Sage

The announcement of the firmware update seems a relevant moment (incidentally my birthday), to give a little update on the GH5 project I alluded to here and on Bmcuser. I had expected in July that it would be ready to deliver before the end of August, but alas it continues to lay waste to timelines. Bringing it into reality has proven exceptionally complicated and challenging, 7 months in. I will release as soon as possible, which may be awhile yet. I suspect it will be of enormous value to anyone interested in the GH5. More soon

Something special coming for the Gh5 before the end of August*

I'm dying to share! Asap I will say - the project has occupied the greater part of my attention for the last four months, and is the result of a considerable amount of research, time, and investment

I've got something special coming for the GH5. You may feel differently soon

The Colorchecker Photo seems to work better with the automatic match features (and Video is great for manual) Btw, got something neat coming up for yall

Does anyone have Studio? I can send you two tiny files I am confident at this point that there will be a tone shift, but the contrast shift is still in question

Do 8 bit V-log files look exactly like 10 bit V-log files? Or is there a contrast & tone shift (10-bit w/ more contrast, more yellow)?

Thank you for taking the time to share this workflow. It really makes the files workable

What do you think of the new GE Reveal HD+ LEDs?

Denis Villeneuve!

Just got the camera, and found there is a little wrinkle in the workflow. The MXFs are being marked with a .9375 pixel aspect ratio in the re-wrapping of C4K files (4096x2160); this results in a slightly squeezed frame, into a 16:9 ratio. It should be a 1.896 ratio (1.0 pixel aspect ratio), and at the moment I am correcting it with 'interpret footage' in Premiere at the outset.

Have you considered FHD? The GH5 1080p is really great I gather Also, when the time saved *creatively is factored in, the proxies are not a bad deal. You really will edit faster. With a Ryzen, I think a case could be made for rendering proxies

That is a tricky situation. It should be faster than realtime, but not less than half. Approximately 1.3-3 hours. Assuming no PC bottlenecks. Write to SSD (a different drive than the read), try overclocking the 3770 if its a K - they are very generous with the extra clock cycles, I run at 4.5 with a cheap 212 Evo cooler You'll will want to edit proxies from an SSD, they are small enough to fit and will have instantaneous performance

Proxies! I know you don't want to transcode, but there is nothing like it You edit better. Simple as that. JKL ftw With your specs, transcode time won't be significant Recommend: H.264, Exactly quarter resolution for C4K, or full resolution for FHD, 22mbps for quarter C4K/20mbps for FHD, 6 gop, (max render quality off) all stored on an SSD

Its often mentioned in the press that the 2nd SD slot may be used for recording proxies, but I find no mention of this in the manual. To those with the camera, is this possible?

That makes the most sense as to why its not working; that would mean that Panasonic has coded the GH5 file header to specify the wrong IRE placement, according to their own V-Log specs (or it is misread by Premiere, due to differing header format from Varicam). And it is a matter of file header ID, as the wrapping method's success reveals (explains the inconsistent handling of levels across NLEs)

Excellent, that is what I was missing. This is even more puzzling; why wouldn't it work, given the 64-940 envelope? Adobe used to be fast with this kind of thing

Yes, the rewrap approach is probably the best way to go (I don't have the camera yet) The information above, and the little discovery by Justin is of particular interest to me as I am working on a little project involving V-log mapping at the moment That compressed luminance recording of v-log and v-log L is even beyond the 64-940 (0-100%) IRE envelope. I.E. exposure bottom* is intended to map at 128. And that is with the 64-940 envelope

Indeed; its confirmed. In this white paper: http://pro-av.panasonic.net/en/varicam/common/pdf/VARICAM_V-Log_V-Gamut.pdf - the values of 7.3% IRE and 61% IRE cited as corresponding to the 10-bit values of 128 & 602 respectively exactly match the 64-940 luminance level placement.

Thanks for sharing this Justin. To further this, it looks like the 64-940 (0-100%) IRE placement matches the official Varicam specification for V-log, as shown in the graphs on this page: http://www.slashcam.de/artikel/News/Naeheres-zum-Panasonic-GH4-V-Log-Update.html It would make sense that it is the data arrangement that Premiere is expecting and was coded to handle previously (standard Varicam levels)

The GH5 internal codec is most definitely limited. In that 10 bit footage, I see distinct color patches in skin tone when very zoomed in. In motion, full-scale, it is not too big of a deal to me. Its 'good enough' for an internal codec. What is more concerning is distinct banding in the sky at the beginning of the footage. I have watched a lot of GH5 footage, including more 10 bit v-log footage from Pampuri where the sky *wasn't banding, so I have a theory.. Perhaps more later. I will likely purchase the camera after NAB

Lol To be honest, I think there is enough there to really get something good. I've been playing with this 10 bit *V-log footage from Emmanuel Pampuri -

This was hilarious Andrew, thank you

1) Great write up Andrew, you have a talent for capturing the zeitgeist 2) Solid site redesign ++