Ok, right off the bat can I just clarify that I am a canon 7d owner. Like anyone who's spent a not inconsiderable amount of money on a nice new toy I should be happy when I read an article that shows that new purchase in a positive light, right? I should, and usually would, but a friend has just sent me an article from the latest issue of TVBay (035) that has really got my goat. It's completely, factually wrong!!
The offending article is "Right tools for the job - RED or dead" by Senior Engineer, Andy Grant. The gist of the piece is that the RED camera revolution may have burnt itself out by kickstarting the likes of canon into making cameras such as the 7d - cameras that would have been the stuff of dreams pre RED. OK, interesting view. Don't entirely disagree - RED are naturally going to face some competition, but not from the 7d and certainly not for the reasons given in this piece.
Fact or fiction 1 - "Both (5d Mk 2 / 7D) scale their video down to 1080x1920 video and have an uncompressed HDMI output... this can then be recorded on a wide range of easily available devices including the sony SRW-1 virtually uncompressed HDCAM SR recorder"
- er I think you might be confusing your dreams with reality there. Yes these cameras do have an HDMI output and yes it would be great if you could use it to capture uncompressed data but they can't! They won't even let you record a clean, non-UI version of the live video feed, let alone an uncompressed one. Maybe a future firmware update, either an official one from Canon or from the likes of MagicLantern, may unleash that potential in the future but this statement certainly isn't accurate now.
Fact or fiction 2 - "...the image is scaled in camera (RED ONE). From experience I've found this to be poor quality"
- Andy, you run a rental house. Have you even used the RED ONE camera? There is NO scaling in camera. The different shooting modes are windowed - meaning cropped and therfore retaining a pixel-to-pixel ratio. Please tell me you haven't been looking at the proxie files and treating them as ready to use, master files? If so, no wonder you can't get good greenscreen keys off them. Try implimenting a proper workflow and scaling the r3d files to your desired output using Red Alert or Red Cine - trust me you will see a difference. That is proper scaling. "From what I've seen the canon looks far better in HD" - ok, now I know you're just joking. You are, aren't you?
I am impressed by the footage I can get out of my 7D but as a user and DIT for the RED ONE I know there just is no comparison. The 7D uses lots of tricks and fast and dirty "scaling" techniques such as line skipping. Just see this excellent post on the true resolution and artifacts that such processes result in.
I could go on and talk about rolling shutter but I think I've made my point. If this article was a deliberate ploy to cause controversy then I think it will work - but don't expect your readers to stick with you if the "expertise" is so full of holes...
Friday, December 11, 2009
Monday, November 02, 2009
7D Greenscreen Keying Test
A couple of days after getting my 7D I did some DIT work on a shoot for Animals on Blue, which involved shooting various big cats against a greenscreen using a RED ONE camera. Not only was this a great opportunity to put the 7D's stills capabilities to the test, it was also a chance to grab some video from the sidelines - including footage that I later realised I could use to test how well 7D footage holds up for chroma keying.
First it's important to note that I'd didn't plan for this at the time of shooting. The camera was hand-held, set to auto white-balance, iso 400 and the Lynx that we were filming was stubbornly sitting outside of the lighting sweet spot. That said, I was really impressed with how easily I was able to pull a key. Importing the raw MOV file into After Effects CS4, I applied Keylight and a few clicks and tweaks later I had managed to pull a key that was pretty encouraging. With a bit more tweaking in post, and more attention to my camera's settings at the time of shooting, I think it would be possible to get results that are cleaner still. Not bad at all for a camera that shoots to 8bit 4:2:0 h264.
Check out the results for yourself:
First it's important to note that I'd didn't plan for this at the time of shooting. The camera was hand-held, set to auto white-balance, iso 400 and the Lynx that we were filming was stubbornly sitting outside of the lighting sweet spot. That said, I was really impressed with how easily I was able to pull a key. Importing the raw MOV file into After Effects CS4, I applied Keylight and a few clicks and tweaks later I had managed to pull a key that was pretty encouraging. With a bit more tweaking in post, and more attention to my camera's settings at the time of shooting, I think it would be possible to get results that are cleaner still. Not bad at all for a camera that shoots to 8bit 4:2:0 h264.
Check out the results for yourself:
Some of my stills photos of the big cats can be found here
Wednesday, October 28, 2009
Field of View and Focal Length Confusion
A conversation I've found myself having alot recently is one trying to explain the concept of field-of-view and focal length. The trigger is nearly always a digital SLR and what its particular sensor size does to focal length compared to normal 35mm. There are two sources of confusion immediately, right there.
Firstly, focal length does not change. It is the distance between the the sensor/film/plane of focus and the end of the lens. A 50mm lens on one camera will therefore have the same focal length as a 50mm lens on another. What does change is the field of view - the angular extent of what can be seen. The smaller the sensor or film size of a camera the narrower the field of view for a lens of the same focal length - i.e. less of a scene can be seen. To get the same field of view on an APS-C camera (e.g. 7D) as a 50mm lens would give on a full frame one (e.g. 5D mk2) a 32mm lens would be required. I can see why many people try to explain this by saying things like "a 32mm lens on a 5D becomes a 50mm on the 7D" but the focal length does not suddenly become something else.
The second cause of confusion revolves around the terms "35mm" and "full frame". What they mean depends on whether you are talking about the world of stills or motion photography - a fact not helped by the fact that the conversation is inevitably about shooting video on a DSLR. If you're comparing a DSLR and the visual qualities inherent to the size of its sensor, then I think the moment you talk video you should only compare it to 35mm feature film. There are countless posts online deriding the Canon 7D for having a smaller sensor than its "full frame" 5d Mk2 sibling but I bought the 7D specifically because of that fact. That "small" sensor is pretty much the same size as "full frame" 35mm motion picture film - which is a huge advantage when you want consistancy between lens focal lengths and the resulting field of view, which I do. Not only is the 7D capable of capturing some great images in its own right but it's also an extremly useful director's view finder. Vistavision (the film equivalent of the 5D's sized sensor) is a niche format, so if cinematography, practicing or learning, is the principal intended use of your DSLR then I say go 7D - all other advantages and disadvantages aside.
Anyway, for those of you who, like me, find themselves switching between formats on a regular basis, I thought it would be useful to put together a little table comparing the field of views for some common focal lengths on some of the usual suspect sensor/film formats. Couldn't have done it without this great calculator
Firstly, focal length does not change. It is the distance between the the sensor/film/plane of focus and the end of the lens. A 50mm lens on one camera will therefore have the same focal length as a 50mm lens on another. What does change is the field of view - the angular extent of what can be seen. The smaller the sensor or film size of a camera the narrower the field of view for a lens of the same focal length - i.e. less of a scene can be seen. To get the same field of view on an APS-C camera (e.g. 7D) as a 50mm lens would give on a full frame one (e.g. 5D mk2) a 32mm lens would be required. I can see why many people try to explain this by saying things like "a 32mm lens on a 5D becomes a 50mm on the 7D" but the focal length does not suddenly become something else.
The second cause of confusion revolves around the terms "35mm" and "full frame". What they mean depends on whether you are talking about the world of stills or motion photography - a fact not helped by the fact that the conversation is inevitably about shooting video on a DSLR. If you're comparing a DSLR and the visual qualities inherent to the size of its sensor, then I think the moment you talk video you should only compare it to 35mm feature film. There are countless posts online deriding the Canon 7D for having a smaller sensor than its "full frame" 5d Mk2 sibling but I bought the 7D specifically because of that fact. That "small" sensor is pretty much the same size as "full frame" 35mm motion picture film - which is a huge advantage when you want consistancy between lens focal lengths and the resulting field of view, which I do. Not only is the 7D capable of capturing some great images in its own right but it's also an extremly useful director's view finder. Vistavision (the film equivalent of the 5D's sized sensor) is a niche format, so if cinematography, practicing or learning, is the principal intended use of your DSLR then I say go 7D - all other advantages and disadvantages aside.
Anyway, for those of you who, like me, find themselves switching between formats on a regular basis, I thought it would be useful to put together a little table comparing the field of views for some common focal lengths on some of the usual suspect sensor/film formats. Couldn't have done it without this great calculator
| Horizontal Field-of-View (degrees) for given Focal length (mm) | |||||||
| Format | Size (mm) | 17 | 24 | 35 | 50 | 70 | 100 |
| 35mm (motion) | 21.95 x 16 | 65.7 | 49.1 | 34.8 | 24.8 | 17.8 | 12.5 |
| Super-35mm (4-perf) | 24.89 x 18.66 | 72.4 | 54.8 | 39.1 | 28 | 20.2 | 14.2 |
| 35mm (stills) | 36 x 24 | 93.3 | 73.7 | 54.4 | 39.6 | 28.8 | 20.4 |
| Red One | 24.4 x 13.7 | 71.3 | 54 | 38.4 | 27.4 | 19.7 | 13.9 |
| Canon 1D IV (APS -H) | 27.9 x 18.6 | 78.7 | 60.3 | 43.5 | 31.2 | 22.5 | 15.9 |
| Canon 5D II (full-frame) | 36 x 24 | 93.3 | 73.7 | 54.4 | 39.6 | 28.8 | 20.4 |
| Canon 7D (APS-C) | 22.3 x 14.9 | 66.5 | 49.8 | 35.3 | 25.1 | 18.1 | 12.7 |
SohoMarky is dead! Long live Digital-Negative!
All right, all right - I give in! In response to my friends' nagging requests to extract some of the technical ramblings lurking in my brain and insistance that I put fingers to keyboard and put some of them "out there" - I have killed off the old, neglected (by me) "sohomarky" site and relaunched it as "Digital-Negative". A blog where I hope to post some digital-cinematography related geekery on a semi-regular basis. And so it begins...
Subscribe to:
Posts (Atom)
