Making the most of the iPhone, GX85 and GH5 and shooting in the real world

[kye]

I've gone through a journey over the last few months, and have come to a new clarity in my work.

As most know, I shoot video of my family but want to make it look as much like high-end TV and films as possible.  This involves shooting in completely uncontrolled situations with no re-takes.

My new found clarity is this.  

Priority 1: Get the shot
You can't use the shots you didn't get.

Priority 2: Get shots in the best way
Shots that aren't in focus, are shaky, don't have good composition etc aren't optimal.  Also, having shots where subjects are aware and nervous of the camera, people in the background are staring at the camera, etc are also not optimal.

Priority 4: Get the nicest image quality
Insert all the normal camera stuff here...  DR, colour science, etc.  No, I didn't mis-number this, it's priority 4 because it's so far away from priority 1 & 2 that there is no Priority 3.

Priority 1 is about feeling comfortable pulling out the camera and shooting, and having it be easy and fast to use.  No fuss.
Priority 2 is about stabilisation and focus, but is also about camera size.  Smaller is better.

This means that iPhone > GX85 > GH5, but unfortunately for priority 4 it's the opposite.

Overriding principle: Use the biggest camera that won't draw too much attention in the situation.

My first challenge was to see if using the iPhone was going to even be feasible.  I mean, the image quality that I saw from it was horrendous.  @mercer laughed when I called it "MAXIMUM AWESOME BRO BANGER FOR THE 'GRAM" but it was potentially the most savage insult I could think of - and the way to make images look as cheap and amateurish as possible.  I own the iPhone 12 mini, which has the normal and wide cameras, but not the tele.  One thing that escaped even my attention was that this has HDR and 10-bit support natively, so it's got at least some potential.

Long story short, I filmed a test scene with all three cameras, as well as a colour chart I made in Resolve and a colour checker and attempted to match them.  Matching to the GH5 was obviously not feasible, but I like the GX85 so I matched to that.  Matching the GH5 to the GX85 was as simple as making a few basic adjustments.  Then came the iPhone.... I tried matching the scenes first, with limited success.  Hues and saturation and Luma behaviour were clearly different and not in easy or nice ways.

That's when I turned to the colour chart and tried to look at what was going on.  At first, the iPhone vector-scope was horrifically tangled, but in a stroke of luck I noticed that the luma curve was pushed up to brighten the mids, and when I brought it back to being linear, the vector-scope immediately fixed itself and was stunningly plain and straight-forwards.  Win!

This is the before/after of the greyscale from the colour chart:

While this isn't a perfect linearisation, check out this before/after of what it does to the vector-scope from my Hue vs Sat test pattern in Resolve....

(I'd removed a few rows in the mildly-saturated region so I could more easily compare between cameras.  This is the GH5 in HLG without a conversion to 709 for comparison)

If I apply a CST to the GH5 (from rec2100 to rec709) and include saturation compression, we get this horrific thing:

I've used that CST technique in the past and it's been fine on real footage, so it's not as bad as it looks.  But the moral of the story is that the iPhone now has easily manageable colours.

I was absolutely stunned when I realised that grading the iPhone footage could be as easy as working with the GH5 footage.  Plus, it's 709 footage in 10-bit, so it should be prone to less banding and artefacts than the GH5 footage, which is 10-bit log.

Next step is to use the colour checker to match the hues and saturation levels of the GX85.

Here are the vector-scopes before matching..  GX85:

GH5:

iPhone with curve:

Obviously the iPhone is nothing like the others, which is what you'd expect - Apple gave it a very distinct look and it's very different to the Panasonics.

I also took shots of the colour checker in -3 stops and +3 stops.  This showed that on the underexposed image (-3 stops) the iPhone is less saturated than the GX85/GH5 and on the overexposed image (+3 stops) the iPhone is more saturated.  Resolve has a curve for this.  I couldn't match it completely as it required a brutal adjustment but I pushed it in the right general direction.

After applying some hue and saturation adjustments, we get quite a good match.  iPhone:

The plots at -3 match well, but the plots at +3 are all over the place, with the GH5 clearly having some rather strange issues.  I'll ignore these for now - the goal is to make the iPhone into a camera that can be used for filming more than just memes, it doesn't have to match perfectly.

I also applied a gaussian blur to the footage to un-do the horrendous sharpening that is applied.

Here are some sample images from a recent trip with the above adjustments applied, without even adjusting each image individually - this is just the starting point.  As I shoot auto-WB I'd be adjusting each image individually, main focusing on skin-tones.  It should also be noted that these adjustments apply equally well to the normal, wide, and selfie cameras, and also apply to the 1080p120 and 1080p240 modes from both the wide and normal cameras.  These shots include a mix of these.

It looks like a real camera to me!  Not the best in the world, but it doesn't make me regret shooting with it, so firmly in the usable category.

[kye]

So, where does this leave us?

Well, the iPhone wins Priority 1 hands down - so many people are vlogging or whatever in public now that it's almost invisible and I feel very little friction in using it.  It's always handy in an easily accessible pocket, and the camera is fast to use from the lock screen.  Once you're in the app it changes FOVs almost instantaneously, and will even change camera while recording, so that's hugely usable too.

It's also pretty amazing at Priority 2.  The stabilisation is simply incredible - I do the ninja walk pretty badly but it manages to give practically perfectly stable images even when walking on stairs etc.  The AF is also fast and reliable, and because it has a very deep DOF it doesn't have the problem of choosing the wrong thing to focus on.  It's got a large screen that is pretty easy to see in full-sun, even if you're trying to get a high or low angle and aren't looking straight at the screen.

[kye]

The GX85 is the next challenge.

For Priority 1 it needs to be kept handy and accessible.  This means either being kept in-the-hand or kept in a pocket - keeping it in a bag adds access time unless the bag is on the front of my chest.  Both of these mean that the rig has to be kept as small as possible, which essentially boils down to lens choice.

The 14mm f2.5 lens is an absolute gem in this regard.  In the 4K mode (which has a 10% crop into the sensor) it's got a 30.8mm equivalent FOV.  I edit and deliver in 1080p, like all sensible people in the real world who haven't confused their ass for their elbow do, and so I can also use the 2x digital crop feature, which gives a 61.6mm equivalent FOV.  These two FOVs are hugely handy for showing people interacting with the environment around them - environmental portraits.

It also has good low-light, good close-focus distance, and has a bit of background defocus if the subject is close.

I have the GX85 configured with back-button focus.  This means that I hold down a button on the back when I want to engage AF and hitting the shutter button doesn't engage it.  This works brilliantly in practice as it means that I can focus once for a scene and then shoot without having to wait for the camera to AF.  I also have the viewfinder set to B&W and have focus peaking enabled in red, making it easily visible.  The histogram is also really handy to know what is going on too.

For Priority 2, this setup works well - the tilt-screen is great, AF is super-fast, IBIS is impressive and very functional, and I find using it in real-world situations to be easy, fast, and very low-friction.  It still gets some attention, but it's not excessive.

I find that for most shots I want to stop down to ensure that everything is in focus.  This is because my work is about the subjects experiencing the location and the interactions that are going on.  A nice portrait with a blurred-beyond-recognition serves very little purpose as it could have been shot anywhere at any time and therefore has no relevance.  This lens can give a satisfying amount of background defocus for mid-shots if required, and especially for macro shots, which are occasionally relevant in an edit.

Here are a few grabs SOOC.  

In grading I would typically lower the shadows to the point where the contrast is consistent (assuming it makes sense for the scene - lots of these have haze which you have to treat differently) and I would even out the levels of saturation etc.  I'd also sharpen or soften images to even out the perceptual sharpness too.

Interestingly enough, most of these images have enough DR, and even have elevated shadows, despite the camera not having a log profile.

These provide a really solid foundation to grade from.

[kye]

The last is the GH5.

Firstly, it's noticeably larger than the GX85, so much so that it's not pocketable (despite what BM might claim) so it's less convenient to transport, and it's not "palm-able" either so you're walking around obviously holding a professionally-sized camera, regardless of what lens you use.

I tried taking it out with the 14mm lens one day instead of the GX85 and I found it more cumbersome to carry, more conspicuous to passers-by, and even subtly more imposing on my family.  When I accidentally found myself carrying it through anti-government protests (long story) I was thankful that the riot police weren't looking for anyone that looked like foreign media.

I typically pair it with manual-focus lenses, and found this to be quite cumbersome compared to the AF of the 14mm lens, but with faster lenses you need to worry about what it will focus on, so it's a whole other ballgame which even the animal-eye-AF cameras get wrong because they don't know which cat you love the most.

While the EVF is reasonable, but not great, the screen is cumbersome to use unless you're shooting from eye-level as you need to flip it out to be able to tilt it, not only making the camera a lot bigger (physically and perceptually to everyone) but also less manoeuvrable and more fragile if you get bumped etc.

Obviously the image is the nicest out of all three cameras, but you really pay for it in size and weight.

For Priority 1 and 2 it's definitely the last-resort, and I'd only pick it over the GX85 in very specific circumstances now.

[kye]

This leaves me with the best setup being either the iPhone by itself, or the iPhone + GX85 which means that the iPhone can do the wide and the GX85 can do the normal and tele FOVs.

The 14mm is good when combined with the 2x digital zoom, but still lacks some flexibility.  When I got home I put on the 12-35mm f2.8 zoom and while it's far more flexible, going a bit wider and hugely longer, the additional size puts it half-way to the size of the GH5 and severely hurts Priority 1 and 2.

This left me with a choice:

1. iPhone wide + GX85 + 14mm
2. iPhone wide + GX85 + 12-32mm f3.5-5.6
3. iPhone wide + GX85 + 12-35mm f2.8
4. iPhone wide + GX85 + wide-zoom like 7-14mm or 8-18mm

The question was if the iPhone wide camera was good enough in low-light.  I reviewed my footage and basically it's good enough indoors in well-lit places like shops or public transit, good enough in well-lit exteriors like markets, but very very borderline in less-well-lit streets at night.

For my purposes I'll live with it for the odd wide and will lean heavily on NR, which will give me the ability to have a longer lens on the GX85 which extends the effective range of the whole setup.  

I did a low-light test of the iPhone, GX85 and GH5 and found that the GH5 and GX85 have about the same low-light performance, iPhone normal camera is about GX85 at f2.8 and the iPhone wide is about GX85 at f8.  Considering that iPhone-wide/f8 was very borderline, if I aim for something like f4 or better then the lens should be fast enough in most situations I find myself.  I don't really find myself wanting shallower DOF than the 14mm f2.5, and could even have less without much downside.

This means that the option of a zoom lens is not out of the question, and the 12-32mm pancake seems a very attractive option.  It's larger than the 14mm when it's on and extended, but when it's off it's basically the same, and it would give a huge bump in functionality.

I took a number of timelapses on the trip too, and having a zoom would make these much easier as well.

I'm contemplating the 45-150mm as well, but we'll see.

[mercer]

Interesting post. I really like the GX85 shots. It's a shame that Panasonic hasn't upgraded the GX85 and G85 with 10bit.

But I think the biggest takeaway is that Canon sucks.

[ac6000cw]

On 4/6/2023 at 10:38 AM, kye said:

I'm contemplating the 45-150mm as well, but we'll see.

The Pana 14-140mm f3.5-5.6 is not much larger/heavier and has better OIS (but it's much more expensive used). It's been my main 'travel' lens for years.

On 4/6/2023 at 12:32 PM, mercer said:

It's a shame that Panasonic hasn't upgraded the GX85 and G85 with 10bit.

Apart from market segmentation and heat issues, I suspect the processing chips used in the lower-end cameras can't support it. They also have major crops in 4k and pixel-binned (probably) FHD, versus uncropped and over-sampled video in G9/GH5/GH6.

[kye]

On 4/6/2023 at 7:32 PM, mercer said:

Interesting post. I really like the GX85 shots. It's a shame that Panasonic hasn't upgraded the GX85 and G85 with 10bit.

It is a shame, but the fact that it's a 709-style profile seriously helps out the 8-bit.  If it was an 8-bit log profile then you'd be stretching it for a 709 grade, but this isn't the case with these cameras.

The iPhone is a 709-style profile and is 10-bit.  That's (very-roughly) equivalent to a 12-bit log profile..  very nice!

The more I use the GX85, the less I find it wanting TBH.

16 hours ago, ac6000cw said:

The Pana 14-140mm f3.5-5.6 is not much larger/heavier and has better OIS (but it's much more expensive used). It's been my main 'travel' lens for years.

In some ways this would be a good option, but I really really appreciate the smaller size of the 14mm and the 12-32mm (at least when it's not in-use).

I found that I would "palm" the camera when carrying it around:

This really helped me not attract un-due attention and protected the camera from bumps in crowds etc, but kept it at the ready when needed.  I could do this with a longer lens but it makes it significantly larger, and makes pocketing it a lot more challenging.

I wouldn't really miss the room range above 32mm, as that's 70.4mm FFequiv when combined with the 2.2x crop-factor of the GX85s 4K mode, when walking around, as I edit in 1080p and could punch-in to 141mm FFequiv.  When I was using the 14mm f2.5 I punched in using the 2x digital-zoom all the time and didn't notice any loss of image quality at all.  Maybe there would be a small loss in low-light but I never noticed it in the final footage where there wasn't a direct A/B.

16 hours ago, ac6000cw said:

Apart from market segmentation and heat issues, I suspect the processing chips used in the lower-end cameras can't support it. They also have major crops in 4k and pixel-binned (probably) FHD, versus uncropped and over-sampled video in G9/GH5/GH6.

The GX85 only has a 1.1x crop into the sensor for the 4K mode, and the 1080p mode doesn't have any crop at all.  I haven't really experimented with the 1080p modes TBH as I wanted the 100Mbps bitrate of the 4K mode.  

I do wonder how much would be supported if we had full access to all the modes supported by the chip.  Obviously the chip in the GX85 can support being given a 4K read-out, can apply a colour profile and do whatever NR and sharpening is done to a 4K file, and can compress a 4K output file at 100Mbps.  To imagine that it might be able to compress a 1080p image at 100Mbps isn't that far-fetched.  Who knows what else might be available on the chip.  ALL-I codecs, 10-bit, etc.

[kye]

I'm into the edit from my last trip.

Putting everything on the timeline results in a 5h22m sequence.  Resolve interprets still images as a single frame, so they aren't padding out the edit time.  My first pass, where I pull in just the good bits of the clips got it down to 1h35m.  

The way I edit is to use markers to separate locations and sequences within a location, which lets me organise the footage (Resolve isn't great if you're shooting on multiple cameras without timecode, so things are often out-of-order).  For example, we went to an aquarium and saw the otters getting fed, saw the sharks getting fed, etc.  In a sense, each of these is like a little story, and for each one I have to establish the scene, then have some sort of progression in the sequence that addresses the "who was there", "what did they do", and "what happened" sort of questions.

I've identified over 40 location markers and over 40 sequence markers within those locations.  Some locations only had one sequence, but others were full day-trips and had a dozen seperate sequences.  That's over 80 stories!

My next steps are to work out which stories get cut completely, then to confirm the overall style of the edit.  My challenge is always how to get from one location to another and establish the change.  I typically shoot lots of clips when walking, on buses / trains / taxis, etc for this purpose.

As it was South Korea, known for K-pop, K-dramas, and the super-Kute things (their reality TV has lots of overlays like question-marks when people are confused or exploding emojis when people are surprised), I'm contemplating a super-cute style with lots of overlays, perhaps using animated title-cards that show where we are and what we're doing.  This would be an alternative style of establishing the location.  I could even have little pics of who was there (sometimes the kids came with us and sometimes not).

Then I'll identify the best shots from each sequence, which other shots are required to tell the story, and then cull the rest. Normally that shrinks the timeline significantly again.

[Davide DB]

On 4/11/2023 at 6:48 AM, kye said:

I really really appreciate the smaller size of the 14mm and the 12-32mm (at least when it's not in-use).

Even the 14-42mm PZ is really compact when not in use and it's very cheap. That small extra tele could be useful too. There's even a cheaper rebranded version on Aliexpress.

[ac6000cw]

1 hour ago, Davide DB said:

Even the 14-42mm PZ is really compact when not in use and it's very cheap. That small extra tele could be useful too. There's even a cheaper rebranded version on Aliexpress.

Yes, it's a nice lens (used it as my main video lens back when I had Pana G6 with a 'power zoom' toggle on the body). It doesn't support dual-IS though. The Oly 14-42 EZ pancake is also 'OK' and has zoom and focus rings instead of toggles, so is a bit nicer to operate. I think build quality is better on the Pana lens though.

On 4/11/2023 at 5:48 AM, kye said:

The GX85 only has a 1.1x crop into the sensor for the 4K mode, and the 1080p mode doesn't have any crop at all.

Oops - yes, quite correct, I was thinking about the mid-range 20MP Pana cameras (GX9, G95, G100 etc.) when I wrote that, which have 1.25x crop in 4k. No excuse really as I own a GX85...

On 4/11/2023 at 5:48 AM, kye said:

I do wonder how much would be supported if we had full access to all the modes supported by the chip.  Obviously the chip in the GX85 can support being given a 4K read-out, can apply a colour profile and do whatever NR and sharpening is done to a 4K file, and can compress a 4K output file at 100Mbps.  To imagine that it might be able to compress a 1080p image at 100Mbps isn't that far-fetched.  Who knows what else might be available on the chip.  ALL-I codecs, 10-bit, etc.

I suspect the processing chips could do 100Mbps All-I 1080p, but having lots of video mode choices in a mid-range camera (aimed at ordinary buyers/users instead of video enthusiasts like us) I think would just be regarded as confusing.

[fuzzynormal]

Seems like you have a process that works and you're content with it.  Very cool.

I'm in the same boat.  Since I'm low low low budget, I shoot what I can with what I can and align as best as possible in post.    I just spent a month and a half filming with a shitty 500mm lens simply because that's all I could afford.  It's not great, but it's not a deal breaker either.  So, off into the field I went and I used it.

And also, combining multiple cam footage with different lenses is not that hard unless, as a filmmaker, you're incredibly intent on having an extremely tight cohesiveness to the IQ --and are desperately striving for seeking out that extra 5% of IQ.

Your test prove that consolidating various footage is viable, and my anecdotal experience follows.

I know a lot of us here really want to find the perfect recipe for all of the above, throw in some secret sauce to make it all work, and that'll make us sit back in the editing seat and go "golly, doesn't that look wonderful!"

However, since consumer IQ tech is pretty damn good now, as a documentarian my goal isn't about the tech, more often it's simply get the shot that tells the story, then tell that story.

These days, when it comes to IQ, I worry much much more about the floor than the cieling.

[kye]

11 hours ago, fuzzynormal said:

Seems like you have a process that works and you're content with it.  Very cool.

This is mostly true.  I subscribed to the 'subtractive' model of editing where you start with all your footage and then remove the parts you don't want, making several passes, and then ending in a slight additive process where I pull in the 'in-between' shots that allow it to be a cohesive edit.  I'm aware there's also the 'additive' model where you just pull in the bits you want and don't bother making passes.  Considering my shooting ratios (the latest project will be 2000 shots / 5h22m likely to go down to something like 240 shots / 12m - either 8:1 or 27:1 depending on how you look at it) and the fact I edit almost completely linearly (in chronological order of filming) I might be better off with an additive process instead.

I've also just found a solution to a major editing challenge, and am gradually working through the process of understanding how I'll include it in my editing style.

11 hours ago, fuzzynormal said:

I'm in the same boat.  Since I'm low low low budget, I shoot what I can with what I can and align as best as possible in post.    I just spent a month and a half filming with a shitty 500mm lens simply because that's all I could afford.  It's not great, but it's not a deal breaker either.  So, off into the field I went and I used it.

And also, combining multiple cam footage with different lenses is not that hard unless, as a filmmaker, you're incredibly intent on having an extremely tight cohesiveness to the IQ --and are desperately striving for seeking out that extra 5% of IQ.

Your test prove that consolidating various footage is viable, and my anecdotal experience follows.

If you're not deeply attuned to the subtleties of the image (as I know some people are) then its quite feasible to match footage across cameras, and even do image manipulation in post to emulate various lenses, at least to the extent that it would be visible in an edit where there are no side-by-side comparisons.  The fact that a scene can be edited together from multiple angles that were lit differently and shot with different focal lengths from different distances is a statement about how much we can tolerate in terms of things not matching completely.

11 hours ago, fuzzynormal said:

I know a lot of us here really want to find the perfect recipe for all of the above, throw in some secret sauce to make it all work, and that'll make us sit back in the editing seat and go "golly, doesn't that look wonderful!"

However, since consumer IQ tech is pretty damn good now, as a documentarian my goal isn't about the tech, more often it's simply get the shot that tells the story, then tell that story.

I had a transformative experience when I started breaking down edits from award-winning travel show Parts Unknown (as that was what most closely matched the subject matter and shooting style I have).  I discovered a huge number of things, with some key ones being:

• Prime (which streams high-quality enough 1080p that grain is nicely rendered) showed clearly that the lenses they used on many episodes aren't even sharp to 1080p resolution, having visible vintage lens aberrations like CA etc
• They film lots of b-roll in high frame rates and often use it in the edit at normal speed (real life speed) which means it doesn't have a 180-shutter, and yet it still wins awards - even for cinematography
• Many external shots have digitally clipped skies
• Most shots are nice but not amazing, and many of them were compositions that I get when I film
• The colour grading is normally very nice and the image is obviously from high-quality cameras

This made me realise that the magic was in their editing.  When I pulled that apart I found all sorts of interesting sequences and especially use of music etc.  But what was most revealing was when I then pulled apart a bunch of edits from "cinematic" travel YouTube channels and discovered that while the images looked better, their editing was so boring that any real comparison was simply useless.  

This was when I realised that camera YouTube had subconsciously taught me that the magic of film-making was 90% image and 10% everything else, and that this philosophy fuels the endless technical debates about how people should spend their yearly $10K investment in camera bodies.  Now I understand that film-making is barely 10% image, and that, to paraphrase a well-known quote, if people are looking at the image quality of your edit then your film is crap. 

When you combine this concept with how much is possible in post, I think people spending dozens/hundreds of hours working to earn money to trying to buy the image they like, and spending dozens/hundreds of hours online talking about cameras without even taking a few hours to learn basic colour grading techniques is just baffling.  It's like buying new shoes every day because you refuse to learn how to tie and untie the laces and the shop does that for you when you buy some.

11 hours ago, fuzzynormal said:

These days, when it comes to IQ, I worry much much more about the floor than the cieling.

Absolutely - that's a great way of putting it!

My consideration is now what is 'usable', with the iPhone wide angle low-light performance being one of the only sub-par elements in my setup, and, of course, why IQ is Priority 4.

[kye]

I wanted to expand on some things but didn't want to clog up the other threads, so am back to this thread.

I previously said this:

On 3/28/2024 at 8:04 PM, kye said:

My current thinking is this:

Full tourist mode is the GX85 and OG BM camera setup, and is where I don't care about weight and want the ultimate options:

• GX85 + 14-140mm for walking-around during the day
• GX85 + 12-35mm for walking-around well-lit places at night
• GX85 + primes for less-well-lit places at night
• BM for high-DR situations where I have time to manually expose and shoot (lookouts at sunset, for example..  these are strangely frequent on my holidays 🙂 )
• Equipment:
  • GX85 + 14-140mm
  • P2K or M2K with 12-35mm F2.8
  • Primes: 7.5mm F2, 14mm F2.5, 17mm F1.4, 58mm F2 with SB
• This all fits into a camera insert inside a generic backpack

Stealth tourist mode is the GX85 and is where I want to be stealthy and shoot quickly:

• GX85 + 14-140mm for walking-around during the day
• GX85 + 12-35mm for walking-around well-lit places at night
• This fits into a sling bag, which are now fashionable, how convenient!

Non-tourist mode is the GX85 and 12-32mm zoom (and maybe the 14mm F2.5 if it'll get dark) slipped into a jacket pocket.  If I'm travelling this is what I would take if I was leaving the hotel but didn't want to take a bag.

The Full Tourist Mode includes "GX85 + 12-35mm for walking-around well-lit places at night" which I wanted to test, so last night while the wife was at a concert, I wandered around town with the GX85 and 12-35mm in aperture priority with auto-SS and auto-ISO and took a bunch of test shots.  Everything was hand-held, and I'd assume it's at 360 shutter the whole time, which I don't mind, especially considering the Dual IS means there isn't a lot of hand-shake.

The TLDR of it is that:

• I think the theory works - the GX85 + 12-35mm at F2.8 is good enough for "well-lit" places at night
• Most of the places I wanted to shoot were ISO 400-1600 and the noise is OK from the GX85 at those levels
• The DR is a challenge at night, which is something I hadn't considered - by the time that you don't clip the highlights you're making the shadows a lot darker than they appear in real life

Here's some shots SOOC.  Where there are two shots, it's because it wanted to expose in a way that clipped a bunch of stuff and so the darker shot is me reducing the exposure with Exposure Compensation to the point that the highlights aren't clipped.  
As these are frame grabs from VLC, the highlights might be more clipped than they would be in the grade because the GX85 clips at 105% so VLC might be cutting off that top 5%.

Here's what I mean:

My recollection was that the ambient light on the street was around or below the top one, but the signs and lights in the shops looked even more dynamic and sparkly and sharp than the bottom image shows.

Same with these:

The mall also wasn't this well-lit:

On 3/28/2024 at 8:04 PM, kye said:

Full tourist mode is the GX85 and OG BM camera setup, and is where I don't care about weight and want the ultimate options:

• GX85 + 14-140mm for walking-around during the day
• GX85 + 12-35mm for walking-around well-lit places at night
• GX85 + primes for less-well-lit places at night
• BM for high-DR situations where I have time to manually expose and shoot (lookouts at sunset, for example..  these are strangely frequent on my holidays 🙂 )
• Equipment:
  • GX85 + 14-140mm
  • P2K or M2K with 12-35mm F2.8
  • Primes: 7.5mm F2, 14mm F2.5, 17mm F1.4, 58mm F2 with SB
• This all fits into a camera insert inside a generic backpack

I also said that I could fit the GX85, BMMCC rig and the 14-140mm, 12-35/2.8, 7.5mm F2, 14mm F2.5, 17mm F1.4, 58mm F2 with SB lenses into a camera insert.  Here's that:

The GX85 and zoom goes into one side, and the primes go into the other:

Obviously I still need to DIY some protectors to keep the lenses protected from each other.

Then, with the BMMCC monitor folded all the way forwards it can go pretty flat:

That has the 14-42mm kit lens on it as a stand-in for the 14-140mm that I haven't bought yet.  The 14-140 is a bit thicker, so the monitor won't lie as flat, but the difference isn't that much.

The BMMCC rig goes on top of the lenses:

...and we're done.  I normally carry around a blower and a small lens brush (makeup brushes from the supermarket are really soft and have short handles to fit into camera bags) so I might need to finesse this a bit, but that's the basic concept.

On 3/28/2024 at 8:04 PM, kye said:

Stealth tourist mode is the GX85 and is where I want to be stealthy and shoot quickly:

• GX85 + 14-140mm for walking-around during the day
• GX85 + 12-35mm for walking-around well-lit places at night
• This fits into a sling bag, which are now fashionable, how convenient!

These easily fit into this ebay-special sling bag I got:

I won't be winning any fashion competitions with this, and it's larger than I'd like, but I'm on the lookout for one smaller and more appealing.  It's not that I'm a fashionable person (ha! what a concept) but if you want to blend in with the locals in fashionable places then you have to look the part.

Anyway, work still to do, but progress is positive.

[ac6000cw]

5 hours ago, kye said:

The TLDR of it is that:

• I think the theory works - the GX85 + 12-35mm at F2.8 is good enough for "well-lit" places at night
• Most of the places I wanted to shoot were ISO 400-1600 and the noise is OK from the GX85 at those levels
• The DR is a challenge at night, which is something I hadn't considered - by the time that you don't clip the highlights you're making the shadows a lot darker than they appear in real life

Here's some shots SOOC.  Where there are two shots, it's because it wanted to expose in a way that clipped a bunch of stuff and so the darker shot is me reducing the exposure with Exposure Compensation to the point that the highlights aren't clipped.  
As these are frame grabs from VLC, the highlights might be more clipped than they would be in the grade because the GX85 clips at 105% so VLC might be cutting off that top 5%.

For me, out of the first two shots, artistically the upper/brighter shot is the better one - the 'bright lights' are properly bright (even if some are a bit burnt out) and the people are visible enough to add more interest to the scene. The lower/darker shot personally I think looks too dull (and less attractive as a result), even though I suspect it's a more accurate representation of the scene.

As ever, it depends on how 'accurate' or how 'attractive' you want the shots to be (after they've been tweaked/graded/edited) i.e. the artistic choices...

[eatstoomuchjam]

Did you ever consider adding (or replacing one of the other lenses) with the Olympus 75/1.8?  It's also pretty tiny, pretty cheap, and IMO it's one of the greatest lenses ever made for Micro 4/3.

[eatstoomuchjam]

Oh, that and why prefer the 7Artisans 17/1.4 over the smaller (and pretty great) Panasonic 20/1.7?

[kye]

15 hours ago, ac6000cw said:

For me, out of the first two shots, artistically the upper/brighter shot is the better one - the 'bright lights' are properly bright (even if some are a bit burnt out) and the people are visible enough to add more interest to the scene. The lower/darker shot personally I think looks too dull (and less attractive as a result), even though I suspect it's a more accurate representation of the scene.

As ever, it depends on how 'accurate' or how 'attractive' you want the shots to be (after they've been tweaked/graded/edited) i.e. the artistic choices...

These images are really about what the camera captured - with colour grading there is a huge latitude to change things in post, which is where the taste and preference comes in.  The challenge is what you have to work with in the first place.

In the thread about colour grading 8-bit footage I showed that there is huge latitude to push things around in post - to the point that you can bring things up by 4-stops without ruining the image:

Of course, if your highlights are clipped then there's basically zero ability to change it in post!

After reflecting on things after posting the images above (and also of reviewing a bunch of other shots I didn't post) I think that the images look and feel like reality somewhere in-between the two extremes I posted.  I can't recall how much I had to stop down to pull the highlights back into range, and it probably varied depending on the situation, but it might be a case of maybe shooting a stop under what the GX85 wants to shoot at that might give a nice middle ground.

I must admit that the highlight blooming on the lens is quite nice, giving a sort-of magical feel to the lighting, while not making the image to surreal in daylight situations or being impossible to use with strong light-sources.

14 hours ago, eatstoomuchjam said:

Did you ever consider adding (or replacing one of the other lenses) with the Olympus 75/1.8?  It's also pretty tiny, pretty cheap, and IMO it's one of the greatest lenses ever made for Micro 4/3.

14 hours ago, eatstoomuchjam said:

Oh, that and why prefer the 7Artisans 17/1.4 over the smaller (and pretty great) Panasonic 20/1.7?

There's always more lenses to add to the kit!!

I used to shoot exclusively with manual primes and I worked out that for the travel work I do that the following were the right lengths:

16mm equivalent FOV:
This is used for WOW!! shots..  buildings and landscapes. It also does crowded interiors too.  Going wider than this makes things look too distorted, and going longer doesn't give that WOW!! factor.

I forgot to mention it in my equipment summary, but I have the iPhone for the super-wides during the day (it isn't good enough at night unfortunately) so that compliments the walk-around zoom lenses.

35mm equivalent FOV:

This is for environmental portraits, which is essentially the main subject of my work.  I shoot mostly travel with family and friends and the subject of these films is the people interacting with the environment.  Combined with the distance I am normally standing away from the subject, this is about the right FOV.  

If I went wider then the people are smaller in the frame than I'd like and if you get closer to compensate then lens distortion becomes undesirable.

It's also good for landscape shots too:

80mm equivalent FOV:

For things that are further away.  When I first started travelling with the 3-prime setup I took the Helios 58mm F2 with a non-speed-boosted adapter but found the 116mm equivalent FOV on the GH5 to be too long - it was too much of a gap and I found myself swapping from the 17.5mm prime to the Helios and then swapping back.  The gap between the 35mm FOV and 116mm is a 3.3x increase in focal length.  I have since bought a M42-M43 speed booster which remedies this.

So now that I've moved to the GX85 and the crop factor in 4K is 2.2x, the FOV of the lenses is:

• Laowa 7.5mm F2 is 16.2mm FOV
• TTartisans 17mm F1.4 is 37.4mm FOV (2.3x previous)
• Helios 58mm F2 with SB is 90.6mm FOV (2.4x previous)

The 20mm F1.7 would be a 44mm FOV, which is a bit too tight in my experience.  

Of course, if I use the 14-140mm F3.5-5.6 zoom during the day and 12-35mm F2.8 for walking around "well-lit" places at night, the only use for the primes I would have is for very low light situations where I can shoot slower and take the time to MF properly.  In terms of what those situations demand, I'm not really all that sure.  

The Olympus 75mm 1.8 might be a useful addition for super-tight low-light shots, although it might be a pretty specific lens.  I want to capture what I see, so it's not a case of going out with a lens and finding the frame, it's a case of going out and experiencing things and capturing the most that I can.  What use cases would the 75mm f1.8 capture that the 14-140mm F3.5-5.6 couldn't?

If I had the budget, the 20mm F1.7 is so small that I could easily fit it into my bag, but the 15mm F1.7 is also a lot smaller than the TTartisans 17mm F1.4 as well that either one would work just as well, except that the 15mm is wider.  The 4K from the GX85 can always be cropped slightly without much impact to image quality too.  

I already own the 17mm F1.4 so it's not a case of having a preference, although the 20/1.7 is 2.7x the price of the 17/1.4 and the 15/1.7 is 4.4x - so budget might have had something to do with it!  I was using only fast manual primes, but now I've come to value AF over DOF, so the primes are relegated to low-light duties, however if I could get fast AF primes then that would be desirable too.  It's probably something I'd look at upgrading to over time, assuming I stay in the MFT system, of course.

[eatstoomuchjam]

Gotcha - if you're more comfortable with the 35mm equivalent FOV, a 17ish vs a 20ish mm m43 lens makes total sense - and if you value AF, it's not your best option by far.  From what I remember, it's a bit noisy and slow.  I'm sure the Olympus 17 or the Panasonic 15 is faster.  For me, the 14/2.5 and the 20/1.7 were no brains to keep in my kit when I was still using M43 because they were just so tiny.  It felt silly not to bring them, especially since they're both really decent lenses.

For me, the 75/1.8 was always useful for either portraits (though one has to stand a little further away than I like) or for landscapes (as I age, I like telephoto landscapes more and more - just choose the little bit of the scene that I want).  The fast aperture let me get pretty sharp stuff even when shooting from a moving car or train without having to crank the ISO on a smaller sensor.  At this point, I have a Summicron-M 90/2 ASPH so unless I'd need autofocus, I'd just prefer it for that sort of landscape shot (and on FF, it's a really nice length for portraits, to boot).

Anyway, if you don't mind the aperture limitations of the 14-140 when it's racked to 75mm, you're definitely set there.

For me, my travel kit contains a few redundant focal length primes - though they're less for the faster aperture now and more for being smaller/lighter/nondescript.  Does 1 extra stop on the Fujinon 63/2.8 make any substantial difference than the 32-64/4 racked out?  Not really.  Is there any appreciable difference in quality on the prime?  Not really.  Are people more likely to ignore me when it's on there?  Yes.

[kye]

5 hours ago, eatstoomuchjam said:

Gotcha - if you're more comfortable with the 35mm equivalent FOV, a 17ish vs a 20ish mm m43 lens makes total sense - and if you value AF, it's not your best option by far.  From what I remember, it's a bit noisy and slow.  I'm sure the Olympus 17 or the Panasonic 15 is faster.  For me, the 14/2.5 and the 20/1.7 were no brains to keep in my kit when I was still using M43 because they were just so tiny.  It felt silly not to bring them, especially since they're both really decent lenses.

The size and cost of the 14/2.5 and 20/1.7 sure make them compelling lenses to own and carry around with you, that's for sure.  I don't know what the AF is like on the 20mm but all I do is frame up a composition, do a single AF-S via a custom button, then hit record and maintain the focus distance during the clip.  I shoot short clips for the edit, so I don't need or want AF-C.  From that perspective the 20mm might be just fine.  The alternative is manually focusing with peaking, which would likely take longer than most AF.  

Speaking of AF speed, I read something the other day - it said that Panasonics DFD sped up their CDAF, and I realised that I never see a CDAF Panasonic camera doing that thing where the focus racks the whole way to one end and then the whole way to the other end before acquiring focus, it just seems to do a quick jitter and it's done.  I never thought about that being DFD but I guess it is.

5 hours ago, eatstoomuchjam said:

For me, the 75/1.8 was always useful for either portraits (though one has to stand a little further away than I like) or for landscapes (as I age, I like telephoto landscapes more and more - just choose the little bit of the scene that I want).  The fast aperture let me get pretty sharp stuff even when shooting from a moving car or train without having to crank the ISO on a smaller sensor.  At this point, I have a Summicron-M 90/2 ASPH so unless I'd need autofocus, I'd just prefer it for that sort of landscape shot (and on FF, it's a really nice length for portraits, to boot).

I watched a lot of landscape photographer YT (Thomas Heaton etc) and they made a strong case for landscapes being shot with wide lenses and ultra-long lenses.  Some of those shots that show just the jagged peak of the mountain or the lone tree or castle on a distant hilltop can be the most stunning.

5 hours ago, eatstoomuchjam said:

Anyway, if you don't mind the aperture limitations of the 14-140 when it's racked to 75mm, you're definitely set there.

Some time ago I realised that DOF depends not only on aperture but also on focal length, so although a variable zoom gets slower as it gets longer (making the DOF deeper), it's also getting longer as it's getting longer (making the DOF shallower) and so I did a bunch of math to calculate DOF of the same composition.  I posted the results in some other thread somewhere here, but the summary is that a lot of variable aperture zooms are almost constant DOF lenses, when taking the same composition (ie, if you double the focal length then you'd be twice as far away for the same composition).

Here's the table of the 14-140mm lens.  The "Mid DoF" column is the DoF of a mid portrait shot (chest and up) and the "Close DoF" is just top of shoulders and up:

It's not constant DoF but it's pretty close.  

I then realised that for my environmental portraits, where I want the subject in focus but the background should at least be recognisable, I didn't want something that just had the subject floating in a sea of mush.  Also, nailing focus is more important to me than shallower DoF, and if the focus isn't going to get it perfect every time, and also pick the right focus subject each time, or if there are two people next to each other but slightly different distances from the camera because I'm not standing exactly 90-degrees to the line between them, then I'd rather the DoF be a few meters rather than the shot be missed.

The other reason to have a fast lens is the low-light capabilities.  FF obviously has the advantage because, all else being equal it gets 4x the amount of light onto the sensor, but this has to be balanced against the DoF which will also be radically shallower for the same T-stop.  So if you don't want to shoot with a razor-thin DoF in low-light then you have to stop down.  I find that in practice this would level the playing field in many compositions.  Not all of them of course, and the seemingly greater investment in sensor technology from Sony in the larger sensors is also a factor, but it makes the topic more complex and far less one-sided than it might first appear.

5 hours ago, eatstoomuchjam said:

For me, my travel kit contains a few redundant focal length primes - though they're less for the faster aperture now and more for being smaller/lighter/nondescript.  Does 1 extra stop on the Fujinon 63/2.8 make any substantial difference than the 32-64/4 racked out?  Not really.  Is there any appreciable difference in quality on the prime?  Not really.  Are people more likely to ignore me when it's on there?  Yes.

Yeah, it's definitely a case of "get the shot" first, "make the scene better by not making everyone uncomfortable" second, and "have a rig with a great image quality" third.