March 24, 2009
By Chuck Gloman
This exercise started out as a “shootout” between the Nikon D90 and Canon EOS 5D Mark II to see which DSLR camera captures better video. Because the playing field was not exactly even, this article will focus more on how to shoot a video using a still camera.
As a class project for my Basic Productions class at DeSales University, I had freshmen shoot and edit a video using the two newest digital SLR cameras – both shooting high-definition video. Although I should be comparing apples to apples (or camera to camera), I tried to keep things as consistent as possible although the two cameras clearly are not equal.
The first camera available to the public was the Nikon D90. The Nikon is a 12.3 megapixel camera with a 1.5 lens factor (because of the sensor size, you multiple your lens by one point five – if shooting with a 50mm lens, the D90 crops that to a 75mm). The video mode on the D90 is 720p 24p — high-def’s lowest resolution, but shooting at 24 frames progressive for that “film look.” The lens is an 18 – 105mm zoom so you have a decent wide angle to moderate telephoto. The other features on the camera include at 920,000 pixel, three inch LCD screen; built in microphone and speaker for sound (with no input for an external microphone or headphones); manual focus when shooting in video mode; recording of up to five continuous minutes onto a SD card; a disabled viewfinder when shooting video; and an exposure lock to keep your exposure set without the automatic gain control kicking it. The street price including the lens is $1,299.
Canon’s EOS 5D Mark II arrived later and offered quite a bit more. With a street price of $2,700, you got a 21.1 megapixel full frame camera with a 24 – 105mm zoom shooting 1080p footage at thirty frames per second. The features of this camera include: a 920,000 three inch LCD; built in microphone and speaker with a quarter inch jack for an external microphone; recording for up to twelve minutes onto a Compact Flash card; and like the Nikon, manual focus when in video mode, a disabled viewfinder, and an exposure lock.
Every shot in our 90-second comedy was recorded with the Nikon and Canon side by side. Both cameras were set at 400 ISO, manual white balance, and same lens setting. The only difference was that we used an external shotgun microphone on the Canon although both cameras did not allow control over the audio levels.
Always trying to be fair, the script consisted of interior and exterior scenes, spoken dialog, voice over, moving camera, and temperature, sound, and lighting extremes. I purposely had the students shoot exteriors in the snow (difficult to do indoors) to see how both cameras handled the extremes of exposure and contrast. Katie operated the Canon and Patrick the Nikon. Both camera’s support, a Liebeck tripod, were raised to the same level and “action and cut” were called so the two cameras would stop and start at the same time.
The Shoot
The first few shots were classroom interiors with a mixture of our studio’s florescent lighting high above the grid (acting as fill) and two 2K Fresnels as the actor’s key lights mounted on the ceiling grid. As expected, the images looked identical on both cameras because our ISO, lens, and white balance were exactly the same. Moving outdoors in the snow brought the challenges of exposing for the talent’s faces rather than underexposure because of the snow. The snow did appear brilliant white because of its overexposure, but neither camera seems to have an issue with it. 
The real issue was with the sound. Indoors, the Nikon D90’s sound was muffled because of using the on-board microphone and hearing only handling noise. We did not use the external microphone on the Canon EOS 5D Mark II indoors and also heard handling noise. Surprisingly, we also heard another problem with the Canon. While running, we could hear background camera noise much like constant scratching as if it were a film camera and one of the internal film spools was bent and scrapping the interior as it revolved. I’ve never heard digital SLR camera noise before and Sara was not scratching the microphone as we shot.
Outdoors, all we heard on the Nikon was wind noise with no usable audio. With no monitoring capabilities on either camera we could not hear this during the shoot. Being that the shotgun microphone inside a fuzzy when shooting with the Canon, the volume was lower (we used a professional low impedance microphone rather than a cheaper high impedance model) but sounded far superior without any handling or “film camera reel” noise as we had heard without the microphone attached.
Once the video project was shot, we returned inside and defrosted. Footage from both cameras was transferred to Corel Video Studio Pro X2 because it handles 720p and 1080p natively as is the first NLE software taught to our students as freshmen. The Nikon files were transferred from the SD card as motion Jpegs and the Canon’s CF files appeared at QuickTime movies. Neither had any problem loading into my Sony Vaio Pentium Four dual-processor laptop. Canon’s 1920 x 1080 files did take three times as long to transfer and load and the final result was striking to say the least.
The Nikon D90 footage was extremely soft, almost to the point of being blurry. It was much looking at standard-definition footage (but still in widescreen). As a stark contrast, the Canon EOS 5D Mark II footage was stellar – looking exactly like 1080p footage should look. I knew that cameras would not look the same – you cannot compare 720p footage with 1080p even though they are both high definition, the 1080p footage will clearly be the winner. With seven more megapixels and another $1,500 added to the price tag, the Canon cannot be touched in this arena. The students edited both projects and exported them to the same DVD. Corel allows you to export a still frame from the timeline with one mouse click. Note the differences between the Nikon and Canon footage. Although both frames are not the exact same millisecond in time, I believe you get the idea of the clarity between the two in these untouched frame grabs from the edit.
Comparisons
First, look at the interior scene with both cameras. The video on each was sharp, but when a still capture was exported, the evidence speaks for itself since the F-stops and the depth of field were identical. The D90’s parameters as a direct frame export were: 1280 x 720 (the 720p), a width of 17.778 by 10 inches in height at 72ppi (pixels per inch). The EOS 5D particulars were: 1920 x 1080, 26.667 by 15 inches in height also at 72ppi. Even without the specifications, a night and day difference.
The exterior grabs were not as stark in their differences, but the D90 is clearly a little softer with much less detail in the snow.
A Lesson Learned
I did this exercise with my students for several reasons: I wanted to give them access to the latest in digital technology from the giants in still photography; I desired the “test” between the competitors to be as close as possible (much like Motor Trend used to compare the Mustang/Camaro and Lincoln/Cadillac) – the only real differences to note was the price and high definition resolution; shoot and edit a complete movie with digital footage from still cameras; and how to prepare them for future tests like theses when they have to make the determination in the real world on which manufacturer’s product from which to choose.
Although the deck may have been stacked slightly in Canon’s favor (it clearly is a better camera for this type of production), I wanted the students to see for themselves how the cameras performed instead of just taking my word for it (I’m wrong too many times).
I look forward to more projects like this in the future to see exactly what the new technology can do. The only true test to determine which works better for you is to try them both out in the type of production you do. The Nikon D90 is a great still camera and shoots high-definition video. Instead of grabbing a still frame from the video – if you need a still, that’s what the camera was originally designed to do. The only time it will fall apart is when you compare it side by side with the EOS 5D which is the better camera here at twice the price.
If on a limited budget, go for the Nikon. If you have more to spend and really need and expect the best, choose the Canon.
Chuck Gloman is an awarding-winning Producer/Director of Photography with over 800 commercials to his credit, and is program director of the TV/Film Department as well as a member of the faculty of DeSales University. He may be reached chuck.gloman@desales.edu.
| COMMENTS (13) | | 04/01/2009 | | Another option for the scratch noise would be the CF cards. Some flash memory based cards I have used have a barely audible "tick" noise when the CF cards have any write / read activity. This is most likely the result of noisy electronic components, but the most likely reason is the IS as noted above. The lack of mic level controls will (on purpose by Canon) keep the 5D from edging into the more serious end of videography of their A1 .... but as noted above, simply using an external audio recording device will eliminate that restriction allowing the 5D to focus on obtaining great images (which it does very well). |
| | 03/31/2009 | | Can you upload some of your test frames and footage for comparison? |
| | 03/25/2009 | | Video? Not really; at least for the Canon which shoots actual 30 FPS. I can't say for the Nikon. There is no HD 30 FPS video standard; 59.94, 29.97, 23.98, yes. Think the difference doesn't matter? Think again. No sync sound, horrible frame rate conversion artifacts, We did extensive and expensive tests for use on a feature. Conclusion - unusable for anything other than MOS B-roll where you can tolerate the slowdown from 30 to 23.98. |
| | 03/25/2009 | | Anyone who wonders about the amazing super kick-*ssness of the Canon's video needs to go check out this video (and make sure to also watch the "making of" clip):
http://www.usa.canon.com/dlc/controller?act=GetArticleAct&articleID=2326
Sold. |
| | 03/25/2009 | | the review needs to be redacted, per the audio comments. gives a wrong impression of the canon and shows a lack of knowledge re: lens stabilization noise, a well noted effect easily found with any web research. simply turning off the stabilization, or using the mic jack for placement of mic a bit away from lens fixes noise completely. |
|
|
|
More...
|
|
|
|
|