The end of 35mm adapers and hv20s....

Interesting thought about this rolling shutter: A few years ago I was using a machine vision camera with a rolling shutter in a homebrew hd camera experiment. No body used CMOS chips then so the problem was hard to wrap your head around.

The interesting thing was, when we would pan up... people would suddenly turn into midgets for the first frame. I think D90's rolling shutter is a huge step forward, and as someone mentioned... RED ONE does do this.... but no hardcore reduser would ever admit it.

Increasing the shutter speed, and the ISO speed would effect the difference of the rolling shutter. That footage is clearly 1/24 or 1/48... personally I always shoot that 1/60 cause it looks better. I'd challenge someone with a RED ONE to post similar footage say... exposure as open as it can be, 1/24 shutter.

This is a flaw of a CMOS chip... but it isn't that bad so... I'm still siked.
 
I personally suspect - though my knowledge of rolling shutter is quite limited - that the D90 sensor has a readout rate of only 24 fps. Therefore the rolling shutter should be more than twice that bad than with the RED or the EX1. (I think they have a read out speed of 60fps.) SCARLET will even read out at 120 fps. and therefore will be the best CMOS camera on the market with less rolling shutter than any other up to date.

But imagine a D90 with twice the read out frequency (48fps.) and therefore variable framerates! That would be still worse than RED or EX1 but quite close.
 
Thank you again sugoroku and dfreerider. Those birds look good to me! The only thing again is the exposure change but otherwise I really like what I am seeing. And the stills? Sweet! Can't wait to have it in my hands.
 
I really want to see skew examples from the RED camera now. If this is still a problem here, i'm willing to be it is still a problem for red.
 
dave12781 said:
Looks to me as if RED has an even bigger problem with the image tilt.
Click to expand...

That footage was taking with Build 12 back in January.
They are now on Build 16 firmware, and the skew issue has been improved significantly.
And apparently it will be improved even moreso with the new sensor to be included in Scarlet, Epic, and upgraded RED Ones.

Skew is just the nature of the beast with CMOS sensors (for now at least).
Just like film cameras - much can be done to minimize it, but it will always be a consideration.
(barring some sort of technological breakthrough of course).
 
Luis Caffesse said:
That footage was taking with Build 12 back in January.
They are now on Build 16 firmware, and the skew issue has been improved significantly.
And apparently it will be improved even moreso with the new sensor to be included in Scarlet, Epic, and upgraded RED Ones.

Skew is just the nature of the beast with CMOS sensors (for now at least).
Just like film cameras - much can be done to minimize it, but it will always be a consideration.
(barring some sort of technological breakthrough of course).
Click to expand...

Even still... I don't see how a firmware upgrade can improve the failings of a sensor. Seems to be a worse problem on Red, and maybe even Nikon will come out with an upgrade to eliminate this problem.

All said... I don't see how red could make this problem go away unless they use a global shutter, or switch to a CCD type chip.
 
joe 1008 said:
So, the D90 compared to EX1 and RED, the rolling shutter is it worse or equal in your opinion?
Click to expand...

I would say so. I'm going to change my consideration of this to a "non-issue". And one I didn't have to spend 17k on to find out.

Again... GO NIKON!

And as for what REDUSER has to say about it... seems as if they recommend increasing the speed of the shutter... and starting production on "Gladiator 2: Gladiatordome!"

That totally deserved a drum hit.
 
dave12781 said:
Even still... I don't see how a firmware upgrade can improve the failings of a sensor.
Click to expand...

That's a question for Graeme or some of the other guys at RED.
All I know is that if you read the comments from RED owners - they all seem to agree that the skew is significantly improved in recent builds.

Granted - RED also recently announced a new sensor (and among other improvements they claim less of a skew issue) -
so it seems that there is only so much that can be done via the firmware upgrades.

I don't see how red could make this problem go away unless they use a global shutter, or switch to a CCD type chip.
Click to expand...

That's exactly what I said - they can't make it 'go away' just like we can't make it go away in film cameras. It can be improved and minimized - it's not an all or none situation... but barring some technological breakthrough CMOS sensors will always have this issue that needs to be dealt with.
 
Tried working with the original file (the only one that was working) from dpreview in PPro. Is anyone having an easier time working with these MJPEG clips than HDV? The file from dpreview is lagging on my computer without any added filters or effects, whereas native 1080p HDV works just fine.

Color correction-wise, forget trying to up the saturation. The macroblocking shows up at the slightest increase. Though I must say, the highlights in the close-up-clip of the black and white bird were handled pretty well by the camera. There seems to be information even in the brightest highlights.
 
Last edited:
You must log in or register to reply here.
Back
Top