GH4 Panasonic V-Log-L - the Seven-Bit Solution for the GH4?

L

Lpowell

Veteran
Panasonic's recent announcement of the long-awaited V-Log-L firmware upgrade for the GH4 prompted me to examine the technical details of its implementation. V-Log-L is a 12-stop version of the 16-stop V-Log tone curve designed by Panasonic for the Varicam video camera. It is calibrated for use with standarized Varicam LUT's, applied to the V-Log footage in post production. According to Panasonic, their design goals for V-Log-L on the GH4 were to increase dynamic range from 10 to 12-stops and to enable Varicam V-log LUT's to be used interchangeably with V-Log-L GH4 footage:

http://m.shop.panasonic.com/cameras...hangeable-lens-ilc-cameras/DMW-SFU1-VLOG.html

V-Log-L is also included in the Panasonic DVX200. Barry Green recently posted an explanation of how V-Log-L differs from the V-Log tone curve:

http://www.dvxuser.com/V6/showthrea...corder&p=1986571024&viewfull=1#post1986571024

Panasonic's V-Log tone curve chart shows 16 stops from -8 to +8, with zero calibrated to 18% gray reflectance. Note how the flat log segment of the curve extends down to about -4, with a rolloff knee down to 0% reflectance in the bottom 4 stops of the shadow range. This is the part of the dynamic range that must be accurately exposed in order to match the non-log segment of the V-log LUT you select.

For the GH4, however, a no less important issue is the green rectangle Barry drew on the chart at stop 4. That's where the GH4's 12-stop V-Log-L dynamic range ends with highlight clipping, 4 stops short of the 16-stop range of the Varicam's V-Log tone curve. In order to match the V-Log curve, Barry recommends setting exposure for highlight clipping at 79 IRE. That would appear to sacrifice the camera's 80-109 IRE range for the sake of allowing you to use Varicam V-Log LUT's with GH4 V-Log-L footage.

The vertical calibration on the chart is a 10-bit scale from 0-1024. On a camera with an 8-bit codec like the GH4, you would have a scale from 0-256. That means highlight clipping would be recorded at around 182, with no image details recorded in the range 183-255. At the other end of the scale, the bottom 4 stops of shadow detail would be recorded in the range 32-48, just 4 bits of data to cover 4 stops of dynamic range. Even if we assume, for the sake of argument, that the GH4's H.264 encoder can preserve this data with flawless precision, the decoded result will contain image data limited to a 32-182 range. That leaves 106 unused data levels, nearly an entire MSB of bit-depth, resulting in an effective 7-bits of encoded data.

In conclusion, 8-bit V-Log-L appears to squeeze out 12-stops of dynamic range by sacrificing tonal range in both highlight and shadow regions. As V-Log was originally designed for encoding at 10-bits or greater data precision, the use of a 10-bit external recorder with the GH4 in V-Log-L mode should produce a significant improvement in image quality over the camera's internal H.264 encoder.

attachment.php
 
Last edited:
There have been custom LUTs designed for SLOG2 that allow for intentional overexposure by 1 or even 2 stops. These LUTs reduce noise and provide a great look in SLOG2 enabled cameras if you decide to lose detail in the shadows. Perhaps GH4 specific LUTs for VLOG-L will be designed for those who don't intend to intercut with a Varicam. These custom LUTs would allow highlights to be placed higher and thus using a more full range of the 8 bits.

I'm not denying 10-bits will provide better results, but I think exposing hotter and using a custom LUT will solve the 7-bit problem.

Edit: Wait, I think i'm understanding now. V-LOG-L places absolute clip at 79IRE, so it literally doesn't use anything above that, thus squishing everything between 32 and 182.
If there is no way to adjust that in-camera, then 8-bit 4:2:0 is not a good recording method for V-LOG. Hmmm.
 
Last edited:
Is this confirmed, or just wild guessing by styding this chart? I have seen the cine-D vs V-log, and it seemed to pack a lot more DR. Not sure if the atomos shogun was used or not.
 
Polaroid22 said:
Is this confirmed, or just wild guessing by styding this chart?
Click to expand...
the most useful chart about vlog-l vs. vlog from pansonic was recently published in a german article:

1405-PIC5-PIC5-12336_PIC5.jpg

i perfectly agree with LPowell concerning this stupid waste of 50% values, given the 8bit recording and the actual dynamic range of the sensor in this camera.
 
Last edited:
This all seems to make perfect scientific sense on paper. What happens when people are able to deliver much nicer looking footage from 8-bit internal V-Log recordings? Will the numbers matter anymore?
 
wwjd said:
This all seems to make perfect scientific sense on paper. What happens when people are able to deliver much nicer looking footage from 8-bit internal V-Log recordings? Will the numbers matter anymore?
Click to expand...
² everybody that got a GH4 with v-log seems really enthousiastic about it. So maybe it is just as good as they say it is? Time will tell, when they finally ship this and more people can do tests.
 
Stop overthinking this guys. Name another sub$1500 4k camera that has 12 stops and the ability to record 4k internally or 10bit out to recorder? $99 bucks is chump change to anyone who uses the gh4 for video use and you're getting another 2 stops DR with this release. I've pushed the hell out of the gh4 cine d profile with great results. It's a $1500 camera that is giving you what was $80,000 just a couple years ago.
 
this technical breakdown is a waste of energy. If it can produce prettier images, which from beta versions shows it can, then thats all that matters.
 
What jumps out for me is the improvement in color rendition with the new footage. Color makes such a huge impact on the final product. Worth the price of admission imho. I don't know if this is post induced or improved color science, but I have never liked GH4 footage until seeing V-Log samples. Looking forward to the GH5.
 
I think it's important to recognize that the tonal areas in which precision is being compromised (which any LOG curve must do) are the areas in which our eyes can't distinguish fine gradations. (Especially highlights.) The sensitive midrange is being afforded the most bits. This isn't revelatory, this is what a LOG profile is for...sacrifice precision in the extremes in order to squeeze in more dynamic range. That's a trade-off that will almost always yield a better PQ.
 
Polaroid22 said:
Is this confirmed, or just wild guessing by styding this chart? I have seen the cine-D vs V-log, and it seemed to pack a lot more DR. Not sure if the atomos shogun was used or not.
Click to expand...
Bit depth does not impact the captured dynamic range. Both 8-bit and 10-bit will capture the same 12 stops. The 10-bit version will have more precision between the extremes, but the extremes remain the same.
 
12 stops of DR squeezed by a LOG gamma into an 8bit codec looks great on the C300/C100. It also A7s S-Log internal recording, which is 12-13 stops of DR in an 8bit codec.

This theoritical argument has been there for a very long time that we sacrifice ''tonal percision'' when we record LOG on an 8bit camera, yet all of us are shooting it and not getting in ''tonal precision'' issues vs REC709.

It's all about the image, and the GH4 V-Log image adds 2 more stops of DR and better colour rendition in internal 8bit 4K, no one can argue with these results. And people who want these improvements will get it.
 
joe12south said:
I think it's important to recognize that the tonal areas in which precision is being compromised (which any LOG curve must do) are the areas in which our eyes can't distinguish fine gradations.
Click to expand...

Ebrahim Saadawi said:
This theoritical argument has been there for a very long time that we sacrifice ''tonal percision'' when we record LOG on an 8bit camera, yet all of us are shooting it and not getting in ''tonal precision'' issues vs REC709.
Click to expand...
Apparently some folks here have difficulty distinguishing between "theory" and "implementation" when a discussion about log profiles actually uses math to evaluate them. Of course, it's the GH4 and as we know, it's all about the mojo, right? Besides, who cares about 7 or 8 or 10 bits when you've got logarithms working for you?
 
Ebrahim Saadawi said:
12 stops of DR squeezed by a LOG gamma into an 8bit codec looks great on the C300/C100. It also A7s S-Log internal recording, which is 12-13 stops of DR in an 8bit codec.

This theoritical argument has been there for a very long time that we sacrifice ''tonal percision'' when we record LOG on an 8bit camera, yet all of us are shooting it and not getting in ''tonal precision'' issues vs REC709.
Click to expand...

What Barry Green (and Lpowell) are saying is that VLOG-L does NOT work like C-LOG or S-LOG. Both C-LOG and S-LOG use the full 0-255 range of 8-bit. Their maximum brightness is 109 IRE.
VLOG-L, on the other hand, has a range between 32 and 79 IRE. It is literally impossible to raise the highlights above that. This means VLOG-L is squeezed into a much narrower tonal range than Canon's or Sony's LOG solutions. So to use those as a defense of 8-bit LOG is comparing apples to oranges.
Of course, empirical proof counts for something too, so I'll reserve judgement till I get to do my own shooting with VLOG and see how it performs. But I'm hoping that Barry's needling of Panasonic about this forces them to remap the VLOG-L curve to the full luminance range of the camera.
 
You must log in or register to reply here.
Back
Top