Incredible low-light video footage from Canon

Started Mar 4, 2013 | Discussions
padang
Senior MemberPosts: 1,063
Like?
Incredible low-light video footage from Canon
Mar 4, 2013

Canon getting back to 1st place in sensor design ?

http://www.canon.com/news/2013/mar04e.html

Photato
Senior MemberPosts: 1,810Gear list
Like?
Re: Incredible low-light video footage from Canon
In reply to padang, Mar 4, 2013

Apparently it's a 2.1 MP sensor.

Wonder why?

 Photato's gear list:Photato's gear list
Canon IXUS 310 HS Canon EOS M Apple iPad Mini +1 more
Reply   Reply with quote   Complain
Felix11
Regular MemberPosts: 113Gear list
Like?
Re: Incredible low-light video footage from Canon
In reply to padang, Mar 4, 2013

Canon says:

"In addition, the sensor's pixels and readout circuitry employ new technologies that reduce noise, which tends to increase as pixel size increases"

Which is interesting because people generally think that noise is greater with smaller sensor elements.

Are Canon marketing just being clever here by using noise when we will read what they mean in terms of signal to noise ratio?

SNR must be very high on this sensor, nonetheless the noise will also be slightly higher due to increased sensor element area or border - is that correct?

 Felix11's gear list:Felix11's gear list
Ricoh GR
Reply   Reply with quote   Complain
Sonyshine
Senior MemberPosts: 5,690Gear list
Like?
Re: Incredible low-light video footage from Canon
In reply to padang, Mar 4, 2013

He he! Its just a low light astro sensor!

 Sonyshine's gear list:Sonyshine's gear list
Nikon AF-S DX Micro Nikkor 40mm F2.8 Sony Cyber-shot DSC-QX10 Nikon 1 V2 Tamron SP 70-300mm F/4-5.6 Di VC USD Nikon 1 Nikkor VR 30-110mm f/3.8-5.6 +6 more
Reply   Reply with quote   Complain
Eric Fossum
Contributing MemberPosts: 853
Like?
Re: Incredible low-light video footage from Canon
In reply to Felix11, Mar 4, 2013

the noise will also be slightly higher due to increased sensor element area or border - is that correct?

Not necessarily, and probably not in this case.

-- hide signature --

*Have fun, make some money, and learn along the way*

Reply   Reply with quote   Complain
onedawg
Forum MemberPosts: 57
Like?
Re: Incredible low-light video footage from Canon
In reply to Photato, Mar 4, 2013

1920 x 1080 = 2.074M

Who cares if it's only 2.1 megapixels if you're going to use it for video? I think it's kind of exciting although it does exclude future 4K video.

Reply   Reply with quote   Complain
JJ Rodin
Contributing MemberPosts: 790
Like?
Re: Incredible low-light video footage from Canon
In reply to Sonyshine, Mar 4, 2013

This sensor will revolutionize the video camera world if it performs well in all other aspects!!

NO, astro photography as an example was used to emphasize its low light capability, and the ultimate low light test is astro photography, (Them stars are not so bright huh?)

Have you ever made a 10, 20, 30 min exposure of the sky, your sensor will likely start banding, heating up and produce an awful mess of a picture, this however would do a fabulous job, just not for a still image.

Them big pixels will nearly ALWAYS win the contest of better capture, if all things are equal!

Canon 1 Sony 0,    Sony, your turn!!

Sonyshine wrote:

He he! Its just a low light astro sensor!

Reply   Reply with quote   Complain
jaygeephoto
Junior MemberPosts: 28
Like?
Re: Incredible low-light video footage from Canon
In reply to padang, Mar 5, 2013

Demo video is intriguing but dull.  Whenever I think of great low light photography I cannot forget the candle lit scenes from Stanley Kubrick's Barry Lyndon, using a NASA spec. 50mm 0.7 Zeiss lens.

Now that's some stunning low light photography - still gives me goosebumps.

http://www.youtube.com/watch?v=hJahPDkvivc

Maybe we'll see this sensor in RED cameras: http://www.red.com/

Reply   Reply with quote   Complain
Cy Cheze
Senior MemberPosts: 2,028
Like?
Applications of Large Pixel CCD for Video
In reply to padang, Mar 5, 2013

Existing security cameras delivery very poor resolution in low light.  The Canon sensor might greatly improve results.  However, the best deterent effect offered by any security camera is to make the camera's presence very obvious.  A vandal-proof metal box with a blinking red light might suffice, as well as be much cheaper than the exotic new sensor.

However, the problem with extreme low light is not just the low light, but also the intrusion of glaring hot spots, whether a full moon or a lamp.  The challenge to dynamic range defies even the human eye.

"Color balance" and "white balance" are also somewhat oxymoron at night too.  Not much cyan to squeeze from moonlight or sodiuim vapor lamps.

A sensor optimized for nocturnal security cameras might be difficult to integrate into a device also intended for daytime use.  Owls, for good reason, operate mainly at night; while eagles hunt by day.  Some bats dispense with vision altogether and intercept insects at night using sound alone.  Infra-red imaging is smudgy, but important to reckon people or beasts hidden not only by dark but also by objects or vegetation.

Reply   Reply with quote   Complain
pgb
pgb
Senior MemberPosts: 1,264Gear list
Like?
Re: Incredible low-light video footage from Canon
In reply to Felix11, Mar 5, 2013

Canon says:

"In addition, the sensor's pixels and readout circuitry employ new technologies that reduce noise, which tends to increase as pixel size increases"

Which is interesting because people generally think that noise is greater with smaller sensor elements.

Are Canon marketing just being clever here by using noise when we will read what they mean in terms of signal to noise ratio?

SNR must be very high on this sensor, nonetheless the noise will also be slightly higher due to increased sensor element area or border - is that correct?

Not according to canon. If smaller pixels are cleaner why wouldn't
They make a 50mpx dalr now?

Reply   Reply with quote   Complain
JJ Rodin
Contributing MemberPosts: 790
Like?
Re: Incredible low-light video footage from Canon
In reply to Felix11, Mar 5, 2013

For someone who knows a little about silicon & chips, the more silicon area you have the more potential for noise, more=more. It is almost that simple.

Silicon is a conductive material (ie electrons are easily freed and produce an electrical signal), and extraneous EMI/EMF (all the electrical crap flying around in the air - cell phones, sun, broadcast towers, earth, etc) will hit silicon and free up electrons (noise), unless very well shielded, and consumer chips are NOT that well shielded (why do you think NASA/space electronics are so expensive, MASSIVE shielding, not only for noise but to NOT get fried by gamma rays and such !!).

Whereas when using the minscule sized pixels (photo diodes) in 18mp 1/2.3" sensors, the noise can be greater than the signal (signal to noise ratio).

IOW, the electrical energy converted due to the desired photon produced electricity is not much more or even less than the random noise in the silicon layers caused by many factors.

That is why Sony has to 'watercolor' their 18mp 1/2.3" sensors because it is a noisy mess, even 12mp Canon sensors have lots of noise, you do have to use NR to compensate for sharpening quite a bit !

It is always about signal to noise, look at knowledge, how much 'crap' do you know that you would just a well like to flush away! But you can't!

Well signal is what we want, until some new unknown physics/material/tech occurs, NOISE will have to be dealt with, so just ask her to 'be quiet' more often!

Felix11 wrote:

Canon says:

"In addition, the sensor's pixels and readout circuitry employ new technologies that reduce noise, which tends to increase as pixel size increases"

Which is interesting because people generally think that noise is greater with smaller sensor elements.

Are Canon marketing just being clever here by using noise when we will read what they mean in terms of signal to noise ratio?

SNR must be very high on this sensor, nonetheless the noise will also be slightly higher due to increased sensor element area or border - is that correct?

Reply   Reply with quote   Complain
Eric Fossum
Contributing MemberPosts: 853
Like?
Re: Incredible low-light video footage from Canon
In reply to JJ Rodin, Mar 5, 2013

I am sorry but this is just incorrect.

-- hide signature --

*Have fun, make some money, and learn along the way*

Reply   Reply with quote   Complain
Cy Cheze
Senior MemberPosts: 2,028
Like?
A Sure-Fire Way to Cut "Noise" Over Pixel Size Debate
In reply to Eric Fossum, Mar 7, 2013

Eric Fossum wrote:

I am sorry but this is just incorrect.

This is a perennial hot argument that inspires all sorts of conjecture and hypotheses, whose tests are either nonexistent or opaque.

Are there any "fair" side-by-side comparisons of JPEGs shot, at various ISO levels, with devices which are identical in every respect, except for pixel count or size?  Presumably, the differences, if any would be greatest at ISO over 400 in the case of a 1/2.3" sensor, or perhaps smaller.

Or is any comparison difficult because cameras are all designed to cut "noise"?

Several Canon videocams employ 2MP sensors which do quite well in low light.  Some presume this is because fewer pixels mean less noise at low light, but you infer that is not the case.

What of the elimination of low-pass filters on some new cameras?  Was the "received wisdom" about the filter's role to cut aliasing or moiré unfounded?

My myopic, peon, worm's-eye observation is only that most low ISO pictures, wether shot with a 4mp CCD camera in 2003, or a 25mp CMOS in 2013, look the same to me.  Friends see no difference between a phone-camera shot and other images anyway.  To recognize the face or pet is enough.

How a 12,5000 ISO shot looks when zoomed to pixel level seems like judging a car by how it "handles" in city traffic at 100mph: in practice, unknown movements of the traffic, and driver error, offset any marginal improvement in agility.   If and when one actually "needs" ultra ISO, something else is apt to flaw the result: light balance, flare, blow-outs, or sheer lack of viewer appeal.

With due respect for expert opinion, I vaguely suspect that adding pixels or high ISO is driven more by marketing than by utility.  See-in-the dark cameras or video for security cameras or nocturnal wildlife tracking is quite interesting, or maybe the intrepid noight owl paparazzi, but not for the usual commercial photographer.  Meanwhile, plenty of myopic peons will be curious to view the output of such devices with their worm-eyes.

Reply   Reply with quote   Complain
Keyboard shortcuts:
FForum MMy threads