but couldn't the noise in te trailer be caused by a bad compression for the trailer to put on internet? Dark footage looks imo pretty bad on the internet most of the time.

But then again, it's not really my expertise

Trailers are a double-edged sword. If the distributors know the movie is bad they might up the budget for the trailer and sometimes the trailers are better than the movie! Other times the trailer promo company gets one-lights and do not have the budget or time for extensive grading.

Technicolor TDI, which did the transfer to 35mm, has their own noise/grain reduction software. It's always a trade-off though, especially when things move in the frame, but I guess it wasn't noticeable here.

I saw the trailer online and in the theater and both looked the same, as did the Miami Vice trailer and the theatrical released feature. But the theatrical released feature of Zodiac looked clean and smooth with none of the normal signs or noise removal, like blurriness or loss of fine detail. Only the noise seems to disappear. David the thing that struck me was that there was no trade off, it looked great and had no noise. I really want to find out how they did it because it was very well done.

In general, the 1080p trailer will have more detail and thus more grain than the 35mm prints that go into mass distribution. The washed out blacks and compression artifacts from h264 don't help.

In general, the 1080p trailer will have more detail and thus more grain than the 35mm prints that go into mass distribution.
How is that possible? Isn't film greatly higher in resolution than just humble 1080?

If not, this gives me huge hopes for a 35mm blowup of a feature I am shooting on an FX1.

I'm saying that the 35mm trailer was just as bad as the online trailer. The online trailer might have been slightly worse but not by much. They did somthing to remove the noise in post I just dont know what program they used.

If you blow up an FX1 to 35mm it will look the same as it did before because the issue is compression. FX1 footage is already compressed about 75:1 compared to HDCAM SR footage. Firstly the FX1 will only record color information every 4 frames. It will only record 1080i (so you will have to deinterlace it and loose half of your vertical resolution). The chips inside are actually only 960X1080 so because of that you lose half your horrizontal resolution. It uses MPEG2 encoding so every 15th frame is a "real" frame, and all the others are based off of that one frame (basically it figures out how the frame changes and saves that information). It dosnt shoot 24p. Basically if you blow up FX1 footage it still looks like FX1 footage.

Zodiac was shot 4:4:4 full HD. Meaning its 1080p24 and every frame has color information and is its own contained frame - not dependent on others for its information. Its like shooting 16mm without any grain. The compression comes when its being turned into a quicktime movie and put up on apple's web site. The compression in the FX1 comes right when it lays information down on tape.

I personaly think its better to shoot pregressive SD24p then interlace HD60i. U get more information per pixel and u get to use them all, where as in interlaced HD you will essentialy have to drop half of the pixels (more really).

As with the rest of things they do in movies: They probably faked it.

There was probably a lot more light than you think and they knocked it down in editing with plugins.... That way they get the resolution and no artifacts but they also get "the look".

Thee's a very good article in POST magizine about the cmaera used (Viper) and how the achieved the film's look:

http://www.postmagazine.com/ME2/dirmod.asp?sid=&nm=&type=Publishing&mod=Publications%3A%3AArticle&mid=8F3A7027421841978F18BE895F87F791&tier=4&id=222741C1B78640368D9229819772A1B8

It's a good read....
john
evilgeniusentertainment.com

How did they remove it without loosing detail?

There is an article in the May 2007 issue of "American Cinematographer" about this. They used the serivices of DTS in Burbank to basically analyze a bunch of frames and derive information to upres and sharpen the image.