XF100 and 300 paired in low-light recital

[Tim Bakland]

Hi, all,

This should serve as both a sampling for folks wanting to see the two XF300 and XF100 paired (and in a VERY low-light setting) and also as a way for me to solicit any help on best setting to lighten up the picture a bit (if possible).

It was shot at 1080/30p, 50mbs, 30fps, Gain at 6.0db, wb3100 k -- all settings identical on both cameras. The XF100 is the long-shot, the XF300 is the close-ups.

I know, looking back, that I could have gotten more light at 720/24p. But, given what it is now, could I get some advice for how folks might tweak this footage in post to lighten up (especially the close-ups) if possible?

Actually, considering how dark the venue was (it was as dark as appears here pretty much), I'm happy with how it looks. But I'd still love to gain a stop in post if possible:

XF300 and 100 in very dark light on Vimeo

[Michael Chenoweth]

Looks great, Tim!

I'm still dialing in the XF100 - overall, love that little cam. Travels well, shoots better and it's light as a feather.

Footage looks beautiful though.

Mike

[Glen Vandermolen]

Thw wide shots looked correctly exposed, considering the venue. The close up of the singer looked darker than his images in the wide shots.Which camera shot which?
Did you set up any lights?

Still, a nice looking video.

[Tim Bakland]

Hi, Glen,

Thanks - no, no lights -- it was a last minute thing and they didn't want the bother.

The close-up on the singer is darker as the zoom in loses a stop (it's the XF300) -- and his darker hair/complexion doesn't help.

Again -- I'm thinking that 720p would have lightened things up? How about 24p? Any advice would be helpful.

[Spiros Zaharakis]

It hurts when the lower cost camera beats the expensive one unless of course you didn't pay for it.
The XF300 looks darker but this could be just an effect of the zoomed-in lens (the lens becomes darker as you zoom in and you have to compensate using either a slower shutter speed or more gain) but it appears to also be noisier (unless I'm seeing vimeo compression artifacts). Maybe responds better to color grading though.
IMHO the footage is at least half stop under on the XF100 and one stop under on the XF300. I'm sure some of it can be corrected in post.

It seems that we were posting at the same time (I'm very slow at typing).

[Spiros Zaharakis]

I can't see a reason why 720p would have lighten things up.
24p on the other hand could give you a bit more light as a side effect since you will be using 1/48th shutter speed (assume you can dial-in 1/48) but what would definately give you more light is the use of an even slower shutter speed.
Idealy shooting at 24p with a shutter speed of 1/24th (or 1/25 if this is what is available) will give you the best low light performance but even when you shoot at 30p you can gain one more stop of light if you drop your shutter speed to 1/30th.
Why don't you also try the +9db setting? it may be O.K. in an emergency situation (perhaps with the help of Neat Video)

[Lou Bruno]

I think we have to keep in mind that this venue did not have video lighting. No key or backlighting. This is an important issue in video. Basically, these scenes are ambient light.

In the XF-300, one must be careful of grain at about the 12 DB range. I have avoided same by using the top DB with minimal gain via the CUSTOM DB setting in the menu. I have discovered that 10.5 DB can max-out the lighting w/o substantial gain. Again, this is in the main menu-CAMERA SETTING. I assign this DB setting to the "H" DB position.

Nota Bene: I have used the XF-100 with the XF-300 and I can state that the XF-300 is a far superior camera as it relates to sharpness, color and definition. Probably due to the 3-chips and outstanding large lens. The lens is longer than the XF-100, so naturally, the F-stop scenario comes into play. If the same focal length was used in the above video utilizing both cameras, there would be better lighting as the closeup requires more light, even with the better camera-XF-300.

[Glen Vandermolen]

An owner of the XF300 has proven that using 720P in these camera will add to the exposure:

XF300 Resolution Sensitivity Test - Canon XF Video 300/305/100/105 on Vimeo

In this situation, i would have added lights. You say that it was a last minute thing and they didn't want the bother. i would have overruled them and taken the time to set up two lights, just to bring up the general light level. At times like this, you as the DP have to make the decisions that enhance the video image.

Barring that, on the tighter shots, I would have boosted gain. I'd rather live with more noise than a noticeably darker image. But that's just me.

[Les Wilson]

Originally Posted by Tim Bakland

... I'd still love to gain a stop in post if possible:]

Isn't this what the 4:2:2 50MBS is supposed to do for you over the EX which gives you the 1-1.5 stops from the 1/2" chips albeit at a non-BBC compliant 4:2:0 35mbs? That is, you can't get it from the camera but you can get it in post?

Physics being what they are, for sure, other capture modes have better light performance than 1080p in the same camera. If the XF has a built in meter, you can setup a test easily enough. Using the EX1r light meter, I setup a locked down scene and tested each mode. The meter readout is a percentage of 100% and was on the same spot of the scene every time. These were the results showing relative light performance:

720p24: 60%
720p30: 59%
720p60: 40%

1080p60: n/a
1080p30: 49%
1080p24: 55%
1080i60: 48%

[Tim Bakland]

Originally Posted by Glen Vandermolen

In this situation, i would have added lights. You say that it was a last minute thing and they didn't want the bother. i would have overruled them and taken the time to set up two lights, just to bring up the general light level. At times like this, you as the DP have to make the decisions that enhance the video image.

Alas, I tried! But, really, it wasn't an option. They knew the image would be dark. They preferred that to putting up lights at the 11th hour and freaking out the performers (despite my pleas). As it was, there was barely enough time to round up gear, get there for 20 minutes of rehearsal, talk to the director, and go over angles, cues, etc. (They were more concerned with my getting soloists at the right times than they were how it was lit -- so we had to spend our limited set up time doing that.)

So... all that being said and done (and knowing that of course it would have been optimal to light the stage), are there any suggestions for how best to post-process the image to best boost it without too much graininess?

Again: I'm fairly happy with things considering the ambient light. And so are the musicians. But, I'd love to hear if there are suggestions for good tricks in post.

[Robert John]

I have a question Tim, were you zoomed in on the close ups with the xf300? were both cameras opened up all the way to f 1.6? The reason I ask is that if you were zoomed in with the xf300 then it would have been at like 2.8 and the xf100 would have been at 1.6. Or did I miss something here.

[Spiros Zaharakis]

Originally Posted by Glen Vandermolen

An owner of the XF300 has proven that using 720P in these camera will add to the exposure:

XF300 Resolution Sensitivity Test - Canon XF Video 300/305/100/105 on Vimeo

In this situation, i would have added lights. You say that it was a last minute thing and they didn't want the bother. i would have overruled them and taken the time to set up two lights, just to bring up the general light level. At times like this, you as the DP have to make the decisions that enhance the video image.

Barring that, on the tighter shots, I would have boosted gain. I'd rather live with more noise than a noticeably darker image. But that's just me.

Quite impressive!
Technicaly there shouldn't be a difference unless the camera does some kind of pixel binning or something like that that raises the sensor's sensitivity. Or it could be something else that I cannot think of.
I'm now curious to see if this is the case with the XF100 too.

[Tim Bakland]

Originally Posted by Robert John

I have a question Tim, were you zoomed in on the close ups with the xf300? were both cameras opened up all the way to f 1.6? The reason I ask is that if you were zoomed in with the xf300 then it would have been at like 2.8 and the xf100 would have been at 1.6. Or did I miss something here.

Yes, that's exactly the case.

Both cameras were completely opened up. And yes, the XF300 was zooming (that's what they wanted: one camera on close-ups and one on long shot).

So, all I'm wondering is if folks have a prescribed/recommended method of brightening in post.

[Charles W. Hull]

...could I get some advice for how folks might tweak this footage in post to lighten up (especially the close-ups) if possible?

Tim, don't know what editor you use but add any effect that can control gain, offset, and probably gamma. Or use AE. But my favorite is Cineform First Light - you could quickly and easily balance all the clips from both cameras with First Light.

[Tim Bakland]

Thanks - FCP7.