Please Critique My Work!

[Ed Roo]

Here is 24P with -5% Exposure correction and Automatic White Balance correction

A Paul Story 24P -5E WB on Vimeo

I don't think there is any way to reduce the hot spot on his head without knocking everything else out of kilter.

[Andrew Smith]

The auto white balance definitely helps.

Andrew

[Bryan Worsley]

Kind of, but now there's a bit of a yellow shift. So that was 'Automatic White Balance'. Does FCP X not also give the option to select the neutral reference with a colour picker/dropper ? Definitely needs tweaking anyway.

I'll look at the uploaded '24P 0 Exposure' video tomorrow and see what I can come up with. I use DaVinci Resolve though.

[Ed Roo]

Bryan, the Manual White Balance dropper wants an area of pure white to reference. I don't have any white in my video! :-)

Yes, the Automatic White Balance did introduced a noticeable yellow tint.

I have adjusted the color in a subsequent edit, bringing down the Master (yellow-green on the left side of the color bar) 110 degrees 0 % to 95 degrees -15%.

I will post that edit later.

Bryan, what did you see that indicated interlacing in the original video?

You guys are teaching me to see things that I previously hadn't noticed! Thank you all!

[Bryan Worsley]

Originally Posted by Ed Roo

the Manual White Balance dropper wants an area of pure white to reference. I don't have any white in my video! :-)

But you have what I suspect is fairly close to neutral grey - the sweater.

Originally Posted by Bryan Worsley

If it has an white balance tool with color picker maybe try sampling off his sweater for neutral grey, say on the lighter edge of the cuff on his left arm, and see how it looks.

You want to pick a sample that is fairly uniform and receiving direct light from the daylight balanced (5600K) LitePanels behind camera i.e. not areas of shadow or light reflected off his face. Zoom in if needs be.

Of course there are other ways to manually white/colour balance. In Resolve, I prefer to isolate the best 'neutral' reference available in frame with an HSL key, zoom in on it and use the Colour Temperature and Tint tools to manipulate the trace on the Vectoscope. The 'white balance' colour picker does a pretty good job though. Again, I have no experience with FCP X.

Originally Posted by Ed Roo

what did you see that indicated interlacing in the original video?

Interlace lines, often referred to as 'mice teeth'. It's where there is movement between the upper and lower fields. You don't see those lines in a progressive frame. If you step though the original video (exported as '24p over 60i') frame by frame, where there is movement you'll see an alternating sequence of three progressive frames and two (periodically one) 'interlaced' frames (with 'mice teeth'). That's due to the 3:2 'pull-down' sequence. Where the 60i interlacing combines the top field from one frame and the bottom field from the next duplicated frame, it will change nothing. But where it draws the top and bottom fields from dissimilar frames, and there is movement, the interlace lines will appear.

Example:

Interlaced frame - note the fine horizontal 'mice teeth' lines:

http://i.imgur.com/JIuEdt3.png

Next frame - stays progressive (no mice teeth):

http://i.imgur.com/weL2iQO.png

You don't want those 'mice teeth'.

Haven't had time to play with the '24P 0 Exposure' video. Severe snow storms here around Montreal to contend with.

[Ed Roo]

Thank you for those images. Now I can see what you are referring to and how to spot them!

I will further experiment with the white balance and try what you are suggesting.

24P Master 95 degrees (left side of color bar, yellow-green) -5% Exposure

A Paul Story 24P on Vimeo

Another thing I have learned... I need to purchase a Spyder X and calibrate my monitor.

Thank you, again, everyone! I have learned a lot this week!

[Josh Bass]

Going down a rabbit hole here but a proper color accurate setup, from all my research, requires more than just a probe. You need a high quality monitor capable of displaying the full color and luminance range of your target distribution medium, and for an external monitor, a card/box to take the signal from your NLE to the monitor (i.e. cant just hook up a hdmi cable from external monitor to computer). You should also have perfectly neutral grey walls and a certain type if lighting in the room where you grade because those slight color biases on walls or with lighting will skew your judgement (e.g. slightly blue walls will desensitize you to blue and cause you to compensate by adding blue you dont need, etc.).

Thats for a super nice setup, and probably runs thousands of dollars. Depending on your standards/needs it may well not be worth it but the idea is that if you all of the above you have the highest chance of the most number of viewers seeing what you see on your end.

Without all that you can use the fcpx scopes to at least get it objectively “right”/neutral in regards to white balance and exposure levels. You can use the vs curves to isolate skin and push it more blue or orange til skin tones sit on the skin tone line, waveform to place exposure of various elements in the image where you want them (if you can get good enough separation with the various qualifier tools in FCPx).

[Ed Roo]

Josh! Thank you for those suggestions! I have only watched the Canon video describing the vectorscopes on the XF300 camera. This is the first time I have tried them in FCP X. I didn't know about the skin tone line until you suggested it and I saw it on the scope. That makes life much easier! I also find the curves easier to use than the other methods.

Tell me about the box between the computer and the monitor? I wasn't aware of such a device.
I use a Dell P2715Q monitor with a 2013 MacPro quadcore.

[Ed Roo]

This is as good as I can get it using the vectorscopes...

A Paul Story 24P CC on Vimeo

[Josh Bass]

There's nothing especially wrong with that...at a certain point you get into individual taste (some folks like warmer, some like cooler), and depending on what lights you used, some shots simply can't be balanced as the lights may be missing certain parts of the spectrum, a problem with lower cost LEDs sometimes.

The box I mean would be a Blackmagic box or similar device, but this is for an external monitor, not your typical computer monitor. In other words, a situation where you'd have the video image taking up the entire screen on another monitor, for color grading purposes. That's more for hardcore color correction/grading, which might be overkill for your needs. I'd suspect you have a single computer monitor like most folks, in which case you can investigate how to make it more accurate (perhaps the probe, like you suggested).

If you want a super accurate setup you'd need to do all the detailed stuff I mentioned in my earlier post but for many things, good enough is good enough. I occasionally edit these interview/b-roll pieces for one of my clients and I just eyeball and go by the scopes, and I know that's close enough.

There should be plenty of good grading/correction tutorials for FCPx on youtube, and there's always the free version of DaVinci Resolve if you want to take it further.

[Steven Digges]

Originally Posted by Ed Roo

Thank you, everyone for your comments. They are much appreciated.
Background...
Canon XF-300
LitePanels 1x1 Spot (key, left), Flood (fill, right), ENG3 (hairlight) - all 5600K lights
Sennheiser MKH-416 boomed mic
Concrete blocked walled (light yellow) training room, ~ 16 ft wide and ~ 50 ft deep. I choose to set up widthwise.
Narrator ~ 4 ft from far wall seated in an arm chair (I like to use this type of chair so the subject cannot shift around too much); camera positioned ~ 6 ft from subject.

I had to eyeball the exposure, instead of using the camera's built in scope as I usually do, due to being rushed to set up. I normally like to have access to the room where the interview will be conducted an hour prior to the interview so that I have time to set up and check everything before beginning. Couldn't do that this time. I do have diffusers and frames for all the lights. .

Hi Ed,

I have some suggestions for you that might help next time. This is all meant to be constructive and kind. In my opinion you are just a few steps away from a polished interview.

First, your comment about winging exposure concerned me. No matter how rushed or what method you use exposure is too important not to get perfect. With a single subject under controlled light it can be done very quickly and accurately every time (without scopes). I suggest you practice with your kit until you are confident it is quick and accurate every time.

Lighting: I believe there are still a few golden rules that should not be broken. Yes there is always exceptions but in a single head interview lighting and camera focus should always be based around the subjects eyes. In the case of this elderly gentleman you had a tough pair of eyes to deal with but that does not change the rule. As you made overall exposure adjustments in post his eyes were disregarded to the point of becoming black beady holes in his face. So lets fix it for next time.

Set direction; In a rush set up like that where I know I don't have time to properly light the background I would have turned the set lengthwise. You had a lot of distance that could work in your favor. A 15 x 50 foot room is a ton of space to light a single head. By putting a lot of the lengthwise space behind him with no additional lights it can fall of to pure black and look like a studio black.

Light fixtures; I don't want to offend you but I am about as blunt as a 1x1 LED spot light. Please see some humor in my comments. I would only choose a 1x1 LED Spot bare faced as a key if I truly disliked the subject and wanted to be as harsh on them as I possibly could. It becomes a weapon at that point, not a key light. 1x1 LED panels are deceptive. They are a very hard source if not treated properly. So I would have used the 1x1 flood as key and the 1x1 spot as fill. Remember you had distance working in your favor.

As we age our eyelids and eyebrows begin to droop and cover up a portion of the eyeballs. This was certainly the case for this subject. The best way to light a face with recessed eyes is to use a key light as soft and large as you can for the situation. Placement is also critical. Generally speaking I run the key lower than I would for a healthy young face. Any height that causes shadows on the eyeballs from the eyebrows is too high. Your high and harsh key light hid his eyes and drew big black lines from the bottom of his nose to the corners of his mouth.

Audio; Even though a Senn 416 is a very good mic it is not a good choice for a concrete bunker. But more importantly the overall quality and feel of the video could be easily improved by fixing your audio. The 416 on the subject sounds OK and then you ask a question and the video screams "Oh no, cheep camera guy audio production!" I believe even the technically uneducated viewer picks up on this dramatic audio change. There is an easy fix to take it up a level. You should be speaking into your own mic. Something as simple as a basic vocal mic like a SM58 you can put right in your grill will work. By properly recording your audio the viewer won't even know if someone is conducting an interview or if its you doing everything. Not who or what matters but it will sound much more professional.

Like someone else said I usually don't respond to critique threads. Its hard not to sound snarky. This is my two cents meant to give you food for thought.

Kind Regards,

Steve

[Pete Cofrancesco]

With softboxes you can lower your key to illuminate the eyes without blinding the subject. There is no denying the benefits of soft boxes its just a matter if you can or want to spend the money and time to add them to your kit.

Last year filmed a video testing out led panels with and without light modifiers, never had a chance to fully produce and post it online but here are couple of stills that show the difference in light quality. The same light and camera white balance were used. This particular softbox warms the light.

[Josh Bass]

It's not just about eye shadows but overall softness..the shadows everywhere on the face will be softer and overall the light will "wrap" around the face more with a softer source.

On one man band shoots where your time is limited the built-in diffusion panels on many LEDs or the inexpensive softboxes you can buy, the point being something you can set up quickly, may have to suffice. I have the D-fuse which are maybe 24" squares for my Dracast 1x1s, they look quite nice, to me. Closer you can get the light, the better, to the point of talent discomfort or undesirable hotspots from light being too close.

But if you have more time and are willing to invest, what many folks are doing now (at least on the shoots I'm on locally) is putting their LEDs through a 4x4' diffusion. They make collapsible frames for these can different ways to mount them on regular light stands or C-stands.

Depending on how powerful the LEDs are you may need to gang two or more lights together for the required exposure but the 4x4 gives you a super soft light in a typical position for one interview subject. If you angle it more to the side you can keep the softness/wrap but get a more dynamic look (no so flat lighting across the face).

Some DPs when using a large source like that don't even bother with a fill light/source, just let the key and the edge light plus whatever's bouncing around the room fill in. That'll be shoot specific depending on desired mood, etc. Some clients want a high key look.

And like they said placing the talent as far from the background as you can while still getting the shot size/focal length you want helps with everything.

[Pete Cofrancesco]

I wasn't thrilled with the D-Fuse. They are light, compact and very portable. On the down side they're shallow not providing a lot of diffusion and were too big for my panels (not their fault) as a result I couldn't control the angle. They're expensive with the crates but without them the spill is unacceptable. I preferred the Fovitec great diffusion and amplified the light but it didn't have crates.

Like you said the placement is also important. Ideally 2-3ft range. Not only do hard lights without diffusion produce unflattering light they make the subject uncomfortable. Large part of a successful interview is a relaxed subject.

[Josh Bass]

I think the D-fuse are okay for ease of use, and the softness is quite a bit better than the raw panel. My Dracasts allow the yoke to be adjusted such that you can change the angle...you just have to mount the yoke on your stand 90 degrees compared to where you're used to (yoke has both a vertical AND horizontal mounting hole). Goofy, but works.

I don't know that I'd get 2-3 feet away, that's pretty close...you don't want to crowd the subject. Typically I see others (and follow myself) at least about 6 feet away. Sort of leave a "sacred space" around the subject.

NOW, in some more elaborate interview setups, for instance two or more cams where you see both interviewer and subject, yes, I have seen Divas or LEDs armed out more overhead and much closer but for single cam like the example here I'd back that light off. I mean if it doesn't have the output to be that far away with diffusion you do what you have to but I'd probably start at 6ish feet.