It's kind of strange that the most anticipated sample media from the latest crop of DSLRs is actually HD video, but that's the way the market's going, and the new
Canon EOS Rebel T1i and
Nikon D5000 are no exception.
Demystifying Digital has footage from both, and while Nikon's newest seems to deliver essentially the same CMOS jelly-motion results from its 12.3 megapixel sensor as the
D90, the T1i is a pleasant surprise -- we're not seeing much wiggle at 720p, which might be a result of the higher-res 15 megapixel sensor. We'll wait to see full-res samples and some actual still images before we crown a winner, but we'll tell you right now that neither one of these is going to make us leave a dedicated camcorder behind when it's important -- the lack of autofocus and the windblown internal mics just don't cut it. Videos after the break.
Read - Nikon D5000
Read - Canon T1i
posted Apr 15th 2009 3:42PM
Both read links seem to point to same URL... Nice videos though
Keep your pants on!
The Canon's video does seem a lot better, but then again they weren't walking around with the camera taking video. I would like to see some side by side footage to really compare.
I agree. One is being held stationary while the other is being walked downhill - these two videos cannot be compared at all.
At Youtube resolution, its hard to tell either way. The D5000 certainly has the rolling shutter of the D90, the T1i may have it as well, but since its stationary its impossible to decipher.
It's already been proven by Canon's 5D that jellyness is a trait exclusive to Nikon cameras. Apparently the D5000 is no exception.
ummm, the 5D is in a slightly - and let me emphasize SLIGHTLY - different league than either of these cameras. And by slight I mean $2,000, give or take a few C-notes.
So, you know, drawing sweeping conclusions based on a camera that's several stratospheres above these two makes LOTS of sense.
@ticohans
Since Canon has been doing consumer and professional camcorders all along, it just make sense that Canon does indeed have a big advantage on making a video recording function work as it should. The price difference between Canon 5D Mark 2 and T1i/500D is due to the sensor and build quality, I would not be surprised if they share the same video implementation (with a slower processor). On the other hand, Nikon D5000 basically shares the same implementation with D90, which really has very limited use.
Click on the video to jump to the youtube page, then click on the HD button and you get a full resolution look, although it's obviously re-encoded by youtube so what it tells you about the actual footage I have no clue.
How can anyone judge video quality through YouTube? Geniuses! Heard of Vimeo?
+1
Actually, as much as I love Vimeo, the most recent HD video comparison I saw somewhere, YouTube actually had the upper hand in fine detail in their HD videos.
Of course, I'd trade nothing for the Vimeo community and general usability of the website compared to YouTube.
@Milas
If you have a vimeo account, and the author allows it, you can download the original source file from vimeo. You can't do that with youtube afaik.
THANK YOU.
Someone needs to mount both cameras on a cross-bar and shoot the same scene... with MOTION.
I guess we're still waiting for a true head-to-head video test.
It's already been shown that Canon's as much of dramatic jello issues as Nikon's.
this is a clear display that nikon still hasn't addressed it.
Mounting the cameras is the last thing you want to do for this test. You need full frame shift to clearly see the jello effect.
>> "Mounting the cameras is the last thing you want to do for this test."
By mounting I meant having the 2 cameras hooked to something and shooting the same thing. Like both hooked to a Fig-Rig or something. Or even shoot the same shot for the test!
My point is that if these are gonna be the groundbreaking, first 2 clips from these cameras... and one shot is a shaky handheld shot, and the other is a motionless wide shot... it's a pretty shitty test... isn't it?
true, a head-to-head would give you a great comparison, but why would you buy a DSLR camera for video purposes?
>> "but why would you buy a DSLR camera for video purposes?"
I don't know. But if there are people that want it, there are manufacturers that will build it.
It's not that big a deal. Think about it... people want SLRs, right? And would they rather have to carry a FlipVideo or not get *any* video at all?
As long as the video features don't ruin the still photography side of things... I don't care. I guess my next DSLR will include video, huh?
@pete
"but why would you buy a DSLR camera for video purposes?"
Because you can use all those 35mm and APS-C lens to be creative. You either can't do that with a consumer camcorder, or it will break the bank with a professional camcorder.
Also you don't want to pay 3 grand twice to get a sensor and some electronics to shuffle data and a lens when one device already has all those things, plus just a DSLR is already overly bulky to carry, so adding another device is going to be over the top.
WIN!
This isn't digg.
This comes as no surprise, Canon knows about making HD video cameras (whereas Nikon should stick to photography). Pros trust Canon for their HD broadcasting and photographic needs, and the T1i carries that heritage. I'm predicting that the T1i is going to be a hot seller this year.
>> "Canon knows about making HD video cameras"
Then why did they pick a 20fps mode? Plus, I doubt the same engineers who make Canon's high-end pro video cameras were working with the team who built this entry-level DSLR. They took a DSLR and added the ability to make a movie. The T1i uses a 15MP sensor... there's nothing like that in Canon's stable of pro-camcorders... pro-camcorder sensors are 2MP. Sure, the DNA at Canon runs deep... but a pro-camcorder is an entirely different beast than a T1i.
>> "I'm predicting that the T1i is going to be a hot seller this year."
Really? $900 cameras will be hot sellers in this economy?
Really? $900 cameras will be hot sellers in this economy?
Yes.
Erm, and canon P&S's were the ones longest sticking to 640x480 video when the rest of the market already moved on to larger formats.
I can see the logic of your statement but reality isn't logical I'm here to inform you.
You can change the manual focus during a scene, right? I don't really know why you'd ever use autofocus for video, every time it changes it looks crap. Of course, if you can't use either, you're going to have to do more takes to get an interesting range of shots.
Geez. I don't trust autofocus for still pictures. What's the point of having it for video? Especially if you're concerned about noise. Those aren't the hoverlenses of the future you've been waiting for, after all.
who the hell would want to shoot in 720 anyway? its no longer any sort of standard, you've either got standard def or full on hi-def (1080i or 1080p).
i don't care how good 720p looks on a DSLR because I don't plan on buying A) outdated video resolutions or B) video that relies on autofocus or C) video that comes from the shakiest kind of camera available.
also if you plan on shooting hi-def video, you should already know that internal mic's NEVER get the job done. This article makes no sense
What are you gonna do with 1920x1080 video? Burn it to a Blu-Ray and mail a disc to your Grandma? No, you're probably gonna compress it and put it on the web. Outdated resolution? Hulu streams in 640x360... is that weird? 1280x720 is fine.
Oh, and you can have your 1080p Full HD.... but it's at 20fps... yuck.
Most of the people who buy this camera will just dump the videoclips and JPGs from the camera onto their computer and forget about them. If you're a cinematographer... you're not buying this camera.
On this camera, 720p beats the 1080p mode. I love Canon, but the 20fps choice is a bit questionable. 720p is quickly becoming a new web standard. YouTube HD will be good enough for quite a while, and alot of services still aren't even that. File sizes are reasonable with new internet connections, and start streaming relatively quickly. Decoding isn't very taxing either, and will usually playback fine on 5+ year old computers. 720p is here to stay for the long run.
I have no idea what "full" hi def is, please explain Mr. Best Buy salesmen
I guess someone should tell ESPN, ABC and Fox to give it up with their "HD" lies. Broadcasting in 720p! Who wants to watch sports in a progressive format, give me interlaced!
Progressive is better than interlaced. 720p for sports is better than 1080i b/c of the need for faster updating of the pixels.
Jon was being sarcastic.
you're welcome.
Another posted already alluded to this, but I it doesn't what the resolution is if it's only 20 fps. That's absolutely silly. I'd rather have 480i video at a higher frame rate than 1080p at 20 fps. Canon should have at least done 24 fps.
best buy salesman? im sorry, but show me a definition of "High Def" that exceeds 1080p. And broadcasting in 720 =/= shooting in 720. You can always change the resolution of your video to something smaller, but you can't make it any bigger.
Yes, watching on 720 is fine, because all true HDTV's are at least 720p, but to shoot in 720p while paying over a grand for the camera is ludicrous. If anyone plans on buying these cameras, they are looking for high quality photos with total control over the exposure and composition. To say that these people would be making video's for their grandma's or for youtube does not make any sense.
When you say 720p is outdated, just think about how many computers these days can actually edit a 720p video with ease, not to mentioned 1080p. Of course, it would be nice if they have 1080p at 30fps, but 720p is not bad at all, Consider all HD movie rental right now is still 720p.
again, recording 720 is different than broadcasting, distributing and exbiting 720. with 720 as your maximum resolution, you have absolutely no room for error when it comes to composition because what you see is the largest image you can pass along to audiences. if you attempt to crop the images even the slightest bit you immediately see the results: noise, grain, pixelization... whatever you want to call it. there is a reason that film and tv do not record in the exact same format that they broadcast/distribute, and if you can't understand that, than you don't understand video.
also, hi-def wasn't meant for computer screens when it was developed. you can't say that hulu's 480p looks good on everything just because it looks pretty good on your computer screen. just try watching it on a 50inch plasma.
wow that teddy-bear post got taken down pretty damn quick
What Teddy Bear post?
The pedo-bear showed his perverted head and was quickly escorted out of the room by police.
Totally an unfair test but it looks like Nikon still has not figured out how to fix the magic jelly bean effect. For a true test the cameras need to be side by side and be doing the same thing. BTW I am a canon person.
>> "Nikon still has not figured out how to fix the magic jelly bean effect."
It's not magic, but it will probably never be fixed. It's the way CMOS sensors work.
RED has overcome this with a rapid read-reset CMOS sensor. Let's hope Canon does something similar.
@Michael Scrip
>> "It's not magic, but it will probably never be fixed. It's the way CMOS sensors work."
I don't think you should make false claim like that as if you are an expert. I guess you just read that somewhere and blindly believe what it said.
I think Canon is planning to use CMOS sensor on their professional video camcorder in the future, they just know how to do video a lot better than Nikon.
Just look at this video from the 500D's big brother 5D Mark II, I don't see jello effect as big of a problem here,
http://www.youtube.com/watch?v=oHOxyUo9IWs
>> "I don't think you should make false claim like that as if you are an expert. I guess you just read that somewhere and blindly believe what it said."
Ummm, I did read that somewhere... everywhere as a matter of fact. That's how CMOS sensors work. It captures the image from top of the sensor to the bottom, so when you're panning, it will skew or wobble the image.
Here's some more info for you non-believer: (why couldn't you look this up yourself?)
"Rolling shutters are found on CMOS video cameras. This method of image acquisition sequentially records the image in rows from top to bottom, one after the other, so that not all parts of the image are recorded at exactly the same time. Therefore, the subject being recorded may actually have moved a bit from when part of one frame is recorded to when another part of the same frame is recorded. In extreme conditions of motion and due to fast flashes of light rolling shutters can cause such effects as:
Skew: The image bends diagonally in one direction or another as the camera or subject moves from one side to another, exposing different parts of the image at different times.
Wobble: This phenomenon is most common in handheld shots. The rolling shutter causes the image to wobble unnaturally and bizarrely. This typically only shows up in somewhat extreme circumstances."
And:
"On the other end, rolling shutter exposes the frame from top to bottom. Imagine a TV as it writes an image line-by-line. A CMOS sensor captures light though capturing each pixel one-by-one. Each row of pixels (from top to bottom) is captured in sequential order, creating a rolling effect, hence the name "Rolling Shutter". A CMOS sensor will appear to stretch or "skew" the image in either horizontal direction, making straight lines appear to bend in a diagonal fashion.
Do you still doubt me, Tin? Now, Canon might be better than Nikon at the moment... but the jello/wobble problem is because of CMOS.
@Michael Scrip
To avoid the jello effect, it has to make sure the image processor read the data from the sensor fast enough, at least faster than the appropriate shuttle speed. I think everyone except Nikon has figure that out, otherwise, why would they uses CMOS for these,
Sony HDR-FX7
Canon VIXIA HF S100
Panasonic HDC-TM300
@Tin
That list is a bit meager. The FX7 from Sony has epic rolling shutter problems.
Although modern generations show less rolling shutter.
My 9k Sony PMW-EX3 is a CMOS camera, it's also CineAlta and pretty professional.
It has rolling shutter.
My Canon XH-A1 camera (3k), is a CCD camera, it does not have rolling shutter but is HDV.
My Canon 5D MII camera, is a CMOS camera, it is a DSLR, and has more rolling shutter then my Sony
Its just less noticeable on the EX3, in normal usage you wouldn't see it, but when you for example film a car from close by which has vertical lines painted on the side, and you would stop the car, it might only cover 3 frames, you will see that the lines are slightly skewed.
This wouldn't happen on a CCD sensor. They might be blurred, but not skewed.
In other words, CMOS will always be scanning line per line, which will always cause those problems. Its just the quality of the CMOS and the hardware processing the data which defines how visible and annoying it is.