Fast lenses and film emulation can resurrect old cameras (ft. GH2 night footage!)

[kye]

Old cameras have a number of challenges, including:
- weak codecs, often 8-bit low bitrate files
- terrible low-light
- dated colour science and no log profile (rec709 profiles only)
- poor DR
- lack of IBIS or EIS
- etc

At the time these were pretty significant challenges.  Now they aren't the challenges they used to be, because fast lenses and film emulation assist with all these limitations.  Let's take these one at a time.

Weak codecs
Weak codecs, including 8-bit low-bitrate files can be soft, and can be overwhelmed by motion.  By shooting with faster lenses you render more of the frame out-of-focus and therefore the limited bit-rate only has to focus on a smaller percentage of the frame.  Thanks to cheap Chinese optics companies, we are now awash in F1.4, F1.2, and even F0.95 primes.  
The soft image is now no longer a liability, because compared to our modern 4K sensibilities, even 35mm film is noticeably soft by comparison.  This means that by adding film emulation you'll be softening those edges and smoothing over any subtle compression artefacts.  Film often has a more compressed colour palette, pushing hues closer together in many instances, lessening the visibility of artefacts.  It doesn't work magic, but every bit helps.

Terrible low light
Cheap F1.4, F1.2 or even F0.95 primes sure make a big difference after the sun goes down.  That "fast" F2.8 vintage lens you were shooting on back then is 3 stops slower than these things now.  That can really bring a lot of situations back from being unusable to being at, or close to, native ISO.

Dated colour science and no log profile
Rec709 colour profiles are basically a creative filter the camera has applied, and they often weren't that good.  Film emulation takes that image and applies an incredibly large transformation over it, which goes a long way to hiding any imperfections the colour profile might have had.  It's like if you put on a pair of rose-tinted-glasses, you can still see that things have different colours, but any subtle differences aren't visible because the image has had a strong look put over the top.
Also, film emulation plugins often come with controls for exposure and WB etc, which can help to grade the 709 footage, which was a major pain back before we had colour management pipelines.

Poor Dynamic Range
You know what else has pretty poor DR?  Print film!  Kodak 2383 has about 5-6 stops in the linear region, and then everything else in the image is squished into the highlight or shadow rolloffs.  Yes, you can see into those rolloffs a bit, but if your camera has 8 stops then you've got at least a stop to put into each rolloff.  People think film has huge DR, and it did at the time compared to consumer digital cameras, but it was the negative film that had the huge DR, not the print film.  It's very common now for people to shoot on film, scan it, and then do everything else digitally, so they keep the full DR of the negative, rather than taking half of it and pushing it into the rolloffs.
This is a still from Minority Report from 2002:

It's not exactly a dynamic range demo - the streams of light INSIDE THE ROOM are blown out and every item of clothing the main character is wearing is crushed blacks.

Lack of IBIS or EIS
So there's a little shake in the files...  well, film had this thing called Gate Weave, which was where each frame didn't perfectly align in the camera and so when played back there was movement of the whole image.  Once we started doing digital intermediates people started stabilising the images digitally and that went away.  When I went to the cinema and saw Goodfellas projected on celluloid they played a bunch of old ads and movie previews also on celluloid, and some were jumping around all over the place and some were rock solid (which means the projector the theatre was using wasn't the source of it) and much to my surprise, Goodfellas itself had quite a bit of it.
By just using modern tools you can now stabilise things pretty easily, but this will create artefacts if you do it too strongly (especially if the camera had bad RS), but applying film emulation gives you much more leeway.  This is because you can stabilise the image, then apply some Gate Weave, and once the viewer notices your images look like film they'll potentially just accept the shake in the image as being part of the film look.  By adding Gate Weave and getting some grace from the viewer you can potentially increase the strength of the stabilisation you're applying too, with there being more wiggle room, and also because the softening of the image will mean that any distortions in the image will be slightly less visible.

I was inspired to write this partly from my GX85 Super-16 camera project, but also partly by this video of the GH2 shooting at night.
You can still see the ISO noise and macro-blocking creep in as blue hour ends, but he was also using the 9mm F1.7 and 35-100mm F2.8, the F1.7 is reasonably bright, but the F2.8 is pretty slow compared to things like the TTartisan 50mm F1.2 or the 7Artisans 35mm F1.2 primes that are $109 and $97 on B&H.  These won't offer OIS, so your options for these on non-IBIS cameras are to spend more (Canon and Sony both offer 35mm and 50mm F1.8 primes with OIS) or to use a tripod or larger rig of some kind.  

Far from perfect, but much more useable than you'd think.

These cameras have actually gotten better over time as the rest of the ecosystem is better able to support them.  The only reason we don't think so is that our expectations have inflated faster than their potential.