How Far Can ML Push It?

[Kami07]

So I haven't seen the movie but I heard that "The Creator" was shot almost entirely with a Sony FX3. A $4k camera.

It got me thinking, can ML push these older DSLRs to produce something similar? I've seen some great short videos in 4k shot with ML but is that the extent of it? How does other content fair using Magic Lantern? I haven't seen anyone post any series, interviews, action scenes, tiktoks, vlogs, commercials, adult films, documentaries or movies. Just short films, usually scenic and a music video here and there.

If one took the time, could say, my Canon T4i with crop mood produce a 1-3hr cinematic movie? I'd love to see someone knowledgeable and resourceful commit to a project like that. I'm afraid it would take me years of learning to even come close to being able to attempt such a thing. Between ML, A.I software and all the technical stuff I don't know about, I'm sure some great things can be done, even with a small budget.

[a.sintes]

As stated by The Creator DP team itself, the fuzz about the Sony FX3 is just unbelievable: the decision was mostly driven by a practical question of size and freedom of movement, the camera being maybe the less important ingredient to create a "cinematic movie" if you consider the whole equation "scenario+talents+production design+lighting+lens+camera+postprod".

Then think about how good a FX3 is compared to the cameras used to shoot movies in the last 50 years and imagine what Spielberg would have done back then with any potatoe cam or an iPhone...

So if you got 4K$ left you'd better invest them in good locations, costumes, lighting or maybe lenses if you want to achieve a "cinematic" movie.

Note anyway the biggest difference between any ML-augmented camera and the FX3 is the reliability: ML is basically a hack, meaning you WILL struggle with unexpected camera stops, limited compatibility with external monitors and recorders, limited focus capabilities, a painful post-production workflow due to MLV format, card spanning etc.

The main difference between a feature film and a short being money and time: it's ok to spend time with technical issues and to workaround limitations when alone without external/money pressure but barely possible when dealing with a full team of 50+ people and deadlines.

[Kami07]

As stated by The Creator DP team itself, the fuzz about the Sony FX3 is just unbelievable: the decision was mostly driven by a practical question of size and freedom of movement, the camera being maybe the less important ingredient to create a "cinematic movie" if you consider the whole equation "scenario+talents+production design+lighting+lens+camera+postprod".

Then think about how good a FX3 is compared to the cameras used to shoot movies in the last 50 years and imagine what Spielberg would have done back then with any potatoe cam or an iPhone...

So if you got 4K$ left you'd better invest them in good locations, costumes, lighting or maybe lenses if you want to achieve a "cinematic" movie.

Note anyway the biggest difference between any ML-augmented camera and the FX3 is the reliability: ML is basically a hack, meaning you WILL struggle with unexpected camera stops, limited compatibility with external monitors and recorders, limited focus capabilities, a painful post-production workflow due to MLV format, card spanning etc.

The main difference between a feature film and a short being money and time: it's ok to spend time with technical issues and to workaround limitations when alone without external/money pressure but barely possible when dealing with a full team of 50+ people and deadlines.

Interesting. I'm big on working with what you got. I think it would be a great challenge to attempt to shoot something noteworthy using ML. I think for the right type of movie, one could shoot it for less than $4k.

I was thinking about how there are Magic Lantern EOS M vids from almost 10 yrs ago on Youtube. Imagine 10 yrs of footage that you can use for a movie. All with little to no difference in quality. You could get so many great shots doing so many different things in so many different locations. The flashback potential would be insane.

Once I become proficient at photography/videography I will delve into cinematography and get creative with ML.

[a.sintes]

Note anyway there are already a bunch of feature films shot partially or entirely using a Canon EOS (generallly the 5D mark II or III, the latter being able to record 4K RAW 14bits, which is more than the FX3) and ML, e.g. Rubber

shotonwhat is a great resource to find out: 5D2, 5D3

Get creative with your mind, a camera is just a camera and ML is just software.

[Kami07]

Note anyway there are already a bunch of feature films shot partially or entirely using a Canon EOS (generallly the 5D mark II or III, the latter being able to record 4K RAW 14bits, which is more than the FX3) and ML, e.g. Rubber

shotonwhat is a great resource to find out: 5D2, 5D3

Get creative with your mind, a camera is just a camera and ML is just software.

Interesting. Ironically I don't see anything on that site shot with the 650D or 700D but plenty shot solely with the 550D/T2i. I don't plan on investing in a new camera anytime soon and will definitely make due with what I have. The creativity is there, I just need to develop the skills to apply it and bring what's in my mind to video.

[Skinny]

I've done a little bit of youtube videos with 5D2 (nothing special), we also made a test commercial using this camera and a fashion video. And I shot thousands of clips with it in RAW just for fun or testing some lenses or other stuff.
So I can say there are some limitations.

1. It doesn't have an articulated screen, and sometimes it is hard to see what is going on depending on the situation. For example, you can't just flip the screen and record yourself for youtube.
And if you connect an external display, it may not work with RAW video, I don't really remember if you can use it, at least I think in some crop modes it won't work. And when you go into the menu or do other stuff, it is painfully slow - the display stays black for a second or two, while the built-in display works instantly.
Even if you connect an external monitor, you have a big nice display, but you still need to control everything from the camera so you have to get up, go and do whatever you need, but you can't see what is going on because the internal screen is off, while HDMI is connected..
If you have a remote for recording start and stop, you still can not review your footage without going to the camera.. And you need to be able to check your footage quickly because sometimes you don't know whether it was a good take or not, or you are not really sure if it was even recording because it could stop sometimes.
So working with an external HDMI display is not really convenient.

2. When working with RAW it takes some time to switch between x10 and normal view for focusing purposes (because it enters crop-mode with x5 or maybe some other stuff). If you bind the half-shutter for x10 magnification and do it this way, it could crash the camera. So focusing also takes more time with RAW, but could be done.

3. You can not review your raw footage in real-time, sometimes you think it looks OK, but later when you process your clips on the computer you see it is too shaky or something else.

4. You need an expensive fast and big CF card, or even a few cards.

5. You need to have a good computer, not just some 10yrs old garbage, a lot of free HDD space and processing ML RAW takes more time, also copying files from the card takes more time, etc.

6. 5D2 drains battery rather quickly so you need an adaptor and some external battery or a powersupply, that's not a problem but anyway.

7. Sometimes it is difficult to use it with a stabilizer, because you don't have an articulated screen and you can't remote control everything. Even if you use a small HDMI display it will add weight or if you mount it on the bottom of the stabilizer (mechanical one, like steadicam) then you can just kick it with your leg or something, or hit something with it by accident.. Not an ideal solution anyway.


And I am sure I forgot other things that makes this camera less user-friendly for some work... Don't get me wrong, I am not complaining, it is just what I found.
But I love the colors you can get with RAW video and how you can push it in any direction and the image will not fall apart like h.264 would do. There is something very nice in RAW and I absolutely love it


If we could have an option to connect a small external monitor for framing, while built-in screen will stay at 10x zoom for focusing, and some options like maybe a remote control for camera buttons, and maybe better clip review preformance.. Then it could become a very useful camera for some serious work.

I also think you can do some serious projects with ML, but you need to have a team. So you will be focused on the camera only, and other people will do other stuff. But such projects usually have a budget to use more professional gear and they probably won't consider ML-enabled cameras for their work.