Best Settings for Skin Tones on 7D?

matt9b said:
Look at the sky in the shots - and you'll see some bright parts are over exposed. In the super flat mode, the sky retains detail.
Click to expand...

These details are being recorded more or less regardless of picture styles and settings!

They are there, you just don't see them as you do with the superflat settings straight out of cam. You can recover them in post and you'll get about the same amount of detail and information in highlights and shadows as with superflat, the only difference is that superflat doesn't do a good job to skin and faces.

I recovered highlight and shadow detail for the Fuji Pro in After Effects, no CC, to compare with superflat:

1269951719.jpg


1269951751.jpg



Fuji out of cam/original:
1269952260.jpg


Btw, I can't seem to use those picture profiles in movie mode with T2i / 550D altough I can in stills mode. Anyone same experience? Is it not possible to use these profiles in movie mode with T2i?
 
Last edited:
GRENCH said:
Is there a flat setting that falls between superflat and standard. Super flat is a little to flat for my taste?
Click to expand...

Don't forget I filmed my shots in London, lighting and colours here tend to be very flat compared to e.g. the US. So, it depends where you live. I've attached some shots shot with better lighting - the first one is in the US but with same settings, and as you can see, the colours are nicer and it's not as flat.
 
stip said:
These details are being recorded more or less regardless of picture styles and settings!
Click to expand...

No they're not. The 7D applies contrast and settings to the RAW data, and then bakes this into the compressed file during save. It goes RAW data -> picture settings -> save. If your picture settings have high contrast for example, it is more likely to lead to part of the image being over exposed (or zero'd out), and details are not recorded. Take for example a high contrast scene with a light bulb. With film stock, you might see finger prints on the glass on the lightbulb. With 7D and standard pic settings, the whole lightbulb will be over exposed and there is nothing you can do to get that data back.

stip said:
They are there, you just don't see them as you do with the superflat settings straight out of cam. You can recover them in post and you'll get about the same amount of detail and information in highlights and shadows as with superflat

I recovered highlight and shadow detail for the Fuji Pro in After Effects, no CC, to compare with superflat
Click to expand...

No, you haven't recovered over exposed data. You've taken shots where the data was already there. Try taking the portrait clip towards the start and recover the sky, then post back :)
 
Last edited:
matt9b said:
No, you can't recover over exposed data. There just isn't any data there to recover ;)
Click to expand...

True of course, but I don't think superflat extends the dynamic range at all. Look at the silhouette of the trees, latitude gets artificially enhanced with superflat, resulting in these blocky silhouettes. It doesn't seem to me that there is more true dynamic range even compared to standard mode.

1269954645.jpg

1269955625.jpg


original standard mode
1269955659.jpg
 
Last edited:
it's pretty hard to judge detail from jpegs and vimeo uploads, just would like to put in a 2nd request for tiffs from the native files. y'know that way everyone gets a clear picture of what is really going on
 
stip said:
Never said anything about recovering detail in overexposed areas.

That was my point, the data is there, you just don't see it straight out of cam like with superflat.
Click to expand...

Hi Stip, sorry I think we're talking cross-purposes. You've misunderstood my point - I was talking about over exposed areas. You're not - Of course with no over exposure, you're going to be able to recover detail (although really, you're not recovering anything, the info is already there). The point is that with Super Flat, you can get more in, in more situations, esp. where lighting isn't controlled, without over exposing parts of the image. Just like film. This gives you more room to maneuver in post. With the picture styles you've chosen to demonstrate your point, there are no over exposed points in those particular images, as it was a pretty dull day, so you're missing the point. On lots of locations however, there will be over exposed highlights with most settings, that would be retained with Super Flat. For example, even on that dull day, the portrait setting in that clip has prob over-exposed the sky, while the face has good exposure.

There are 2 solutions to this with the 7D:

1. under-expose the image until there are no over-exposed areas, and then bring up the mids & darks in post (Con: lose detail in mid & darker areas)
2. use the super-flat setting with custom white balance & bring down the blacks in post

This is one of the primary problems people face with video, and when trying to make video look like film. Video cameras have more contrast than film, in the upper and lower frequencies. If you don't believe me, check out the recent test between all these cameras and film cameras here: http://www.zacuto.com . As you can see from their tests, the 7D and other DSLRS over-expose the highlights far sooner than film does. The over-exposed parts in the portrait clip I provided earlier also demonstrates this phenomenon.

This is going to happen in many situations outdoors, with almost any setting other than Super Flat. If you still not convinced, try doing the same thing you did earlier, to the portrait clip :) or try shooting a high contrast scenario yourself, where the highlights are over-exposed and see if you can find a work around that isn't the Super Flat setting. ;)
 
Last edited:
Matt, you're right, the portrait setting shot clips detail in the highlights a lot more but if you look at the t-shirt, the blacks, despite the higher contrast, are as bright as on the superflat shot, so the exposure might just differ.

Anyway, the basic point I think I found out is that superflat does not give you a wider dynamic range than you can achieve with other picture styles. While it is, like you mentioned before, the only preset that shows you all detail the camera actually records during live view, it also has an artificial and nasty looking effect on skin tones.
I found most presets with contrast turned all the way down and minor tweaking in post will result in the same dynamic range but punchier results, also after CC.
Of course it's also a matter of taste, the superflat look might be desired.

I'll try to load an example of recovering detail, the vimeo footage might be too compressed to do so.
 
J Davis said:
it's pretty hard to judge detail from jpegs and vimeo uploads, just would like to put in a 2nd request for tiffs from the native files. y'know that way everyone gets a clear picture of what is really going on
Click to expand...

If I get time when I'm back at the studio, I'll see if I can sort this.
 
stip said:
Matt, you're right, the portrait setting shot clips detail in the highlights a lot more but if you look at the t-shirt, the blacks, despite the higher contrast, are as bright as on the superflat shot, so the exposure might just differ.
Click to expand...

The in-camera exposure meter showed bang on 0 in all shots.


stip said:
the portrait setting shot clips detail in the highlights
Click to expand...
stip said:
anyway, the basic point I think I found out is that superflat does not give you a wider dynamic range than you can achieve with other picture styles.
Click to expand...

You've just contradicted yourself. You're saying the portrait setting clips detail in the highlights, but the Super Flat setting retains detail in the highlights. Then you say superflat does not give a wider dynamic range.

Strictly speaking, you're right, the dynamic range is limited by the compression. But then you could also say the same for sensor size, yet people still use anamorphic lenses to increase resolution. Think of it like this: with Superflat, you can cram more detail into that dynamic range, than you can with other settings. You can then expand it in post.

stip said:
While it is, like you mentioned before, the only preset that shows you all detail the camera actually records during live view,
Click to expand...

No, sorry, thats not what I meant. In contrasty scenes, it is the only preset that shows all the detail even in post.

stip said:
it also has an artificial and nasty looking effect on skin tones.
Click to expand...

Not if you perform proper white, grey, and black balance.

Do you think these look artificial and nasty?

http://vimeo.com/9848608


http://vimeo.com/7545683


http://vimeo.com/7776144


stip said:
I found most presets with contrast turned all the way down and minor tweaking in post will result in the same dynamic range but punchier results, also after CC.
Click to expand...

Can you post footage? For an academy film look, punchier results might not be wanted, but footage speaks louder than words :)
 
Last edited:
Matt, just for the record, I'm not trying to be smart or proof you wrong. Like you, I'm trying to find out the best settings to achieve maximum results. If I disagree with you, than there's absolutely no pun intended.
 
stip said:
Matt, just for the record, I'm not trying to be smart or proof you wrong. Like you, I'm trying to find out the best settings to achieve maximum results. If I disagree with you, than there's absolutely no pun intended.
Click to expand...

Yes I understand. I see what you're getting at now - the *compression* has the same dynamic range no matter what pic style is used - so when you cram more frequencies of light into that range, and expand in post, you're going to lose a bit of light frequency detail across the spectrum afterwards. But this is what happens in film, and has a side effect of giving it a more flimic look.

I started off with the same opinion as you re: super flat, but with proper white/grey balance I'm now convinced. I'm interested to know if you think the graded super flat shots above give artificial / unpleasant look on the faces?
 
I think you will find that the best settings to achieve maximum results will vary from setup to setup. From tungsten to fluro to daylight. From dark skin to light skin.
 
matt9b said:
You've just contradicted yourself. You're saying the portrait setting clips detail in the highlights, but the Super Flat setting retains detail in the highlights. Then you say superflat does not give a wider dynamic range.
Click to expand...
Retaining shadow / highlight detail is not the same thing as dynamic range.

Dynamic range is the distance between the darkest and lightest parts of an image.

Superflat actually reduces dynamic range by raising the blacks and lowering the whites.

The problem is that in such a low-res (not film or 4k), highly-compressed, poor chroma format, it is effectively squishing too much information into too flat a range, so that when you go to add the contrast back in post, the information becomes stretched out too much resulting in the image falling apart.

Superflat is not a bad concept, but it goes much too far, especially for people talking about using it all of the time. It should really only affect the shadows / highlights, and even then in just a slight roll-off on the extreme ends of the histogram. It shouldn't touch the middle, at least beyond what Neutral does.
 
This is a framegrab of T2i footage I took at 1080p, I had to downsize to fit max upload size, as files are .tiff. The settings were "neutral" with contrast and sharpness all the way down. This is the setting the creator of the superflat picture style claims to give better results than his superflat.
The camera still recorded more details in highlights and shadows than it appears straight out of camera.

straight out of cam
http://www.dvxuser6.com/uploaded/46888/1269972708.zip

actual detail recorded
http://www.dvxuser6.com/uploaded/46888/1269972763.zip
 
Last edited:
If I remember from several months back the maker of Superflat came to the conclusion that a neutral setting with contrast turned all the way down would be better than using Superflat.

What I'm seeing in Superflat is a slight loss of detail. Don't get me wrong...I kinda like the look (reminds me of Cinemode in the HV20/30/40. But even Cinemode had its slight resolution loss issues). Didn't stop me from using Cinemode however. It was a great help in post work and gave that filmic look.
 
You must log in or register to reply here.
Back
Top