Noob question Re: GH5s and V-Log
I'm researching this camera tha GH5S and have seen a lot about V-Log.
My question is regarding exposure.
When shooting, do you just expose as you normally would using histogram etc? Is there a benefit of using a LUT preview to expose?
I'm just a bit curious because it can be hard to tell visually in v-log no?
Thanks and feel free to point out anything I seem to be misunderstanding
Since you remap colors and swap out color info for more bandwith, you want to use as much as the codec can handle, thus expose close to max. (Or perhaps a bit over.)
You are going to do heavy CC anyways, or you would not shoot Log, are you.
Now, since doing heavy CC assumes you know what you are doing, why are you asking this?
Did you run tests yourself? Are you able to make your footage better than shooting it normally? At what cost?
Kind of confusing reply tbh. I'll try to answer your questions.
"You are going to do heavy CC anyways, or you would not shoot Log, are you."
I want it for the extra dynamic range, so depends what you mean by "heavy CC" - My gigs have some difficult lighting conditions and require me to maximize dynamic range to preserve highlights/shadows and from what I understand V-Log will help with this, no? With the drawback being that it is harder to expose than regular footage.
"Now, since doing heavy CC assumes you know what you are doing, with V-Log"
No it doesn't. Knowing how to CC doesn't assume you know how to shoot/expose in V-Log. Not sure what you mean here? That is why I am asking this lol.
"Why are you asking this?"
I'm asking this because I have never had a camera that can shoot log and I want the extra dynamic range in post but wanted tips for exposing in V-Log.
"Did you run tests yourself? Are you able to make your footage better than shooting it normally? At what cost?"
No, like I said I do not have the camera yet.
From what I've gathered from more research. There is a "general rule" of over-exposing about 1 stop but this does is not always true. Using the Live LUT preview on the GH5s can help you visually tell better along with the using the exposure tools to keep the shadows above 0% and highlights around 70 - 80%. If you cannot do this, then you will have to decide if you'd rather have blown out highlights or lose detail in the shadows.
Does this sound right to you guys? Anything to add on that?
In order to save the extra luma info, you need heavy CC / masking, as the output codec is not capable of displaying it either.
So, this is NOT a trick that you suddenly have no longer blown out windows / candles / lights and having beautifully exposed faces.
Then, to answer your question, you should be close to overexposure. Underexposure will make you loose to much color information.
Now, if you like it or not / what you should do is get a cam, rent one, borrow one, steal one and shoot the same testshots in all varants you think is wise, and see what comes out after processing, and how much time that takes.
(If you do shoulder / run & gun work, also test that to see how easy it is to do moving masks on blown out highlights.)
Then decide for yourself. Let no-one tell you it is great or easy. (I have yet to see a real life scene that indeed looked better then the same shot normally taken.)
It has a purpose, but only for the ones that know what they are doing / have time and money.
For the rest it's the emperors clothes.
Thanks for the detailed answer.
I am getting this camera regardless so I will take your advice to heart testing the same shots and you're right there is a good chance my skills are not up the task of CCing V-Log.
I have a good amount of experience in editing, animating masks etc but I have never dove head first into color. I've mainly been learning through tutorials.
Can you go into more detail about what you mean by "Heavy CC"?
Is the stuff in this tutorial not valid?
His process is basically:
1) Using Channel Mixer for color - Enter a specific set of values that are specific to V-Log
2) Uses RGB Curves to get the contrast right
3) Uses Lumetri Color to make additional changes to stuff like Temp/Tint/Exposure/all those options.
Are you saying this is not going to get me any more dynamic range? Is this process flawed? I will not gain any benefit from this if I shoot it correctly?
I guess I am confused by what you mean by "heavy cc" and how much I need to learn and know to be able to not call myself an expert but simply to benefit from shooting in V-Log.
Heavy CC means, heavy Color Correction.
I'm not going to follow this guy. If the video is 17 minutes and he says it's really easy there is something wrong.
It should be clear in 2 minutes then.
(And I'm a Resolve / Avid guy)
You will NOT get more dynamic range on output, but you WILL get it on input.
So, since the crushed / blown out parts are saved, you need to bring them into the normal range, AND save the overall look. (So the interiour is not all smashed into one level to save the tiny look outside the window.)
Toy yourself first, with actual footage, that is the ONLY thing you should believe.
[Bouke Vahl] "I'm not going to follow this guy. If the video is 17 minutes and he says it's really easy there is something wrong.
It should be clear in 2 minutes then."
LOL! that made me laugh 😉
I agree in regards to the video.
For the main poster, never shy from learning.
If you have the patience to learn to grade.
I'm a VFX guy and never thought I'd get into this wonderful realm 😉
And I have extensive experience with a GH4 VLOG but live in a RED ECO.
Grading GH4 frames are yummy :)
There is no need for any special "skills" if you take a proper approach. Starting from scratch and trying to use "standard" color controls to turn the Vlog image into a sensible Rec709 image is not a "proper" approach. Try this instead: Forget about trying to normalize the shots in Premiere Pro. Open the shots in Resolve. Go into Color Management and enable ACEScct, selecting Panasonic Vlog as the input color space and Rec709 as the output. When you bring the shots into the media pool make sure they are all set to use Panasonic Vlog as the input color space. Open the shots in a new timeline and go to the Color page. Set the Contrast to somewhere between .85 and .9. That's it. You should now see an excellent representation of what you shot. The advantage of being in Vlog is that you can now vary the exposure and balance settings to whatever you need. The advantage of using ACEScct is that you don't have to figure out a "proper" starting point.
I know that many people online (including countless YouTube videos) will tell you to do everything by hand from the original log image. Those people, for the most part, are self taught experimenters and not professional colorists. Professional colorists understand color spaces and color pipelines, and how to interpret them for proper results. It's not about blindly tweaking curves until you like what you see. It's about knowing how the original file was created and putting it into a proper path for the monitoring you're using and/or the format you're delivering to. To do that sensibly requires a proper scientific approach that implements proper transforms to get you from one to the other. That's how high end digital intermediate work is done and it's what you should at least try if you want an optimal result with the least guesswork.
BTW, a suggestion regarding exposure. When shooting in Vlog, the histogram is your friend. It will turn from yellow to white when you're in a "proper" exposure range, but more importantly, it lets you see when you're approaching clipping points in either the highlights or the shadows. Keeping the histogram towards the highlight clip without actually hitting it is a good way to get a solid exposure, although anything in between the two should be fine.
[Mike Most] "There is no need for any special "skills""
[Mike Most] "Open the shots in Resolve."
That is special skill #1 ☺
My beef with this is, you need to do a lot of work that is absolutely NOT needed if you have shot normally.
And, after having your machine do a lot of calculations, the resulting image is not better than it would have been when shooting normally.
The whole shebang with log / luts was made up to overcome shortcomings in the codecs, not be able to record raw.
Now, if you look at the YouTube video from the OP, I notice that it sucks big time.
Why not spend time on rehearsing your takes / think about what you want to tell, edit properly, do another take when an airplane flies over? Spend time everywhere to improve the quality of your show, NOT in this kind of stuff.
To make it worse, his example shots are typical shots that do NOT need Log.
Even worse, the last shot (and poster frame) with the women in backlight, I would have expected the demo to be able to recover the women instead of ending up with a silhouette.
So, it has a function, but only for high end stuff or if you are a geek with lots of time.
(I've seen people spend more time on grading than it takes me to edit and finish a complete piece for national television, including a lot of editorial descisions, selecting music, mixing sound, graphics etc.)
Editing and shooting are typical things where you have to make (a lot of) decisions. Postponing them to post is not the best way. (Unless you have unlimited time / money.)
Thanks guys a lot of good opinions and useful information in this thread.
I really appreciate the well thought out responses.
[Bouke Vahl] "The whole shebang with log / luts was made up to overcome shortcomings in the codecs, not be able to record raw."
Uhhhh.. No. Log coding has nothing whatsoever to do with "codecs." It is a way to bring a larger dynamic range into a limited 10 or 12 bit space so that it can be contained within a reasonable size file. It is an adaptation of what was done for film scans in the past, and is intended to retain all values of an image and leave it up to the user as to what values should be retained in the final image and which should not. If you "shoot normally" (I interpret that to mean record in Rec709) you are taking images with more values than can fit into a Rec709 signal and deciding during production which values at the extreme ends are going to be retained and which are going to be clipped. The problem with that is that you have very limited control of contrast when shooting. You can use filtration, or you can flatten the signal, but it's not optimal to do either of those things. Encoding to a log curve allows you to retain more than the Rec709 signal can hold, so you can adjust contrast/clipping in a more controlled environment. Recording RAW also largely accomplishes this, but at the expense of having to do much "heavier" processing at the back end. It might also be pointed out that most "RAW" formats are not purely the information from the sensor in its original form. Most of them do some type of transform to allow that information to fit into a reasonable sized file, in a number of cases using log encoding of the RAW values to accomplish that. Log encoding is just math, not some elaborate scheme or hoax. It't not a gimmick, it's a way of retaining extended values at the high and low end to allow for more flexibility in finishing. The reason I suggested the ACES approach is because properly used, that approach "tames" the entire process and gives you essentially a "better" version of Rec709 recording due to its use of "proper" color science and sensible rolloffs at the high and low ends (particularly the high end). And you still have the original log image if you wind up using a colorist for final finishing.
Ok, poor choice of words on my part when I wrote 'made up'.
But, it is a hack to fit dynamic range in a codec.
What do you think, that codec developers don't know about math? It's a trade off, and a poor one when it is not needed. I'm not going into discussion about that. (And I've yet to see an A/B comparison where Log defeats normal shooting on average shots.)
Now, if it were your money, would you spend your money on CC'ing or on good lightning?
And that is just the image. Would you spend your money on a high end cam or on a good story?
Would you spend your money on gorgeous images but no time to edit them?
I prefer a good movie on VHS over a bad one in IMAX.
Let's agree to disagree.
While I can agree to disagree on the ultimate value of recording in log in certain cases, it is most certainly not a "hack." It is a way of taking the enormous range of values in the real world and capturing it in an efficient manner. The real world exists as linear light, but your human vision system interprets that huge range of values by limiting the number of variations you can see in the brightest values and emphasizing the lower and midrange values that are more critical to interpreting the world you see. What that means is that the human vision system is logarithmic, just like the "hack" you're talking about. When light is doubled - i.e., increased by one stop - your eyes and brain don't see the result as being twice a bright. If one really wanted to record what a camera is actually seeing as absolutely as possible, one would always record in linear light (i.e., what the sensor actually sees). However, as with human vision, that isn't practical or necessary. Log encoding is one way of approaching that problem, and applying a gamma curve is another. Log encoding has the very real advantage, though, of being much more efficient in its use of available values to better capture a wider range. In essence, the use of log encoding is no more of a hack than the human vision system is.
You win, no questions asked.
Math over good shows, logic over entertainment. Arrogance over argument. (Trust me, on the last one it's hard to beat me.)
It's not binary choices. You can actually have proper craft and good material. You can have top quality images and a good story. You can have better approaches and simple ways to make it sing. All one has to do is want it.
Sure Mike, happiness / wealth / success / food / shelter, well, even health is a choice! And it all comes for free, without having to sacrifice one for the other.
Just tap your heels three times, and it will all come trough.
Happy new year!
Nah, sorry, that was too harsh.
I do agree on large amounts of what you're saying, and you misinterpret what I call 'hack'.
But, you seem to live in a high end world and seem to forget what the majority of people have to face.
Small example, our biggest newspaper does videos now, and even they do NOT shoot in log and have proper white balanced / lighted / recorded, they still are able to annoy me with a video that fails in every aspect of the craft.
So, please, bear with me and do read what I write, and try to understand at least half of it.
Happy new year!
>>Small example, our biggest newspaper does videos now
What's a newspaper? ;-)
An organisation that forces clueless people in this bizz that you have to educate for free so that in a few years you are out of job.
It might be less frightening, it may only be that news gets more annoying than it is right now, think of 30 minutes of video that tell you nothing but will make your ears, eyes and nose bleed. (This site is moderated, so I leave it to that.)