"film look" ironies

[A. J. deLange]

Originally Posted by Steve Rosen

Steve: A friend has a Sony and I don't notice it from his camera at all, but his image is considerably softer too.

That, in a nutshell, describes the trade space. This is hardly the first time this has been discussed in these fora but digital cameras sample the picture and sampled systems are subject to aliasing. The way to eliminate it is to soften the image before it reaches the sensor so that there is no energy at spatial frequencies higher than the pixel spacing. If a manufacturer does that he is faulted for producing a camera with poor resolution. If he doesn't he is criticized for "jaggies", twitter, moire or whatever you want to call it. Canon, in the XL2 caught it for the latter more than most because they had a darn sharp optical system for SD. Subsequent investigations have shown that the SD lenses are about as sharp as the HD one. Their MTF's don't even start to roll off until about 400 lines and are about 90% at 480 lines, the resolution of the sensor. This did allow aliasing in many situations like the ones people have described here. Where the picture didn't contain high frequency components we sang the praises of the camera for it's sharpness. Where it did we moaned about jaggies and crawling except for the guys who understood how to get around them (i.e. diffusion filters, going to the p modes which reduce vertical resoultion, moving the camera to a position where the beat frequencies were not noticeable etc.). I applaud Canon for giving us the option to have the shapness when we wanted it and combat the jaggies when we had to.

With the H1 the ultimate resolution of the sensor is 1080 in the V direction and the lenses still break at about 400 though the response is still 50% at about 600 and 20% at about 800 it's only a couple of % at 1080, the folding frequency. Thus less aliasing is to be expected with the H1 and I have certainly noticed this. Reducing resolution even further by going to the f modes makes it even less likely.

BTW I believe "crawling" leaves to be largely caused by the compression algorithms (DCT in straight DV and DCT plus prediction/interpolation residuals when further compressed by MPEG) though aliasing doubtless does contribute to it.

[Steve Mullen]

Originally Posted by A. J. deLange

BTW I believe "crawling" leaves to be largely caused by the compression algorithms (DCT in straight DV and DCT plus prediction/interpolation residuals when further compressed by MPEG) though aliasing doubtless does contribute to it.

Great explanation. And, aliasing being TINY unwanted detail further stresses the compression system. (As do edge outlines.)

[Gary L Childress]

Just getting back for a moment to depth-of-field and the "film look". Every time I have seen people talk about the use of shallow DOF, it's described in a dry "these are the rules" kind of way. You use shallow DOF to direct the audiances attention. The better creative DPs' or directors I have worked with use short DOF more for the artistic feel of the look than to "direct the audiances attention". And if you look at it as more of a emotional choice than a narrative one, there are just as many reasons to use deep focus shots as short focus. One of my favorite action films is John Frankenheimer's "Ronin". The director's commentary track is really interesting, and he talks alot about using big deep focus shots. He explains his use of deep focus really well. If you have not seen the movie and listened to the commentary track give it a try.

[Steve Rosen]

A.J. I certainly would never be said to have a problem with the sharpness of the image from the H1... The problem, I think, is that in an effort to offer the most resolution possible, the detail default is somewhat exagerated.. as I've said before, a good image isn't all about resolution. As a matter of fact, I tend to use diffusion (1/4 and 1/2 ProMist or SoftFX) most of the time to soften the edges, especially when photographing people... I do this in 35 and 16 as well... and dial the Detail down on the Canon...

Gary. It's interesting that when you study films of the past you see that shallow depth of field didn't really become "in" until foreign films of the 60's, particularly French films, hit US distribution. They were shooting with relatively slow film, minimum, sometimes no lighting, and handholding those old 35mm Eclair Camerettes with their lenses wide open, and were forced by necessity to develop a new style of cinematography, a style that was heavily dependent on skilled assistants pulling focus. Now it has become what many people think of as "film look" - Funny, if they'd had 1/3" cameras like the H1 that style may have never happened...

Just yesterday I viewed a contemprary French film, READ MY LIPS, and boy does that AC have his act together...

[A. J. deLange]

Originally Posted by Steve Rosen

The problem, I think, is that in an effort to offer the most resolution possible, the detail default is somewhat exagerated..

There is no doubt about it - the default is for in-camera sharpening. This shows up very plainly when one reconstructs the edge from a test target for MTF estimation.

[Steve Rosen]

Well, you're way more sophisticated than I am - I go as far as film stock tests in film and experimenting with presets shooting typical footage in video, but that's where my expertise comes to a screaming halt...

I have got to say that the footage from this camera is pretty remarkable.. I just re-edited a show I originally made 2 years ago for a client that I originally shot with my old DSR500 in 16x9. I had to add some new down-rezed footage from the H1, and I was a little nervous about it - Except for having to do some color correction, the footage actually looks better than the DSR stuff, and it was a top-o-the-line 20K DVCAM camera...