[redpigeons]

DSLR "Video" is the future and the present there is no doubt about it.
you need to remember that digital video is a relatively new format.

carlos Baena is one of the most important animators in the world today and head of short film department at Pixar studios .

he is shooting primarily with DSLRs

CarlosBaena.com

[Urkeldaedalus]

DSLR "Video" is the future and the present there is no doubt about it.
you need to remember that digital video is a relatively new format.

carlos Baena is one of the most important animators in the world today and head of short film department at Pixar studios .

he is shooting primarily with DSLRs

CarlosBaena.com

If this is true I am hopeful camera companies will offer another option -- digital HD camcorders with interchangeable lenses. Since digital HD cameras can already take stills, maybe we're heading this direction from both sides.

[redpigeons]

If this is true I am hopeful camera companies will offer another option -- digital HD camcorders with interchangeable lenses. Since digital HD cameras can already take stills, maybe we're heading this direction from both sides.

the Red Camera is exactly this and its the biggest thing in the video world right now

[mutedphotos]

With only 1/7th of pixels read out (14.6MP), an APS-C sensor effectively becomes

a 1/2.3" sensor

in video mode (not for DoF -- but for ISO performance)! This is small P&S size.

The noise of a P&S sensor isn't simply because of the sensor's size but has to do with the size of the pixels. This is a physical aspect of the sensor and changing the amount of pixels recorded should have no effect on the noise aspect of the sensor.

Also, if they're only recording certain pixels, they wouldn't just record the center of the chip as it would affect the field of view.

[Philippos]

Where is your 15.1 MP spec coming from?

The Samsung sensor in the K20D has a total of 15.1 MP, however it uses 14.6 MP of them (referring to them as "active") to form the image.


I have no idea on what the final capturing technique will be. I agree that the entire surface of the sensor should (and I'm sure it will) be used - otherwise it's meaningless.
If someone knows how the burst mode in K20d works, I believe we will have our answer.

As far as sound, I don't really care if it's mono or stereo. You can get a decent Stereo Field Recorder pretty cheap (Zoom makes a 4 channel one for less than $200) and you can get better sound than any pro-level video camera.
This however brings another quastion to mind:
Will the K-7 produce SMPTE time-based video output?

[Gimbal]

How does the k20 21fps mode work? Subsampling? Probably, but then again noise in video isn’t nearly as critical as noise in a still picture since you get an averaging affect by the shear amount of pictures thrown at you. Supersampling over time instead of area, sort of..

Edit, Philippos raised the k20 question while I typed, ops.

[tomtor]

The noise of a P&S sensor isn't simply because of the sensor's size but has to do with the size of the pixels. This is a physical aspect of the sensor and changing the amount of pixels recorded should have no effect on the noise aspect of the sensor.

Ofcourse it has an effect on noise and DR. If you use only a part of the sensor pixels (even if they are distributed over the whole sensor) you're effectively using a smaller sensor. If you combine several pixels from the larger photo sensor as input for a single video pixel you're lowering noise.

[nixcamic]

Even sub-sampling will still give better noise than your average camcorder, just due to the bigger pixels and higher quality sensor. But who says it has to be one or another, what if they sub-sample half the pixels, then super sample those 7MP down to 2? Less noise with way more manageable bandwidth.

[mutedphotos]

Ofcourse it has an effect on noise and DR. If you use only a part of the sensor pixels (even if they are distributed over the whole sensor) you're effectively using a smaller sensor. If you combine several pixels from the larger photo sensor as input for a single video pixel you're lowering noise.

I'm not an expert on sensors so I may be completely wrong here, but my understanding is that the S/N ratio has nothing to do with the size of the sensor itself, only how big the pixels on that sensor are. A 6mp sensor of any size will have a better S/N ratio than a 12mp sensor of the same size, because the 12mp sensor has to use smaller pixels to fit them all in.

IF my understand there is correct, then it shouldn't matter how many pixels on the sensor you actually use. The size of those pixels remains the same, and thus the S/N ratio should remain the same.

Seems logical at least, but I may be wrong. And I don't know how this effects DR, just thinking in terms of noise.

[Gimbal]

Actually it’s the other way around, sensor size is what matters, not pixel size (well pixel size also matters, but not as much). That is, if you look at the result at the picture level. If you judge at pixel level, pixel size will take a bigger role.

[mutedphotos]

Actually it’s the other way around, sensor size is what matters, not pixel size (well pixel size also matters, but not as much). That is, if you look at the result at the picture level. If you judge at pixel level, pixel size will take a bigger role.

I'm not following. How is that?

[Gimbal]

Have to get some sleep now, but check out http://www.dxomark.com/index.php/eng/Insights/More-pixels-offsets-noise!
Although in this case where we subsample it would be better with bigger pixels.

[rparmar]

As far as sound, I don't really care if it's mono or stereo. You can get a decent Stereo Field Recorder pretty cheap (Zoom makes a 4 channel one for less than $200) and you can get better sound than any pro-level video camera.

True enough. And the Zoom is at the low end of quality. There are many better choices, though of course they are more expensive. One thing for sure: you won't be using the built-in camera mic for anything.

Will the K-7 produce SMPTE time-based video output?

That would be highly unlikely. But it would instantly notch up the credibility of this feature.

[Philippos]

I'm not an expert on sensors so I may be completely wrong here, but my understanding is that the S/N ratio has nothing to do with the size of the sensor itself, only how big the pixels on that sensor are. A 6mp sensor of any size will have a better S/N ratio than a 12mp sensor of the same size, because the 12mp sensor has to use smaller pixels to fit them all in.

IF my understand there is correct, then it shouldn't matter how many pixels on the sensor you actually use. The size of those pixels remains the same, and thus the S/N ratio should remain the same.

Seems logical at least, but I may be wrong. And I don't know how this effects DR, just thinking in terms of noise.

Unfortunately, it is not only a matter of size, it is also a matter of good electronics.
The Philips sensor Pentax planned to use in it's MZ-D (aka K-1) was a real full-frame sensor, bigger than the biggest full-frame DSLR sensors in use today. And it only had 6 megapixels. Fancy BIG pixels? This is THE SHIT! A real mosaic.
With your logic, it should have ZERO noise even at ISO 100,000. However it's design and electronics were so poor, that the damn thing produced terrible amounts of noise even at ISO 100.
- Pentax of course chose not to produce the camera, but another company the legendary Contax (something like the Rolls-Royce of SLRs), decided to produce a camera based on this chip and eventually threw almost three years of research and design and all of it's money to produce the Contax N-Digital, a product that failed to sell and caused the collapse and death of Contax altogether.

We don't know what the future holds for us. Maybe someone will make a new type of sensor that can produce ten times more pixels with ten times less noise in an even smaller area(something like the single-pixel sensor).

[mutedphotos]

Gimbal and Philippos: You're both adding other factors into the debate. My argument had to do with pixel size alone, and assumed all else was equal, including the electronics (Philippos) and the amount of megapixels (Gimbal).

A 10mp P&S has to use smaller pixels to fit onto its small sensor, so a 10mp APS-C sensor will have bigger pixels and thus better S/N. But, if for strange reason they shrunk the pixel size of the APS-C sensor down to the same size pixels as the P&S sensor (and I guess just spaced them out to fill the whole sensor), I would imagine the S/N ratio would be identical between the two (provided the spacing of the pixels doesn't mess things up, not sure there).

EDIT: Apparently I should read my own comments! I did bring up 6mp vs 12mp affecting noise, which was maybe a bad example of my point. Regarding the original question, if the resolution is the same for video, it shouldn't matter if the sensor only uses a fraction of its pixels, as those pixels are still larger than the ones found on a P&S and thus should still have a better S/N ratio.