Guest post by Jeremy…
Video stabilization is a great tool for tweaking and smoothing video that was shot with out (or even with) the use of some sort of rigging to keep the vibrations at a low rate. I use Final Cut Pro to edit my video, and have always been impressed with the results of the default Smoothcam filter/effect. I was asked by Alain to do a review on CoreMelt Lock and Load X and give my honest opinion.
I ran the video through 2 different tests, the first was a simple pan that was shot without the use of my tripod. The footage was about equal on both parties results. I was impressed with Lock & Loads X (LLX) speed. It took Final Cut Pro over 2 minutes to analyze the footage, while LLX analyzed it in under 15 seconds, with a render of about 20 more seconds. The final end product were about the same. I left everything on the default settings for this run and was impressed with the results.
For the second footage, I decided to test a moving object, an old spinning mill wheel at a local resort. FCP’s Smoothcam took over 5 minutes to analyze the footage, while LLX took only 45 seconds with a render time of 30 seconds after that. These results were not as cut and dry.
The Smoothcam render has a noticeable wave coming from the right side of the screen, at the base of the mill. If your looking at it, the footage seems like it is spinning. I know this is due to the fact of how Smoothcam processes the data.
LLX’s footage was a bit rough around the edges, I had to go in and play with the settings to get a decent shot. I actually ran the footage through two different ways, one in the “lockdown” stabilization mode, the other in the “smooth single shot” option. I was not as impressed with the footage. I did some tweaking on the “lockdown” mode to get it as smooth as possible. I admit, I could be doing something wrong, but compared to the “smooth single shot” the footage was better. The “smooth single shot” mode is the default setting when you start LLX to process the footage, and what I have noticed on the video I shot (along with the video on the demo real on Coremelt’s website) is that when using this method, the stabilized video becomes wavy, almost like your video was being projected onto a white screen blowing in the wind. With the hints of movement, it can make your audience feel seasick. I am going to continue using the trial and if I come across findings that alter my judgement, I will post them below.
Pros
It’s fast, and non intrusive to the editing process. The whole process is very easy and the support from CoreMelt’s website walks you through step by step with great explanations on the science that goes on behind the scenes. Coremelt’s plug ins have a reign of success because they are efficient at what they do. For panning shots, I encourage you to try the LLX plugin.
Cons
I was not impressed with the really shaky footage. The warping effect would deter me from using this plugin on anything more than panning shots. Any action shot I ran through the process was playing too many tricks on my eyes. Even though I got the same experience with the Smoothcam filter, I was more impressed with the footage for shots like this. Even on the footage they have on their website’s demo reel, the you can see there are issues with the periodic wave of pixels going in and out.
Result
As stated before, panning shots are awesome and LLX is a time saver. Is it worth the $150 price tag… not in my opinion. Even the time ratio equation doesn’t work for me. Since Smoothcam can work in the background, it isn’t so intrusive I would consider buying a program to help, I would invest it in better stabilizing gear to get the proper shot in the first place. I am not badmouthing CoreMelt’s programs, they have other great plugins you should try for yourself.
Tired of always answering the same question in every forum I go to, I have decided to make a blog post about it…
So you want to get a 70-200 lens but dont know which one to buy? Should it be the 70-200Lf4
To answer these questions, I would ask two others. Read the rest of this entry »
One of the most frequent question asked by people trying to gear up for cinemaphotography is which lens they should buy, and most specificaly if they should get a lens with IS (VR for Nikonians) or not.
While this question as been answered many times when SLR were only able to shoot pictures, few people have looked into it from a cinemaphotography perspective.
So here it is! My totally unscientific side by side comparison between IS and non IS at two focal length. I do not have a 24-105ISf4 so I had to use the only IS lens in my bag, the 70-200LISf/4. I did two simple test: one at 70mm ELF and another at 200mmELF (coming soon, need editing!).
Disclaimer: My hands are not steady at all, this is why I always carry a tripod around! And to make things worst, I took these clips after working out. This is a worst case scenario test, I was holding the camera without any support but my elbows pressed against my body.
Image stabilization and the 5DMrkII @70mm
Image stabilization and the 5DMrkII @200mm
As a bonus, I have done some post processing in the last segment of the clip to show how much more improvement can be done by applying a software based IS filter. In this case, I used the one from Final Cut Pro.
I guess it is clear how much the IS contributes to the overall quality of the clip. I would love to have somebody do the same test at an EFL of 24mm to see how much of a difference it makes. In fact, I am thinking about getting rid of the 24-70 for a 24-105, a move I would have never dared to think of a few months ago!
Note: sorry for the 200mm video being in low def, seems like I cant be a pro user of Vimeo since I am not USians…
