Horrible Day W/ The LS300

[Lee Powell]

In a severely backlit situation like this it might turn out better to shoot in Rec 709, using the LS300 gamma and black level controls to boost the shadows. You could also adjust the knee to give you a reasonable amount of highlight compression. I used to do this back before J-Log1 was released, and found it worked well with natural lighting. With stage lights, gamma boosts would make LS300 oversaturate on pure colors but that shouldn't be an issue in broad daylight.

[Alex Humphrey]

follow up dumb question. are your two pics from each camera or was it from your point and shoot to simply show the lighting and setup? If so can you upload a still from each camera?

[Aaron Jones Sr.]

Originally Posted by Alex Humphrey

I have been experimenting a bit myself with various degrees of success but have not had the time to do some quality tests.

here is a link to a good article on the subject if you haven't read it https://www.thebroadcastbridge.com/c...h-j-log-part-2

I myself picked up the Lumu light meter ($69 ball that plugs into an iPod touch or older iPhone) and running Cine Meter II from the App Store for aI think $29. It's a good incident light meter on the cheap that also has filter factor adjustments that you can set to match the built in ND on the JVC. I've only used it a little and got much more consistent results though I have not fully put it through it's paces. A Sekonic Cine Meter would be far better for $400-$800 as long as you have built in filter factors to make my life simpler. (avoiding the iPod going into sleep mode or staying on and running out of juice in an hour or so is what I mean by better)

That being said I think your shooting was probably just fine for Rec709, but for Jlog the mid tones should be higher. I also shoot at ISO400 not the recommended 800 since more grain shows up faster at 800ISO.

I expect once I or anyone else establishes a consistent shooting and post style that the light meter will tell you or me how we shoot and post work best.. then go backwards and reset the our high and low levels in camera to match a majority of our shooting. But if you can lay your hands on an incident light meter, even a photo one (no built in filter factors usually like the Sekonic 308) for cheap or borrow since you will want a better Sekonic going forward in career. and establish a more consistent baseline.

If the JVC shot raw or at least 10 bit Pro Res it wouldn't be as a big of deal, but it is.. especially if you are going to be playing with Red. I have to do the same thing with my friends Red Scarlet. So I have to figure out my method as well to be B camera to a Red Scarlet... Good luck and let us know how it continues.

Thank you for the added information with the light meter it was very well received. I will take a look at those link asap.

Originally Posted by Alex Humphrey

Post Thought: Maybe I'm not seeing what your seeing, but I don't think your jpg images looked terrible.. I sort of think you got as good as you could get without a big bounce card or a fill light. I still say all of my above was valid, but I expect a gold bounce card to bounce some life into the face would have given you the latitude you wanted. That being said it didn't look terrible to me. I can see how the JVC had more noise in mid tones than the Red did... IF you had over exposed by probably 1 stop or at least 1/2 stop you may not have gotten as much noise, but I don't know if you could recover much of the sky.

I'm using FCPX (older version) and color Finale plug in that I love.. though latest version of FCPX (need to update laptop and FCPX and lose Color Finale till they do a patch) gives me a lot of pushing and pulling sky and ground. here is a link to my outdated tests to see how much I could push and pull with Color Finale if you are not using it. here is a link to 1 shot with adjusting in post the same image as to how far FCPX with Color Finale could push from a shot with original footage in box 1. http://www.dvinfo.net/forum/attachme...10minutesc.jpg IF I go outside and repeat some testing with my light meter and compare with my zebra I KNOW I could do a LOT better..

But yes... I think from what I saw you can do more Post work to recover the sky and lift a little of the face, but I would never say it was horrible... honestly it's better than a lot of I have seen on HBO and Showtime let alone network... and I do think if you had a reflector or at least fill light your shot would be easier. Unless you are seeing something I'm not.

But yes, hit us back with some updates when you have time to test. I'm out with an injury so I can't walk outside with a camera for another few days at least.

Sorry the those pics were just to show the environment of the light. I took those with my smartphone. They show where the light was coming from and how I was positioned in the shoot.

I will definitely post my findings here as I keep working with the sunny day scenario. I do really good with this cam indoors, and in the shadows outside, but in the sunshine I suck. Either I have to blow out the sky and expose the talent or expose the sky and underexpose the talent. Trying to work on a happy medium. I know I have to introduce more light to the talent to get both but was trying to work it out.

[Steve Ritchie]

I forgot to mention, the native ISO for the LS300 may be 400 when using Rec709, but it is 800 when using JLOG1. If you shoot at ISO 400 in JLOG 1, it clips highlights early. You lose dynamic range.

I have tested the dynamic range in JLOG1 at ISO 400 vs ISO 800. ISO 800 has higher dynamic range. I tested shooting at 400 and lifting shadows, and was unable to produce a cleaner image than shooting at native 800 iso.

[Lee Powell]

Originally Posted by Steve Ritchie

I forgot to mention, the native ISO for the LS300 may be 400 when using Rec709, but it is 800 when using JLOG1. If you shoot at ISO 400 in JLOG 1, it clips highlights early. You lose dynamic range.

Yes, JVC simply doubled the numbers in J-Log1 mode. J-Log1 ISO 400 is just a digitally scaled down copy of ISO 800. Hence

J-Log1 ISO 800 = Rec 709 ISO 400
J-Log1 800% Reflectance = Rec 709 400% Reflectance

Inflated LOG ISO numbers have become an established industry practice. JVC followed Sony and Panasonic's lead in their LOG implementations, which are even more inflated. Sony and Panasonic LOG footage has oppressively low contrast and saturation and often looks bleached out, even after grading. J-Log1 is comparatively mild and easier to grade, much better suited to an 8-bit encoder.

[Aaron Jones Sr.]

Originally Posted by Alex Humphrey

follow up dumb question. are your two pics from each camera or was it from your point and shoot to simply show the lighting and setup? If so can you upload a still from each camera?

Yes, just to show my point of view. they were not screen grabs.

How ever here are screen grabs from that session:

First one sun behind clouds and second one sun beaming out. Both images are screen grabs from the LS300 cam after a slight color correction.

[Alex Humphrey]

first thoughts. Especially for darker skin tones I think it's vital to expose Jlog as 400 (0 gain) and slightly over expose for the scene... (old school film expose for the shadows vs video which is to protect the highlights) I can bring down the exposure a lot and retrieve detail in the sky with curves in my NLE... (via plug in for FCPX or latest FCPX) as well as any other modern NLE.

It is true that we may lose 1 f-stop of dynamic range, but we gain a cleaner mid tones and darker tones without noise, and thats true with all cameras. It also means that with JLOG and exposing just a little over that it will always mean some post work.. though I set up my custom filter preset and drag and drop on timeline and it's accomplishing 3 or 4 tasks at once that have me ballpark about right and quickly.

Also the latest 422 update that I have yet to preform myself might help.

I don't have enough first hand use with the better Cine incident/spot meters from Sekonic, I had the 308S 308C that burned up in the fire... and they annoyed me since they did not have ND filter factors in them.. so I would get a reading like f22 then I would have to add ND on camera and figure out what the meter would be telling me.. (PAIN) that's why I like the newest light meters that have filter info in the meter just like that new iPod app (with Lumu ball plug in meter) If I don't like the f-stop reading in the meter (f22 for instance) I can click on the app, adjust the ND filter factor from zero up to 1/1024 (0, 1/2, 1/4, 1/8, 1/16 etc) that matches our JVC's. and pick our aperture on the lens to match. Then look at skin tones and and probably set custom zebras to buzz on the skin tones that best match our shooting styles. That would be 1 method.

But regardless a good Cine meter is $400 and up to $800 but probably well worth it, and good for all cameras now or in the future. In the meantime I'll work with old iPod touch as a stopgap.

[Alex Humphrey]

There is also a LOT of data there that you can still adjust and push and pull, though going too far at 420 it will start to fall apart.. the latest update 422 I hope will help a lot.. with lifting the shadows and burning in the sky with applying a S curve to the exposure. I'm not saying it will compete with 444 Raw, but I'm amazed at what I can squeeze out of the JVC at 420 that I would not have thought possible.

What software are you using?

[Lee Powell]

Originally Posted by Alex Humphrey

Especially for darker skin tones I think it's vital to expose Jlog as 400 (0 gain) and slightly over expose for the scene.

Read the posts above. In J-Log1 mode, base ISO is 800, and delivers maximum dynamic range. J-Log1 ISO 400 reduces your SNR without any image quality benefit.

[Alex Humphrey]

I have and I am aware of the added DR JVC indicates for ISO 800 (6dB gain) over ISO 400 (0dB gain) and I find the image has too much digital noise in the mid tones and shadows. Indoor or night time I switch back to Rec 709 and do a minor Black Stretch and a fairly aggressive Knee setting. Daylight outside over exposing and recovering highlights J-Log at 0dB or 6dB works as well, but I find the noise in shadows and edge of shadow to mid tones still to jarring at 6dBs and digital looking vs say the Digital Bolex that has perhaps more video noise but looked more organic and less jarringly digital. For me the JVC LS-300 has to mix with our Red Scarlet and I find the digital noise in the shadows is more noticeable than a little less dynamic range than working with Red.

[Lee Powell]

The ISO 800 noise is still there in ISO 400, it's just digitally scaled down so you don't notice it as much. J-Log1 footage always needs grading, and adjusting shadow noise in post is just a routine part of the process. It's better to have more shadow detail to work with rather than chopping it in half.

[John Vincent]

Good thread.... couple of questions:

1. How would the log users grade it? Earlier reviews of the camera/log weren't very favorable.

2. A possible answer are monitors that show you what the footage will look like after grading, or even more interesting (to me anyway) is the monitor Kai was using in his latest GH5 video, which could actually bake-in a log.

3. Am I the only one on earth who doesn't shoot log?

Thanks guys -

[Aaron Jones Sr.]

Originally Posted by John Vincent

Good thread.... couple of questions:

1. How would the log users grade it? Earlier reviews of the camera/log weren't very favorable.

JVC GY-LS300 Working with J-Log - Part 2

Also See This Post for the Leeming Lut by Nick Haman

I bought this lut and it was a godsend.

Originally Posted by John Vincent

2. A possible answer are monitors that show you what the footage will look like after grading, or even more interesting (to me anyway) is the monitor Kai was using in his latest GH5 video, which could actually bake-in a log.

Hmmm...

Originally Posted by John Vincent

3. Am I the only one on earth who doesn't shoot log?

Nope! I never shot log professionally until I got this camera. I'm trying to learn the practice of it so that I can use it on my paid gigs. I now shoot 709 at the moment as I keep trying to fine tune my use with the log. Right now I hit and miss. It is not the camera or the log but me, lol!

I had the GH5 and sold it 2 weeks after having it. It will boil down to what you are looking for in a camera and how it will compliment your style of shooting. I also bought the Zhiyun Crane to use with it. What drew me to the GH5 was the ability to set one of the profiles to auto exposure and auto focus. I could be inside and it would expose properly and soon as I go outside it would adjust with the transition time you set. For me at the moment that was big because even though I shoot full manual, I only did so because the AI of the camera would adjust my shots mid stride soon as there was a little light difference. The GH5 for me started just to be a B-Roll type of camera or a B cam because of the form factor. I prefer a beefier cam that I can sink my teeth into and I have an issue trying to use DSLR type cameras for video. It is just my taste. So you have to decided what works well with you and your style of shooting. In the end they both shoot well. The question is can you live with the limitations that each one has.

[Ed Kishel]

first time I shot J-Log was outside- I had the zebras set to 100 and in full sun they would rarely trigger. The LCD looked VERY overexposed. At ISO 400 and without a cloud in the sky I was really nervous I was messing something up. Lens wide open and still no zebras, even when I pointed it up at the sky. Yikes! If going by the LCD the picture looked like hell. My gut told me to kick on an ND or lower the ISO- but I trusted that the zebras weren't lying. When I got back and graded it- it looked awesome. The NLE scopes showed no clipping. LOG rocks- just trust your other exposure tools and only use the LCD from framing and focus.

[Alex Humphrey]

I finally went outside today with my leg in a cast today and dragged my LS-300 along with my iPod touch and new Lumu light meter (plugs into headphone jack of older iPhones and current iPod Touch with Cine II app) Also with my grey fabric expandable card with my Zebras high of 40% and low of 35% to compare where my meter should be saying middle grey should be. The camera just went up in quality in my eyes simply due to going old school and getting a better reading of the light and a better recording. The noise in my dark - mid tones in J-Log after grading are simply virtually gone at 400/800 and barely noticeable at 1200 ISO. I was trying to ride the exposure to the right in the histogram and keeping detail in clouds etc but I was hit and miss.
So here are my thoughts for anyone following along, and I'll upload pics in a few days after I repeat my testing to prove to myself that I just didn't get lucky today.

1. Incident Cine light meter with filter factors (ND readout and adjustment on the meter so you don't have to do the math in your head) is a must. The Lumu and Cine II app if you have an old iPod touch or older iPhone laying around is great... but with the iPhone/iPod chewing up battery life quickly and going in and out of the app, it's not a long term solution and a quality Sikonic Cine meter above a 308c by a couple steps probably in the $400-$700 to last me the rest of my cinematography/photography life is in order. in the short term the $59 ball and $29 app help out a lot.

2. Using a grey card (fabric pocket spring unfolding loop type, similar to larger reflectors etc) and setting Zebras to Low 35% and high 40% putting in the sun and rolling the aperture till most of the card is in Zebra gives almost the identical reading as my Lumu - iPod Cine AppII meter combination does. In the pas I have tried to ride the histogram to the right as well as using the built in spot meter and spot checking the brightest subject and my results are sometimes great, sometimes terrible.. but hardly consistent.

3. Putting the same grey card in the shade having set the aperture already to match the above I no longer get the hideous video noise the shadows as I was when I was riding the histogram to the right or even the more erratic build in spot meter (that can be useful as a backup but not a primary exposure indicator for me now) how much off was I before? If memory serves maybe only 2 stops above or below what I was metering with Lumu, though I feel it must have been more like 3 stops, just enough to really start breaking the curve of the exposure after applying my grade.

4. Color Finale plug in for FCPX simply rocks. FCPX should have always had these features (FCP7 did) I think the latest FCPX has regained some of the features Color Finale offers, but you get more than simple curves and LUT and sample grades with Color Finale. That being said dropping the Color Finale plug in on the timeline and applying the JVC cube LUT to it makes a quick simple contrasty punchy 709 looking video from my J-Log footage that has a lot more dynamic range than shooting 709 in the first place. That being said I still find it garish and too contrasty. I can drop the effect down in FCPX however the other Luts/looks/grades that came with Color Finale are nice. the 3 Strip color and 20-30 Kodak/Fuji film simulations in their extra downloadable content for the plug in I find very nice. In fact. just add a little sharpening to the J-Log footage and 1 of the Luts to get me close to what I have in mind, and it starts looking close to some of the footage I have seen recently from Blackmagic's cameras have done.

If I don't look to close at the still frames I don't see too much of the 8 bit macro blocking, it's still there but radically reduced.

5. Exposing for the shadows on Aaron Jones shot that started this thread might be how I might have tried to shoot it. With the ball meter pointing back at the lens and being mostly in shade would been 2 or 3 stops below if it was placed in full sunlight or, obviously turning sideways and splitting the difference. If you have never used an incident meter you face the ball towards the lens hopefully in same similar light as your subject and pull a reading. So backlit you are exposing for the shadows for instance since most or all the white ball is in shade. It's not the same reading as if you used a spot meter on a grey card in the same shadow however. The ball is round and picks up light on the top and sides and bleeding around, almost like the light spilling around someone's head who is similarly backlit.

Anyway if your thinking of running out and getting an iPod and a Lumu Meter and Cine Meter II app? Don't just get a good Cine meter.. if you have an older iPhone or iPod touch and want to experiment, you can try the app and watch the videos. If I constantly replicate my experiences today I'll have to start shopping for a handier Cine meter to throw in the camera bag and use the iPod Touch meter as a backup.