THIN + vertical gausian blur = good idea?

D

dodongo

Member
Hello,

I'm wondering if you shoot THIN, apply a small vertical gausian blur in Vegas, and finally output to DVD, can you get a better looking image in the end than if you shot THICK?

I'm basically thinking that you can end up with the same resolution [and lose the flicker from THIN with the blur], but also kill some artifacts by using the gausian blur instead of just going straight THICK.

Is this a crappy idea? Would you end up with a blurrier image [or an image w/ strange-looking horizontal artifacts] by using this technique?
 
Interlace flicker can be reduced with vertical lowpass filtering, but doesn't affect aliasing artifacts. Thick mode means optical anti-aliasing by reading out two lines at the time,. This optical vertical sampling aperture widening not only results in a vertical lowpass filtering (line twitter reduction in interlace display mode) but also reduces aliasing effects wich can not be suppressed by appling blurr on a "thin" signal.
 
andre said:
Interlace flicker can be reduced with vertical lowpass filtering, but doesn't affect aliasing artifacts. Thick mode means optical anti-aliasing by reading out two lines at the time,. This optical vertical sampling aperture widening not only results in a vertical lowpass filtering (line twitter reduction in interlace display mode) but also reduces aliasing effects wich can not be suppressed by appling blurr on a "thin" signal.
Click to expand...
What is this guy? Klingon?

Man, that's a great info, but I understand less than half of it.
 
Andre, thank you!!!

I'm still confused about something, though:

My original thinking was that if you have artifacts in the footage, [digital grain artifacts or the "mosquito" noise], applying a blur would soften them up or make them fade into the picture [and get rid of interlace flicker].

I figured you'd get these artifacts/noise if you shot in THICK or THIN, so why not shoot THIN and then soften the artifacts/noise with the blur. That way I might end up at a similar resolution as shooting in THICK, but with softened artifacts [and no interlace flicker]. I also figured the blur would help to soften aliasing, as well.

I think you were saying that shooting in THIN gives your footage an aliased look that a gentle blur is not capable of getting rid of. [I definitely don't want to end up with aliased-looking footage]! Is that correct?

I'm a newbie bugging a pro, but I'm sorry I'm still a little confused about that.
Thank you!
 
The basic aliasing problem is on the optical/CCD level. Aliasing shows up when e.g. a scene contains details which come close to the CCD pixel pitch. Moire effects (interference) which are no longer related to the detail frequencies will occur and can not be removed by any low pass filtering technique. Suppose (and this can easely be verified ) you have a scene containing horizontal white lines on a black background which come so close to each other (zoom-out) that some lines image just on a sensor row and other in between two CCD sensor rows: white and black zones will be the result which will show even more annoying effects when displayed in interlace mode : this is one form of spatial aliasing. Pro cams have (expensive) optical anti aliasing(AA) filters. Comsumer cams "widen" the sensor area by taking two rows at the time for readout. This overlap technique results in vertical lowpass and AA filter as well
 
So basically - it's safe to shoot thin when there aren't objects like blinds, grills on the front of cars, stuff like that which is close together and stuff that would end up looking jagged like that?
 
Yes, in order to avoid the moire aliasing effects, which I gave as an example. Line twitter (flicker) in interlace mode on sharp (horizontal) edges and even in progressive mode, slow tilts shot with high shutter speed settings will still show twitter effects on CRT displays when applying thin mode. The latter two effects are not only created by the spatial properties of the sensor (discrete pixel structures) but also the time/position effects and are classified as spatio-temporal aliasing effects
 
You must log in or register to reply here.
Back
Top