Firmware 3.02

[Scott Berrington]

The HDMI output is 4:2:2 8 bit, and as far as I understand, it comes from the image processing section of the camera...before the h.264 compression.

As for the update, it doesn't appear to provide any substantial improvements that I was hoping for.

[Duncan Craig]

We can only speculate about how the camera processes it's data.

Regarding updates, I'm not sure how long you've owned the LS300 but there have been lots of updates.
I've done five myself (I have all the versions backed up) and I'm sure the 4K60 update will be along soon.

You want 'substantial improvements' which almost sounds like you won't be happy until the camera can do 4K60. But this feature was only made public quite recently and shouldn't be a deal breaker for buying an LS300.

[Marcello Mazzilli]

Should already have internal 150mbit (4k).
And already should be more luminous than BMPC 4k.
I have it and can't go more than 400 iso without noise

[Noa Put]

"Addition of Natural mode in color matrix (Suitable for condition under blue light)"

This sounds interesting, I noticed the ls300 having quite some issues displaying blue led light, I get that kind of light a lot at weddings and on the ls300 the color glows like it's radio active, my panasonics have much less issues displaying that color.

[Lee Powell]

Haven't tried the "Natural" profile, as it's available only in Rec 709 gamma mode. In my view, there is no shooting situation where I'd use Rec 709 mode on the LS300, J-log1 is a better choice in every case.

Before J-log1 was introduced, I tried numerous combinations of gamma, black level, and color bias in Rec 709 mode, all in a futile effort to record reasonably well-behaved footage under LED stage lighting. Pure blue LED light was a particular problem, provoking the camera's internal light meter to underestimate its intensity, resulting in overexposure.

I'm completely unimpressed by this firmware update, it's little more than a marketing push for JVC cloud. There are vital areas where the efforts of JVC's engineers should be focused, issues that would make a real difference in the functionality of the LS300:

1. Enable the Sharpening control in J-log1 mode. This setting should have NOTHING to do with the selected gamma mode.

2. Fix the "800%" spot meter to use peak detection rather than area averaging. It is currently useless for reliably gauging peak exposure.

3. Make a serious effort to address the OIS support problems with numerous Lumix lenses. At BARE MINIMUM, document which lenses have been tested as fully OIS-compatible.

4. Verify support of ALL currently marketed MFT lenses by Panasonic, Olympus, and others. Document all exceptional support issues, as Panasonic has done with each of its MFT camera models.

[Stein Onshus]

Re the 4.2.0 v 4.2.2 discussion.

I remembered and tried to attach the JVC 4K cam white papaer from 2015 where they say:

"HDMI is uncompressed
As we have indicated earlier in this document, any compressed HD in a standard consumer delivery format is 4:2:0 at 8-bit depth. However, any uncompressed HD format, being exclusively professional (i.e. HD-SDI and 3G-SDI), is generally 4:2:2 at 10-bit depth. HDMI connected video is indeed uncompressed but generally 4:2:0 at 8-bit depth (higher sampling and bit depth are permitted).
Two of the 4KCAM camcorders (GY-LS300 and GY-HM200) feature identical LIVE wired HD-SDI and HDMI outputs, while the GY-HM170 only features HDMI. IP streaming is not included in the GY-HM170."

They are pretty clever not to tell exactly what their professional camera deliver, but with some good will I can interpretate that they deliver 4.2.2 10 bit, but my realism tell me 4.2.0 8 bit.

[Marcello Mazzilli]

I got an answer from JVC and after few exchanges the guy said that I am getting 10bit 4.2.2. and I am getting it directly from the sensor (and is not an up-sampling from the 4.2.0). But he wasn't really convincing as I seemed to know more about it than him.
One thing I discovered (Wikipedia, HDMI) is that HDMI can't be 4.2.0. but only 4.2.2 or 4.4.4., but bits can go from 8 to 12 and over. So we are 100% sure is 4.2.2. What we don't know is if it's 10 bit or 8 bit, and if it's 10bit if it comes directly from the sensor or is just a "fake 10bit", a resampled version of the 8bit signal

[Stein Onshus]

I picked up this to understand the potential in the LS300:

Altasens 4k super 35 sensor AL41410C gives up to

 4512 x 3008 resolution at up to 50 fps
 4512 x 2520 resolution at up to 60 fps
 4096 x 2160 resolution at up to 72 fps for cinematic 4K
 3840 x 2160 resolution at up to 72 fps for Ultra HD video
 1920 x 1080 resolution at up to 240 fps for slow motion playback

I am looking forward to JVCs full utilisation of the hardware.

[William Hohauser]

The imaging chip's full potential may not be realizable due to other constraints in the camera, heat dissipation being the first reason that comes to mind. Other components needed to process such high frames rates may move the camera's price out of the range we expect it to be.

[Steve Ritchie]

Tested out the new 3.02 firmware, but the scaling error / stepping artifacts in 4k are not fixed yet.
Reported this to JVC months ago, no reply. Hopefully they'll fix this in the next major firmware update.
Maybe if more folks chime in to JVC, they might bump this up on their priority list.

Attached stills extracted from 4k (3840x2160) vs C4K (4096x2160). There are no stepping artifacts in C4k, but in C4 (UHD) diagonal lines look pretty ugly.

ISO: 1600
JLOG1
Shutter: 1/48
fps: 24

After looking deeper into this it seems it may have nothing to do with JVC, and there may be no scaling error at all. It seems to be a playback issue with MPC.

[Duncan Craig]

It's hard to tell which one looks better isn't it? A more regular test subject may have helped.