GH4 GH4 V-log?

Sorry, but the technical details you documented on Panasonic's V-Log-L implementation are not "speculation", they are hard specifications
You may have misunderstood -- by "speculation", I'm not saying you're speculating as to what the technical details are, I'm saying that you're speculating on how much of an impact it will make on actual, visible images.

Nobody's saying that full use of 256 levels wouldn't have been better -- of course it would. But the way VLOG-L is implemented appears to be exactly the same as CLOG in the lower 10 stops, and as far as I know, nobody's gone around saying that CLOG is problematical. Furthermore, those who have actually used VLOG-L seem very happy with it. So -- what purpose does it serve to raise alarm based on the numbers?

Let me put it like this -- if people were out there using CLOG and VLOG-L and hating it, then we could figure out why and say "ah, it's because they only use 150 levels, etc..."

But the reverse seems not to work -- it's like you're saying "They only use 150 levels, so you all should hate it." What's the point in that? Why raise alarm, when there very well may be no need for alarm at all (given, again, that CLOG has been working for something like half a decade with not a whole lot of complaints that I've heard). And VLOG-L has been used by various people for going on, what, 9 months now -- and there aren't a lot of complaints coming from those people either.

So why pre-emptively try to scare people? Why not let 'em try it and see if they like it?

As an ex-professional programmer, I can see and appreciate all the technical arguments. But the fact is -- sometimes it just works anyway. Take 24p, for example -- I can make every technical argument on earth as to why 60p should look "better", and demonstrate every flaw in 24p that there is... yet ... people frickin' love 24p. They just do.

Will they love VLOG-L specifically because it uses fewer code values than the maximum? I can't imagine why they would -- but they may love it as it is, and they may not care that it doesn't use every available code value. They may just like the results anyway (as, again, those who have been using it seem to).
 
I am very impressed with the skin tones that sample V-Log files show. What seems to be a critical factor, however, is using the actual VLOG Varicam LUT in the color grading application of your choice.
I used the official VLog to V709 LUT in Premiere's Lumetri, and the result is fantastic to my eye. However, using the default Alexa LUT oversaturates blues and greens, and crushes blacks. I bring this up because much of the VLog footage online seems to have been graded using the Alexa LUT. It's all subjective, of course, and adding the LUT is just a starting point, but boy the official VLOG LUT provides very pleasing colors to my eye. It really does seem to make the GH4 into a mini Varicam.

V-LogLUT.jpg
Alexa-LUT.jpg
 
Will people still use it with the GH4 encoder anyway and manage to produce pleasing subjective results? Of course, but it will demonstrate nothing more than the power of confirmation bias to vindicate any amount of marketing hype.
Wow. Just, wow. So the client who knows nothing about this discussion, who doesn't know 8-bit from cubit, who doesn't even know or care what camera was used, the client who just sees a result that they like – they're subject to "confirmation bias" and "marketing hype"? How?
 
I am very impressed with the skin tones that sample V-Log files show.
I'm right there with you. The improved color has me excited more than the extra DR. No amount of work with any of the existing profiles has made me truly happy with the GH4's rendition of skin tones. What I've seen from V-LOG L samples has put a smile on my face.
 
Someone who had early access mentioned that the Amira LUT was a better match.
I just tried it. It's not much different from the Alexa LUT. At least not the version that Premiere uses. Both the Alexa and Amira LUTS in Premiere/Lumetri push the saturation of blue and green unnaturally. This also leaves a greenish tinge in the shadows of skin tones. So far, the Varicam VLOG LUT, which was created to be used with VLOG, seems to be the best match. Which makes sense.
 
It looks good, but on my monitor the v-log-709 lut looks to crush blacks much more than the Alexa-709 lut. I agree the colors are less saturated on v-log lut though. Either way, it will be an improvement over the current gh4 from what I've seen.
 
Furthermore, those who have actually used VLOG-L seem very happy with it. So -- what purpose does it serve to raise alarm based on the numbers?
Two reasons:

While there are many who only care to evaluate subjective results, there are more than a few who do in fact pay attention to technical details in order to obtain the highest quality images their workflows can produce. I know from experience that people like this appreciate in-depth analyses like yours and mine; it gives them clues on where to look for limitations and pitfalls in their own results, which they might have otherwise overlooked.

In cases where enough people start sweating these kind of details, a critical mass of support sometimes develops around finding practical solutions to improve the commercial tools we have at hand. This is exactly what happened with hacks developed for Canon, GH1, GH2, and Nikon cameras. In each case, the manufacturers' original specifications were initially assumed to be adequate and were rationalized as good enough for practical purposes. But as soon as third-party hacks produced visible improvements in image quality, it became apparent that the technical limitations we had previously accepted were no longer necessary nor adequate for our own purposes. Eventually, the value of these improvements became so widely acknowledged that Panasonic themselves responded with firmware updates and higher bitrates on subsequent camera models.

The V-Log-L update for the GH4 has not yet shipped, so evaluation of the severity and any practical solutions to the issues you raised will have to wait. But thanks to your research on the DVX200, we'll know exactly where to examine GH4 footage for comparable V-Log-L issues. I have no doubt that if there does turn out to be a way to improve V-Log-L image quality, we'll find it.
 
Last edited:
I just tried it. It's not much different from the Alexa LUT. At least not the version that Premiere uses. Both the Alexa and Amira LUTS in Premiere/Lumetri push the saturation of blue and green unnaturally. This also leaves a greenish tinge in the shadows of skin tones. So far, the Varicam VLOG LUT, which was created to be used with VLOG, seems to be the best match. Which makes sense.

Yes, the Varicam VLOG sold me on doing this upgrade. The footage I'd seen so far hadn't impressed. The ALEX LUTS seemed too green. Played around with some sample footage with the Varicam LUT and it looks bang on.
 
That Panasonic provided lut does indeed look really good on my Android tablet. Not a great measure of quality, but a good anecdotal comment since a lot of consumption happens on millions of devices just like mine.
 
Hi Illya,
Could you give us any more detail about setting the exposure with V-log, is the histogram mapped to the log profile so clipping is at 100% (79 IRE)? There is some discussion about the need for a monitor with LUT functionality? Thanks :)

In the version I have currently, if your histogram is climbing the far right side then you are starting to clip. I would need to put it on a 7Q+ or other waveform to tell you the specific IRE. I can do this next week when I am back from vacation if you would like?
 
Back
Top