I like CMOS cameras in spite of the rolling shutter and camera flash problems. CCD cameras have their own set of problems, like if you have a light shining directly in the background. A CCD camera will make a band that extends the height of the frame. A CMOS camera just overexposes the direct light area. CMOS cameras have way more dynamic range as well (even if they are a little noisy in the darkest shadows). Both sensor types have their good and bad points, but overall, I prefer CMOS.

I do wonder is that was not offset by the CRTs in our TVs also do the same thing i.e. taking the same time to present the image as the electron beam scanned down the tube.That is definitely true, although it still makes me wonder why I don't see the problem when I play an old VHS tape on my modern display.

BTW, you comment also suggests a "fix" for the CMOS problem, namely create displays that cannot only display interlaced natively (I see no reason why alternative rows of pixels cannot be addressed independently), but do so in a fashion that will match the nature of the CMOS scanning rate.

Finally, another nugget I discovered is that if you go to the main Deshaker page:

Deshaker

and scroll down to where the CMOS rolling shutter is discussed, you will see that Thalin (the Deshaker programmer) has found that there is indeed quite a difference in how each camera operates.

"you will see that Thalin (the Deshaker programmer) has found that there is indeed quite a difference in how each camera operates. "

Indeed, different cameras have different readout times and with the one camera they can very depeding on the use of digital zoom and higher frame rates where resolution is dropped. In both latter cases less of the sensor is scanned resulting in lower readout times and less skew.
It is worth a mention that despite this problem the high speed video cameras use CMOS sensors.

Resolving the skew problem can go from relatively simple to pretty much impossibly complex. I'm going from a post on Reduser from a few years ago by a VFX specialist. Consider what happens if you're tracking a fast moving car as it is driven down the streets of a city. The car itself suffers no skew as it is static in the frame. The buildings suffer different amounts of skew and the front of the buildings and the sides of the buildings are also different. Trying to track and correct that becomes a very complex problem in a 3D space. It mightn't look all that bad but if you're trying to do set extensions or building replacement its vital to get it corrected.

Thankfully the one time I've shot footage that I knew would require stabilisation although I didn't have much time to plan for the event I did have the presence of mind to grab a Z1 and set the shutter speed at 1/250. I shot around 40 minutes from a Bell Longranger copter and the wide shots after stabilisation of the city, Harbour Bridge and Opera House would have been usable if it wasn't for the poor overcast weather.
I think we had some discussion about that just after I had shot it and the final conclusion was that when it comes to shooitng from copters you need to get below 500' and there is value to be had from gyro mounts neither of which were available to me.

Bob.

re: Laurence & johnmeyer's suggestion to study cmos wobble & shutter speed.I made a few tests and I find that I really have a hard time getting my Panasonic TM700 to show any Jello effect at all. I tired 1080 60p, 1080 60i and also tried 1080 60i on my Canon HG21.

I set up a turntable (remember those? [chuckle]) and set the speed to 45 rpm, then turned the manual control as fast as I could (maybe 50 rpm?). Here's a link to video in full speed & 25% SloMo at 60 fps: Rolling Shutter It's really hard for my eyes to see any Jello effect at all.

Upon closer examination, it does exist, however. Note the following screen captures:





My conclusion: at least for my cameras, this effect is so small it's just not something I'm going to worry about. Now someone will probably jump in and say, "If your motion is faster, the problem will be more apparent.", but my counter is... this motion is still faster than anything I indend to shoot.

So, as all posts in this forum... YMMV.

...Jerry

"Now someone will probably jump in and say, "If your motion is faster, the problem will be more apparent.", but my counter is... this motion is still faster than anything I indend to shoot"

As I don't shoot chase scenes then nothing I "intend" to shoot has a problem induced by my camera's rolling shutter. Where I do strike a bit of a problem is with what I don't intend to shoot but cannot avoid.

If somehow the camera / tripod is kicked or the floor jumps as someone stomps past me on a sprung floor then the skew adds to the impact of the motion. The other thing is when the stills guy(s) fire off their flash gun(s). The latter is quite a problem and it can look really bad when shooting interlaced video.

Bob.