Canon Rebel T2i / 550D demonstrates 1080p video mode, collects glowing early preview
By Vlad Savov 
posted February 25th 2010 6:13AM
posted February 25th 2010 6:13AM
posted February 25th 2010 6:13AM
A look back on popular stories from today in a specific year.
Now that we've thrown 'em off the trail, use the form below to get in touch with the people at Engadget. Please fill in all of the required fields because they're required.
Too much megapixels.
@Mikkis74
That can be minimised if needed. You obviously dont know much!
@kuruptedbhoy Well no, the sensor pixel density/size cannot be changed. I would be more than happy to buy an DSLR with 12MP as I believe that's enough for an APS-C sensor. That said, I believe that compacts should have no more than 6MP.
@SlimSpaceman
but it's not a compact it's a dslr...
@SlimSpaceman
I believe that todays 550D APS-C is not the same APS-C made 2,3 or 4 years ago. They have been optimised over the generations. todays 550d APS-C is as close as you can get to the 7D's
@Mikkis74
Maybe its just the post processing, but I'm not entirely happy with the way this camera handled color in the video. It either seemed slightly washed out, or a bit dodged. I really want this camera, and perhaps the color can be corrected, the depth of field control is incredible. I would have liked to see more high motion, or something of the like. this video was really well done, definitely a huge step forward in camera technology for indie cinematographers
@xconan Actually I prefer my DSLR to have about 6 MP, max 10. I want low light performance that is so good that it has no problem with photos where my eyes can hardly see anything. You won't get that with high MP cameras. There is a reason why the full frame D3s has 12 MP. This has a much smaller sensor and 18 MP.
Want want wantttt
550D is a great camera. noise control is great for its class. makes me wanna buy one, 2 bodies, 5D + 550D =)
@ahchar
"makes me wanna buy one"
This, ladies and gentlemen, is what's known as a 'marketing win'.
If it is done right why is it to much, yes it is not about the pixels only but maybe somebody does want that wall size portrait of themselves on the cheap.
@TheOne, you do not need the 18MP for a wall size print. Unless of course you going to look at the print with a loupe...
@TheOne Too much pixel on a smaller sensor it not a great thing.
You definitely need an Austrian branch! I as an Austrian reader would appreciate that!
Anyway, I am looking forward to replace my 450d with this new 550d, there is really not much difference to the beloved 7d.
@fdomig de.engadget.com should serve all German speaking countries... what would be the advantage of making a dedicated Austrian site?
@hq
It would satisfy our national pride! hehehe
After living in the UK for too long those pictures really made me homesick. Vienna is such a gem.
This camera value is just incredible !!! I have a 7D, and i would choose it over the 550D but this new Canon gives a lot more for the money !!!
What about an AUSTRALIAN Engadget, peoples?!
...
...
I know we're boring, but still...
@Wheee Engadget will probably be blocked in the Australian Great Internet Reef anyway, I've seen several models holding gadgets with small boobies, so it's obviously a sinful website!
@Wheee
What Engadget Really, Really, Really, needs is A Camera and Camcorder section. I'll jump on the band wagon if they decide to.
yeah yeah very nice... now where the frig is my 5D Mk2 60fps firmware update!!?
Wait so 30 fps is required now? After all the moaning about lack of 24 fps in the 5D? Haha
@Yemble: Think they were talking about how before the T1i only had 20fps. Now the T2i can go all the way up to 30fps but it can still do 24.
Sure, come here. You can set up your Austrian HQ right here in my room!
Yes 24/25p is the requisite for cinematography, don't blame the poor journalist :) for him higher numbers are better.
higher pixel densities aren't a bad thing, unless one makes the naive decision to compare at 100% where output may look slightly noisier and softer per-pixel. Once resized to equivalent magnifications though, noise output from like sensor technologies at different pixel densities often favour the higher pixel density. Obviously adequate engineering (in both the software and hardware) is required to take advantage of pixel density but the 5D mk II accomplished it quite well, as does the 7D and obviously the D3x. It does, however mean one needs a higher resolving lens to truly take advantage of it, and most budget-conscious zooms barely approach the nyquist frequency of previous generation sensors, so a lot of those extra pixels go to waste.
@DanKennedy
Higher pixel density has a major drawback from some people- more noise at higher ISO. That's why the D3S performs slightly better then the 1D MIV at those higher ISOs. It's why many manufacturers are going back from 15 to 12MP with the point and shoots.
@EM1
You simply did not read/understand my post.
Show me an equivalent magnification (ie. not 100% per-pixel on both cameras) image from two concurrent generation cameras at identical ISO in identical conditions and I'll believe you, otherwise the science and evidence is simply not in your favour.
With current sensor technology there is negligible space between photosites, this is thanks to intelligent layout of electronics on the silicon and gapless microlenses.
By all means if you have evidence or a rational explanation I will accept it. Otherwise, the megapixel myth has been overcome by technology in more serious cameras. It does, however, still apply to the very cheapest point & shoots as they have limited software/hardware enhancement to deal with increasing pixel densities.
And in addition you obviously aren't very connected to the industry, I am.
Latest generation Olympus, Canon and Nikon compacts all sport a new 14.2 megapixel sensor, most likely of Sony origin, so much for scaling back to 12mp hey?
The G11 was different, because Canon switched to a higher sensitivity sensor - Sony ICX685CQZ is in the G11. This sensor has an optimised component layout and readout configuration for high-iso shooting. Added to the gaining popularity of 'experts' toting the megapixel myth, many people including this website especially always think they're onto some 'inside information' because they know better than the manufacturers what gets a better picture and oh more pixels is always more noise... Rubbish, spend a while shooting with numerous cameras of different resolutions at the same sensor size from the same sensor manufacturer and you quickly descover that you get less/same noise at the same print size from the higher density sensor, with the added benefit of increased resolution especially applicable to low-iso shooting.
Added to the fact that sites like dpreview always compare at a per-pixel level which means little to the resulting shots at equivalent magnifications. Faulty testing methodology is what makes armchair experts like you able to spread information with assumed authority.
@DanKennedy
Armchair? I wouldn't make assumptions like that. I've tested many a camera, cameras up and including the H3D and Phase One P65.
Of friggin course comparing two cameras with the same sensor is going to yield simmilar if not matching results.
So the testing DxO mark does is BS? Or this site that uses science: http://theory.uchicago.edu/~emj/pix/20d/tests/noise/noise-p3.html
Please backup your segment with fact and proof. If we're all just sitting in our armchairs sipping on ignorance, why don't you enlighten a few of us with links and proof? Cause if I am wrong, I want to see proof. Your arogance on a blog does nothing for your credibility right now.
@EM1
Yes, actually DxO mark is using incorrect testing methodology. On a per-pixel level. MFDB's and yet you still don't understand sensor technology? How embarrassing. My arrogance comes from being fed up because others can't see the wood for the trees, if you consider that heat production is no longer a factor (especially with CMOS sensors) for most photographers and the amount of photons being captured doesn't vary, surely you can imagine that a higher pixel density doesn't increase the noise in an equal magnification print?
I didn't say same sensors, I said same sensor technology, please read my post carefully as it is fundamental to your understanding.
A big factor for noise however is the CFA and the necessary amplification to gather sufficient light through a highly effective one. As a P65 user you should know that MFDB's don't have outstanding high-iso because their priority is colour reproduction in a studio/controlled situation or tripod-bound. I'm sure some googling can explain this to you.
Read noise CAN affect noise, and this is why we have slight amount more noise at higher pixel densities of the same technology... WHEN VIEWED AT 100%. But we have so many more pixels from the higher density sensor, that when we downsample to the resolution of the lower density sensor, we have effectively less noise. I honestly don't understand why you cannot comprehend this and why you need DxOmark and other 'authorities' to determine such things for you (incorrectly as it were).
http://www.fredmiranda.com/forum/topic/684759/0
Here is a real-world example which shows the D700 has slightly worse noise at equivalent magnification than a 1DsIII, and the D700 is considered more recent sensor tech and a superior low light camera.
Read and learn my friend. Maybe you can understand those 10's of thousands of dollars of digital equipment you say you use.
@DanKennedy
Intelligence coupled with arrogance usually consists of the latter.
@newtype
Fair enough, I overreacted because I knew the backlash was coming, and it was late at night on a 42C day and I was tired.
However, your opinion of me doesn't change the fact that I am right, the science and evidence supports that. I don't see you providing any useful information (checked your post history) beyond an opinion.
@DanKennedy
Now just imagine what sensors without the pixel rush could do. What do you want to do; cover your wall in a larger than life size print of yourself and count the pores on your nose from an inch's distance?
@Cistron
Wow big sigh, you didn't read/understand my post eh?
Technically, I want to shoot landscape, promotional pictures for artists and portraits, as this is my profession. If you don't need resolution good for you, just dial the setting down in the camera, I promise it'll still be better than any 6 megapixel camera you've used in the past.
I need to print large up to about 32x48" and sometimes larger. I don't want to stitch as I it's not practical in all situations and for this I need high resolution sensors and the best optics available. I would use an 8x10 view camera but sheet film is horrendously expensive and the size isn't practical for me. I use 18mp and 21mp 135 format digital cameras, and see nothing wrong with them going to higher resolutions still.
@DanKennedy
I am confused about a few things then if you don't mind explaining them without being rude. I read the link which explains a bit more thuroughly what you said which now makes sense. Cause it is all about the final output that all this needs to be judged.
My question is then, why the higher resolutions? If the 18MP camera performs equally as well as the 12MP when it's downrezed, printed, etc, why keep cramming more photons into the sensor? It just seems like the more expensive route from the manufactures standpoint. Or am I still wrong or missing something here?
@DanKennedy
sorry, meant photosite not photons.
@newtype
Something coupled with anything always consists of the latter. And the former.
@EM1
The only thing you're missing is the point of low-iso performance. A sensor that has a stupidly high resolution should still be able to provide a very clean low-iso image. So if you have adequate light you get a resulting file you can print quite a bit bigger without a loss in quality or noticeable softening. There is some smart resizing software out there (like genuinefractals) but it's still not as good as having the information there to begin with.
Of course this means one needs higher resolution lenses, which is why we see consistent refreshes of ranges that have been covered before, admittedly they also add other technological enhancements at the same time.
It's very hard to see this in practice until you take the same picture with the two cameras and print it to the same reasonable size, then the noise benefits or (if you're printing REALLY big) the resolution benefits, become apparent.
Sorry for being rude before.
@DanKennedy
when did camera manufacturers like Canon start using microlenses? I thought those were only for niche cameras/ backs like the P30 in order to create a sharper image and help with the moire pattern with fabrics and such. I was under the impression that the use of microlenses also meant no tilts or shifts can be used because light would hit them at such a weird angle the image would be distorted. With so many people in love with the T/S lenses these days, I don't understand. Or is or should I say was because the P30 back didn't have "gapless" microlenses, and that's what created distortion and the move to rid the gap corrected the issue of light falling into the photosite improperly.
@EM1
Couple of tricky questions there :p Microlenses are so close to the pixel surface that unless they are being placed far off-centre (like in the Leica M8/M9) to counter an extreme angle of incidence they aren't going to cause any issues with any lenses mountable on an SLR.
With the Leica M9, because there is no mirror, the lens sits extremely close to the sensor, this proximity of the exit pupil, means that when using a wide angle lens, the light strikes the sensor at a very extreme angle. If the microlens was directly over the pixel a lot of light would be scattered/reflected, by shifting the microlens closer to the centre of the sensor it can correctly focus light into that pixel.
Canon tend to use a uniform microlens array, I haven't heard of them using shifted outer microlenses as Kodak do in the M9 sensor but they might, their primary function is to gather light and direct them into the pixel. Canon have used microlenses since the D30 (not 30D) in 2000.
I'm not sure if microlenses will affect moire, as moire is an inherent problem in the grid-like nature of sensors when resolution in patterns approaches the nyquist frequency (the resolution of the sensor itself). Incidentally, Canon and mass-market DSLR manufacturers use a special filter called an anti-aliasing (AA) filter over the sensor. What this sensor does is blur any detail that reaches the nyquist frequency, preventing moire.
A lot of digital backs (like the P30) and the Leica M8/M9 don't use an anti-aliasing filter and sometimes suffer from moire in extreme circumstances, but it can be quite well corrected in CaptureOne or third party software. I haven't ever had a problem with it.
Another advantage of higher resolution sensors is the nyquist frequency goes up! meaning that it is harder and harder to encounter moire from regular objects like grills and fences, and instead it's only a problem on the finest fabrics, even then it's very well controlled by in-camera processing these days.
T/S lenses should perform just fine regardless of microlenses etc. Even the 17mm TS-E doesn't create an extreme enough angle of incidence on the sensor to create smudging or falloff. SLR wide angles tend to be retrofocus, meaning they aren't symmetrical and take light from an extreme wide angle and spit it out at a less wide angle. With rangefinders like the M9, many wide angles are symmetrical so the angle of light exiting the lens is as extreme as the light entering, this leads to better image quality, but isn't possible with a mirror in front of the sensor. Google the Zeiss ZM 25mm Biogon for an example.
My knowledge of the P30 is a bit limited, but I'd think it's issues were moire/processing related or perhaps an issue with improperly centred microlenses? I'd have to look into it to find out the real cause.
@EM1
Oh, I should add one more thing. The reason why we haven't seen many extremely wide angle lenses for micro 4/3 is because their sensors don't have shifted microlenses, meaning the light can't come from an extreme angle like it can onto the M9 sensor. This means that all their lenses have to be telecentric, meaning the light strikes the sensor almost at a perpendicular angle, in order to maintain sharpness and illumination. This makes lenses harder to design and typically bigger.
@DanKennedy
Everything I assumed about cameras got owned by this comment series.
I want one, I want one!!!
If this camera had an LCD control panel, 100% frame coverage on the view finder and LCD, inbult anti-shake and Mpeg4 AVCHD, it would be a killer machine!!!
"The last thing that I would do is lie to you. If everybody was in on it, I'
d have to be in on it too. I'm not in on it, Truman, because there is no 'it' "
Thats all I think when watching this soundtrack, oh wait, I mean video.
Unless they've fixed the autofocus problems, the video mode in a DSLR is still nothing but a gimmick.
@Yogi
That's true, all the serious large sensor cinema cameras use autofocus lenses, like...... um........ um...... oh wait they all use manual focus and/or anamorphic lenses. It's ok, they still have fast autofocus in video mode on the Panasonic GH-1 for professionals like you, so us amateurs can use the gimmicky cameras :)
Great lenses like the Zeiss ZE range or other adapted manual focus lenses open a world of possibilities to cinematographers longing for an affordable large sensor digital video camera.
@Yogi
You actually bring up a valid point. Photographically, these are priced and aimed at the consumer, but the video mode is going to be very difficult for an amateur to get much out of. I just don't think they do well for anything but carefully planned shots--but they do incredibly well for that.
@Yogi
I don't know what you consider to be a problem, but the T2i does autofocus in movie mode.
@DanKennedy
I am happy for you that you are terrific at manual focus. However, as Chispito pointed out, these cameras are priced and aimed toward the consumer who may not be quite as amazing as you with manual focus. They are marketed toward end users who are stepping up from a P&S digital camera and they just want to be able to take video. For the targeted user, video without a good autofocus is truly a gimmick.
@Jack
I never said that it doesn't do autofocus. The problem is that, on DSLRs that have autofocus in video mode, it's not very good.