Maestro is a nice short movie, shot with a Canon HV20 and a Twoneil’s 35mm adapter. The plot: a young musician stifled by the modern day schooling system tries to find his way to make his mark. HD version & download here.

Canon announced today it will release a firmware update for the EOS 5D Mark II allowing users to manually control exposure when shooting video. The firmware update will include the following manual controls when shooting video:

* Full aperture selection
* ISO speed: Auto, 100 – 6400 and H1
* Shutter speed: 1/30th – 1/4000th second

This, in addition to its exposure compensation, exposure locking abilities, and gamma/color/picture settings that can be modified via templates generated with Canon’s PC/Mac accompanied 5D application.

This new firmware makes the 5D MII the best video camera out there for DV Rebels, and by far the best DSLR video camera out there. Of course, there are a slew of video-specific hardware features missing (e.g. XLRs), but in terms of actual visual quality obtained with its 38 mbps h.264 files, and with its new basic control options, it beats anything in that price range. Especially if you have some good lenses for it.

The only other feature that matters for us DV Rebels and that’s _really_ missing is frame rate support. 24p (=23.976) is still not there, and its 30p is 30.00 fps instead of 29.97. Additionally, 720p at 60p would have been nice too (for slow-motion usage). If Canon fix these two small problems (easy to implement), it has a real winner in its hands. There was a rumor a few weeks ago where someone said that Canon doesn’t want to implement 24p in its 5D in order to not cannibalize its prosumer cameras, but truth is, people are asking for it — a lot. They will have to comply.

In our home, we are 60% ready to buy the new 5D. Before today’s announcement, we were 10% ready. So we are close. But still not there. Add some frame rates that make sense, and then we will buy it with closed eyes. Promise.

When I look to buy a cheap digicam/digirecorder (NOT a camcorder) these are the video features I first look at, in this specific order:

1. Does it have exposure compensation?
2. Does it have exposure locking?
3. Does it have shutter speed control?
4. Does it have color/gamma settings? (low saturation/contrast/sharpness)?
5. Does it have manual white balance?
6. Does it have focus locking?

The first two are the most important video features one should be looking for. Without them, you can ruin your shots no matter how careful you are. For example, the Flip/Kodak digicam/digirecorders don’t have any of these features. The Panasonic ones only have #1 and #5. The new Canon ones have #1, #2, #4, #5, #6 making them a better buy overall (missing only shutter speed). Of course, a choice between 24, 25 and 30 fps would have been nice too.

Notice that I don’t list gain or aperture control. If I was looking for a still picture camera, aperture control would have been much higher up in the list, but for video, shutter speed is more important — at least for the kind of videos I am shooting. Needless to say that for all of my HV20 videos, none was shot in aperture mode. They were all shot either in shutter speed mode, or Cinemode.

Update: JBQ wrote a similar blog post too, about still cameras.

Here’s a small guide editing your DV/HDV clips with ffmpeg under Ubuntu Linux. All cutting, processing, and exporting is done via the command line (oh, the joy!). Sure, this method doesn’t have niceties like fades, titles, effects etc, but it works for basic editing. Please note that not all AVCHD formats from all cameras are supported.

First, you need to install the right version of FFmpeg, that supports all codecs (the default Ubuntu version doesn’t, for legal reasons). Follow the “B” solution from this tutorial to install the full version of ffmpeg on the latest versions of Ubuntu.

Now, find the DV/HDV or AVCHD video files you want to trim, or slice & cut. Watch a video file on a video player (e.g. VLC or Totem), and on a text editor (e.g. gEdit) type the minutes/seconds you want to cut. For example, you can type:
video1.m2t
– section 1: keep from 00:00:01 to 00:00:15
– section 2: keep from 00:02:37 to 00:03:54
video2.m2t
– blah-blah…
– blah-blah…

Then, load your glorious terminal, navigate to your videos’ folder, and type:
ffmpeg -ss 00:00:01 -t 00:00:15 -i "video1.m2t" -acodec copy -vcodec copy "edited-video1.m2t"
This means that ffmpeg will only save to a new file the video from the first second up to the 15th second. Adjust the numbers according to your needs. The transcoding will be completely lossless and done in mere seconds. Do the same for all your other sections/files. The original files won’t be destroyed btw, so you can re-cut them if needed.

After you have done this for all your files, you can put together the various edited files in a single file!
cat edited-video1.m2t edited-video2.m2t edited-video3.m2t > chapter1.m2t
Warning: The files must be of the exact same type! You can’t mix and mash different formats!

Now, you might want to burn a DVD! Here’s how to create a widescreen PAL DVD:
ffmpeg -i "chapter1.m2t" -aspect 16:9 -target pal-dvd "chapter1.mpg"
Use “ntsc-dvd” for NTSC videos instead of “pal-dvd” above. Also, you can have many different “chapter” files, that can act as different chapters to your DVD. You can use an app like DeVeDe to put together the various edited chapters/videos, and burn your DVD.

And here’s how to export for YouTube/Vimeo/PS3/XBoX360 in 720p HD h.264/AAC .MP4. First, let’s put all the chapters in a single video (if you have gone that chapters route):
cat chapter1.m2t chapter2.m2t chapter3.m2t > myvideo.m2t
And then, do the final export:
ffmpeg -deinterlace -i "myvideo.m2t" -threads 2 -f mp4 -vcodec libx264 -level 41 -refs 2 -loop 1 -deblockalpha 0 -deblockbeta 0 -parti4x4 1 -partp8x8 1 -partb8x8 1 -coder 1 -subq 6 -brdo 1 -me_range 21 -s 1280x720 -r 30000/1001 -b 5120k -bt 8192k -bufsize 15000k -maxrate 16000k -g 300 -acodec libfaac -ac 2 -ab 128k "myvideo.mp4"
Use “-r 25″ instead of “-r 30000/1001″ if you are on a PAL instead of an NTSC region.

And one last tip, for those who want the ultra-widescreen cinematic look. Add the code below to your (final) exporting code of ffmpeg, and adjust the numbers accordingly to your taste and create horizontal letterboxing:
-padtop 44 -padbottom 44 -padcolor 000000

Then, sit back, and enjoy the video on your TV. Oh, did I mention before that my TV runs on Linux?

A great HV20 video by fellow Greek, George Hantzis. Very good editing and vision throughout the video. HD version here.

Wow, I had no idea, until the thumbnail I was looking on Digg’s pages looked familiar! Scott McIntyre submitted my jellyfish video on Digg, and he managed to get it to the front page! So far the video has acquired 24 new “likes” and over 5,000 new views! Thanks Scott!

More info on how the Crank 2 movie was shot. They used five XH-A1 cameras as primary, and fifteen HF10 ones as crash-cams, sometimes shooting all at the same time (with several random crew members just grabbing a camera and shooting — mostly on rollerblades)! They used very high shutter speeds, between 1/1000th and 1/2000th, and with “neutral” picture settings for saturation and contrast, while they cranked up the in-camera sharpness. On the XH-A1 they also used its cine gamma settings, but on the HF10 it was not possible to use its Cinemode ability because it can’t be used independently to manual shutter speed settings (customizing the shutter speed obviously takes precedence in such an action movie). That’s yet one of the reasons why I am still rooting for that hybrid consumer-prosumer camera that doesn’t exist yet. But NAB is here, so let’s see what Canon will announce in two days time.

A few more tidbits about the movie:
- The movie is actually insane, but in a good way. Here’s a positive review.
- It seems that the movie won’t make more than $6-7 millions in the box office this week, but when you add in that count the next few weeks’ earnings, the worldwide release, and DVD sales, it will easily recuperate its $12.5 million of cost.
- If there’s going to be a Crank 3, it will be in 3D, the directors have let us to believe. These two directors really are cool.
- And good looking too.

Before and after. Click through for larger versions.

Both pictures by Tabass Co./Tatiane Brito, licensed under the Creative Commons “BY” 2.0.

Sony Vegas plugins used: NewBlue Duocrhome, Channel Blend, NewBlue Pastel Sketch, Magic Bullet, Color Corrector, Unsharpen Mask, Bump Map for lighting. Compositing “overlay” mode used.

The digital revolution is trying to take a serious shape. I was watching this brand new TV cop show “The unusuals” and I was thinking that it looked “different”. There is almost no shallow depth of field in it. I kept thinking that they used a cheap(er) camera, and they did: the $10,000 Sony EX3, with its 1/2.0″ sensor. They also use the much more expensive F23 sometimes, but the EX3 seems to be the camera used mostly. The CC and lighting is minimal and that makes the show look cheap, but I guess this is part of the idea behind the show to make it look like a documentary. Still, it’s cool to see a digital camera used for a TV show, because on most of the rest of the primetime shows, they use film.

High frame rates and 3D video are in our near future. Many times I have cheered for 60p over 24p, but I think this won’t be accepted by many directors until 3D becomes a reality. 3D just looks better in high frame rate, and so this might be the catalyst for the move to higher frame rates. Even James Cameron said so.

In the last year, 3D filmmaking has seen a strong comeback as new digital cameras help out with the complicated workflow. Even the TIME magazine did a special article recently about the 3D comeback. However, the true revolution won’t be here for another 10 years, when TVs will display 3D objects without the need for glasses. The current monitors/TVs that do no-glasses 3D are still not very good, so there’s at least a 10 year maturation period in that technology.

I have thought out the traditional 3D process, and I might give it a shot if I get a second HV20 or HV30. All I need to shoot in 3D is a second HV camera, a clapper for synchronization, and this tripod accessory. Then, the workflow to edit in anaglyph 3D on Vegas is pretty straightforward. Sure, there’s some extra work involved, but it’s not unachievable. Maybe one day Vegas adds 3D editing capabilities by automatically packing two or more different clips into a single track (and I am not talking about the “takes” feature here).

I have already exported a small 2D clip I had around as pseudo-3D anaglyph (by misaligning the two stacked copies of the clip in the timeline by 2-3 frames), and even that worked great when using the red-cyan 3D glasses! Download the Vegas Pro .veg file here to see how I did it (use a clip of yours, and pay attention in the track’s plugins, compositing settings and misalignment of the two clips by a few frames). The pseudo-3D trick with the Vegas Movie Studio versions requires a somewhat different workflow, follow it here.

The whole 3D craze started for me a few weeks ago when I saw the above 3D music video by Golden. I even wrote a review for the band’s album too, should be published soon at the local Bay Area OWL Magazine that I occasionally contribute at. Video requires red-cyan 3D glasses, best watched full screen. Download the original HD file for best 3D results.