1 sensor vs. 3 sensors

R

rtmst3k

Member
Excuse the novice question, but I was curious about the 3-CCD vs 1-CCD (or 1 sensor vs. 3 sensors) comparison. Canon says that their HF11 has a single image sensor that has 3.3 effective megapixels. I think the HMC-150 has
3 1.1 megapixel sensors. Does that mean they should produce the same quality image ? Or is there something more to having multiple sensors?

I know the subsequent compression method would make a difference on what
gets recorded to the SDHC card, but I was curious about the "native" capability of the sensor front-end.

Again, apologies if this is a dumb question...:)

Thanks,
Randy
 
The HMC is in a different league to the Canon. A fair comparison would be with something like the SD100. I would expect that they would be similar, but maybe the Canon would have shallower DoF.
 
Single sensors have bayer filters which only allow 1/4 of the pixels to be striped with a blue filter...so they are extremely noisy in tungsten light, not to mention fitting that many pixels on a small sensr causes a lot of noise, because it's more difficult to supply them with ample signal to overcome the inherent noise with smaller cams
 
Cut and pasted from this Wikipedia article:

"Digital color cameras generally use a Bayer mask over the CCD. Each square of four pixels has one filtered red, one blue, and two green (the human eye is more sensitive to green than either red or blue). The result of this is that luminance information is collected at every pixel, but the color resolution is lower than the luminance resolution.
Better color separation can be reached by three-CCD devices (3CCD) and a dichroic beam splitter prism, that splits the image into red, green and blue components. Each of the three CCDs is arranged to respond to a particular color. Some semi-professional digital video camcorders (and most professionals) use this technique. Another advantage of 3CCD over a Bayer mask device is higher quantum efficiency (and therefore higher light sensitivity for a given aperture size). This is because in a 3CCD device most of the light entering the aperture is captured by a sensor, while a Bayer mask absorbs a high proportion (about 2/3) of the light falling on each CCD pixel.
Since a very-high-resolution CCD chip is very expensive (as of 2005[update]), a 3CCD high-resolution still camera would be beyond the price range even of many professional photographers. There are some high-end still cameras that use a rotating color filter to achieve both color-fidelity and high-resolution. These multi-shot cameras are rare and can only photograph objects that are not moving.

[edit] Sensor sizes

Sensors (CCD / CMOS) are often referred to with an imperial fraction designation such as 1/1.8" or 2/3", this measurement actually originates back in the 1950s and the time of Vidicon tubes. Compact digital cameras and Digicams typically have much smaller sensors than a Digital SLR and are thus less sensitive to light and inherently more prone to noise. Some examples of the CCDs found in modern cameras can be found in this table in a Digital Photography Review article"



I think it was written with still and not video cameras in mind, but a CCD is a CCD so the same applies to both. Long story short... 3 is always better than 1. One thing to remember, 1-chip CCD's with 3.3 megapixels are going to have a lot of really small pixels, whereas 3 1.1 megapixel CCD's of the same size are going to have the same total number of pixels, but those pixels will be a lot larger, as they're spread out of 3 instead of 1 chips, and should then be able to capture more light.

One other thing... the HF11 uses CMOS chips, not CCD's. I don't plan or want to start a CMOS vs CCD debate, but you might want to take a look at the pros and cons of both.
 
Yes, they look different, at different times and lighting conditions. Not hugely, but it's different.

Isn't the Red a single CMOS chip?
 
As is the SI2k used to shoot Slumdog, and the Viper Filmstream used to shoot Benjamin Button, And the Genesis used on 21 and Apocolypto...

Number of sensors is immaterial. Quality of sensors is very important.
 
Yes, I think still photographers learned that some time ago with the Foveon chip. A single chip outperforming many other 3 chip cameras at the time (5+ years ago).

Pretty sure the massively impressive VR Phamtom cameras are single chip as well.
 
Mike Harvey said:
I think it was written with still and not video cameras in mind, but a CCD is a CCD so the same applies to both. Long story short... 3 is always better than 1. One thing to remember, 1-chip CCD's with 3.3 megapixels are going to have a lot of really small pixels, whereas 3 1.1 megapixel CCD's of the same size are going to have the same total number of pixels, but those pixels will be a lot larger, as they're spread out of 3 instead of 1 chips, and should then be able to capture more light..
Click to expand...


Well, 3 1.1 megapixel CCD's don't necessarily create a 3.3 megapixel end image...they just allow you to have 4:4:4 images, TRUE 4:4:4 is not actually possible with single chip cams due to the bayer filtering
The single 3.3 megapixel chip may look sharper in the end though, but much more noisy..

now a 3 ccd cam with 4k CCD's would be awesome.
 
DanHD said:
Well, 3 1.1 megapixel CCD's don't necessarily create a 3.3 megapixel end image...they just allow you to have 4:4:4 images,
Click to expand...

HUH?


DanHD said:
TRUE 4:4:4 is not actually possible with single chip cams due to the bayer filtering
Click to expand...

Absolutely wrong. Tell someone using a Panasonic Genesis they are not getting true 4:4:4 images on a half million dollar camera and see what they say to you...
 
PerroneFord said:
HUH?




Absolutely wrong. Tell someone using a Panasonic Genesis they are not getting true 4:4:4 images on a half million dollar camera and see what they say to you...
Click to expand...

You mean Panavision? That has a 12Mp sensor so it could achieve 1080P 4:4:4 by downscaling the image. I think the other poster was trying to say that a 1080P native single chip camera would not give true 4:4:4 due to the Bayer pattern. So, in a way, you are both correct.
 
studio1972 said:
You mean Panavision?
Click to expand...

Sorry, yes, I did mean Panavision. Been reading Panasonic too long!

studio1972 said:
That has a 12Mp sensor so it could achieve 1080P 4:4:4 by downscaling the image. I think the other poster was trying to say that a 1080P native single chip camera would not give true 4:4:4 due to the Bayer pattern. So, in a way, you are both correct.
Click to expand...

Guess that depends on what you mean by "True". If deBayering doesn't fall under "True" then yes, he's right. But that doesn't stop the manufacturers of the Dalsa, RED, and others using a Bayer pattern from claiming 4:4:4 and it doesn't stop cinematographers from using them like true 4:4:4 for VFX and other work.
 
You misunderstood me Perrone.

1 1080p single chip vs 3 1080 chips would suck. It would not produce true 4:4:4.

and how the hell does 3 1 magapixel chips equal true 3 MP resolution in the image.

The hvx200 doesnt do it, even with pixel shifting.

What is debayering and how is it done? Is that why these single chips have so many pixels to start with but only end with 1080p?
 
PerroneFord said:
Guess that depends on what you mean by "True". If deBayering doesn't fall under "True" then yes, he's right. But that doesn't stop the manufacturers of the Dalsa, RED, and others using a Bayer pattern from claiming 4:4:4 and it doesn't stop cinematographers from using them like true 4:4:4 for VFX and other work.
Click to expand...
Red doesn't claim 4:4:4 sampling on their images. Graemle says that it's more like 1.6:1:1. A Red One doesn't make a 4K 4:4:4 image, the closest you can come to 4:4:4 would be by shooting full frame and downrezzing to 2K.

Same with the Dalsa Origin.

You cannot get 4:4:4 out of a one-chip system, at sensor resolution. No camera could do that.

Red claims 4:4:4 in its RECORDING format, yes, because they don't downsample the chroma as part of recording. But the imaging head is not capable of delivering 4:4:4 at 4K, nor do they claim it will.

The Genesis uses a 3x oversampled chip to get a true 4:4:4 recording. But it does so using a 5760-wide RGB stripe sensor, and downconverting to 1920 wide. If they were allowing you to record the full raw 5760-pixel-wide image from the sensor, it would in no way be 4:4:4, nor would they claim it was.

It's like if you took a DVX image and exported it as a .BMP file - the BMP file is a 4:4:4 recording system, but the DVX image doesn't deliver 4:4:4, it's 4:1:1. So you'd have 4:1:1 data carried in a 4:4:4 recording media. That's the same kind of thing RedRaw does; it carries 3.2:2:2 data inside a 4:4:4 recording format.

If you want 4:4:4 at sensor resolution, you have no choice but to use three chips. Or a Foveon, but even then a Foveon is basically three chips, just sandwiched vertically instead of split by a prism. If you have a single-chip system, you cannot get 4:4:4 out of it unless you downconvert to no more than 1/3 the chip res horizontal (using RGB Stripe) or 1/4 total (1/2 x 1/2) using Bayer.
 
DanHD said:
And Bayer patterns are terrible for tungsten.
Click to expand...
Way oversimplified.

Bayer + a daylight-balanced chip + CMOS = poor tungsten performance.

Bayer + daylight-balanced chip + CCD = okay for tungsten.

Bayer + tungsten-balanced chip + CMOS or CCD = fine tungsten performance.
 
Excuse the oversimplification.

Most consumer and prosumer cams with bayer filters dont perform well in tungsten light because of the lack of pixels sensitive to blue. Even when working with the Red I always ALWAYS use a 5000k balance.

Get your facts right Perrone Red doesn't claim 4:4:4
 
I was just saying that because you say things that sometimes are just wrong and confuse people.
 
You must log in or register to reply here.
Back
Top