Stereo 3D Alignment using different focal lengths

nicolesharp100 wrote on 7/31/2021, 3:43 PM

Does the Automatic Stereo 3D Alignment for Movie Edit Pro 2021, Movie Studio 18, and Video Pro X work with two different cameras using different focal lengths, sensor sizes, and/or focal ratios? For example, instead of buying two identical cameras, it would be much cheaper to use two old but different cameras I already have, but they will not provide identical views. Can the Magix software automatically adjust for differences in focal length/sensor size/focal ratio/lens distortion as well as for differences in physical position (other than horizontal displacement)? I know I could test this myself, but figure it would be faster to ask online if someone has already tried this. Otherwise cropping the second video manually so it provides the same field of view as the first might be painstaking.

Comments

CubeAce wrote on 7/31/2021, 5:15 PM

@nicolesharp100

Hi Nicole.

First there are few 3D content makers I know on this forum so you may not get many replies.

My feeling though based purely on normal videography and photography would be no.

In some of the past replies regarding 3D inquiries it would seem some form of manual adjustment is normally needed to get the best results.

Then as you say there would be differences of quality of output as well as depth of field out of focus areas that you could not match to make such footage seem seamless. As for adjusting for focal length. That would be a manual task and raise quality issues especially if the cameras in question had different sensor sizes and photo-site counts. One could look a lot more grainy compared to its larger sensored sibling due to the need to use a different ISO setting or the use of noise reduction making the image softer.

Ray.

nicolesharp100 wrote on 7/31/2021, 5:56 PM

I was reading an article on this, and I agree it might be difficult without identical cameras. Differences in focal length and sensor size will change the depth of field in addition to the field of view, so even if the field of view was cropped to match, there will still be differences in focus. The other issue is lens distortions, especially with fast wide-angle lenses commonly found on smartphone cameras. There is a health hazard too that if the two videos are not aligned correctly, it could induce headaches or nausea. An image/beam splitter would probably be a cheaper alternative to use with a single lens but I was thinking of setups where you would want the cameras at an increased separation for enhanced depth of objects further away.

I would like to see automatic 2D to 3D video conversion available with Magix software though. This feature is already available with Corel VideoStudio and CyberLink PowerDirector. 2D to 3D video conversion might actually get better results than trying to make side by side 3D video using two non-identical cameras. At least that way the data is presented in a consistent format within a single video. But need some degree of artificial intelligence programming to make sense of distance cues such as atmospheric blur, depth of field, and changing perspective. I was really impressed with some of the 2D to 3D conversion results from the free 30-day trial of CyberLink PowerDirector 19 Ultimate, but without a 3D camera, it's can be difficult for the AI to figure out which objects are what distance, so it only works out sometimes.

CubeAce wrote on 8/1/2021, 1:43 AM

@nicolesharp100

I also hope you realise that most 3D television production stopped in 2016. You can still get 3D projectors but the TV side is effectively dead and virtual reality is a slightly different format.

Maybe you should read this article among many others. It also explains other production problems.

While I think there is nothing wrong with 3D as a medium I wonder about the availability of a decent sized target audience. If you are doing it purely for your own amusement then I think this is great. Support for such ventures is always a good thing generating interest in less popular mainstream mediums.

Ray.