Unexpected natural light pulsing at DVinfo.net

Nick Mirro · November 25th, 2020, 11:34 PM

Hi. I accidentally left the XC15 in aperture priority F8 and autofocus. The clip has 2 parts. The first 20 seconds (no audio) was shot nearly directly into the sun. A couple of steps left and direct intense sun reflection appears on the water.

The autofocus swings wildly. The exposure is pulsing up and down.

The last 12 seconds (seems normal) is shot facing the opposite direction, though with the same settings.

Other points are...

1. The audio is bad in part 1 because one of the transmitters was accidentally left on 12V.
2. Both are shot in log with a Canon 1D LUT applied
3. No ND filter into the sun

I'm trying to understand this camcorder. Is the intense light too much for it? Could the audio voltage issue have thrown the exposure and focus out of whack?

Thanks much for any thoughts on this.

Paul R Johnson · November 26th, 2020, 01:50 AM

Just the mega bright reflections in the sensor auto area that are identified as the overall brightness changing, and anything auto would be confused. Autofocus and auto exposure is historically known for these issues, although in fairness they have been getting better. I assume you weren’t looking at the image when you were shooting?

Every camera responds differently. One of mine jitters the brightness, a sort of wobble that looks odd when it’s adjusting slightly slower than the reflections and can’t keep up. It’s old and over the years I just know when it will play up. Sunlight on waves is where it’s worst. Those little pinprick hotspots are probably your danger area and it will be fine on everything else. I’m old, so autofocus has never worked for me. Always unreliable.

Bryan Worsley · November 26th, 2020, 02:07 AM

Plus, there's some camera movement.

Edit: Yep, stepping through the downloaded video, I'd say it's camera movements catching flare and specular reflection, off the water, from the sunlight coming through trees on the right, just out of frame.

Nick Mirro · November 27th, 2020, 12:09 AM

This is so helpful, thanks much! So in hindsight, manual exposure and focus.

It is log and highs were not terribly blown out. Would ND have helped the first scene with the high contrast?

Yes definitely on the shaking. How does that pick up flare? Also, are you saying the movement also magnifies the specular reflection? (which I just looked up :)

Pete Cofrancesco · November 27th, 2020, 12:47 AM

The pulsing is caused by auto exposure, shooting in manual would solve that issue.

ND filter wouldn't fix this or give you more dynamic range. ND filters are still useful because in bright conditions there are only 3 ways to reduce the light. The first is the aperture but usually its not possible to close the aperture enough and other times you might want the aperture to be wide open for a shallow dof. The second way is the shutter speed. Ideally the shutter speed should be no more than twice the frame rate. If it's set too high it can look odd. So that's why ND filters are used to control the exposure so the aperture and shutter can be left at their optimal settings.

Shooting outdoors is a specialty unto itself.

Bryan Worsley · November 27th, 2020, 07:41 PM

Still think it's primarily flare/glare 'haze' coming from strong specular, and diffuse, reflections off the water on the far right side. The left side of the frame is far less affected. Actually, stepping through the video again, I'm less inclined to think it's camera movement, as there are also strong flashes when the camera is static. Could easily be due to the sunlight breaking through patchy cloud passing over though. There's also what could be a flare spot (light purple) on the tree, just up-right from center, unless it's dust or a dried water spot, which would be more conspicuous at higher aperture.

Paul R Johnson · November 28th, 2020, 02:45 AM

We’re all right. Those reflections of the sun are at almost sun intensity, so when they are in the cameras auto exposure zone they Indicate the exposure is too high and it closes the aperture, or sensitivity depending on setting. As the water moves, these triggers the. Vanish and the camera compensates. I have a small handicam that is terrible in its control. Far too quick and sharp and in some circumstances it wobbles or flickers. Fast moving clouds ramp the brightness up and down very noticeably. Another camera is very slow to compensate and much better. Churches cause me problems with windows with stained glass. As clouds and sun play outside, the changes inside set off the small camera, but annoyingly it’s manual control is terrible.

This phenomenon is quite normal for some cameras, and you work around them.

Doug Jensen · November 28th, 2020, 01:57 PM

Quote:

Originally Posted by Bryan Worsley

Still think it's primarily flare/glare 'haze' coming from strong specular, and diffuse, reflections off the water on the far right side.

You are absolutely correct. The issue seen here has nothing to do with exposure fluctuation.

Paul R Johnson · November 28th, 2020, 04:25 PM

If you cover up the right hand side of the frame, the exposure level of the entire image changes. Something is clearly influencing the camera's exposure. Traditonal point sources that cause flaring tend to only influence parts of the whole frame. This appears to be a frame wide artefact, hence my belief it's electronic in nature, rather than a physical iris movement. Should be easy to replicate in the studio with a grey card and a torch. Sorry Doug, we will have to differ in our opinions.

Bryan Worsley · November 28th, 2020, 05:16 PM

The left side of the image is far less effected though. Examining the histogram behavior (in Resolve) you can clearly see that the luminance 'pulsing' is much stronger in the cropped right third of the image compared to the left, and that the pulses are irregular. Also if it were due to in-camera auto-exposure fluctuation you would expect to see marked fluctuation in the height (intensity) of the specular peak (at 255, when examined at 'Full' data level), and you don't - it stays pretty constant.

Doug Jensen · November 28th, 2020, 05:29 PM

Bryan, you are exactly correct again. This is not an auto-exposure problem. If I could download the original file and run it through my waveform monitor it would be obvious.

Bryan Worsley · November 28th, 2020, 07:16 PM

Needless to say, nigh-on impossible to fix, successfully, in post.

Andrew Smith · November 29th, 2020, 04:00 AM

I'm late to the discussion, but wanted to say that my first thoughts were that the wind was moving leaves/branches of a nearby tree and this was causing the rapidly blocking and then un-blocking of sunlight directly hitting the lens.

You may find that the de-flicker filter from RE:Vision Effects fixes it, per the example at 28 seconds on the demo video. Might as well give it a try.

Andrew

Bryan Worsley · November 29th, 2020, 11:03 AM

I tried it with a VirtualDub 'deflicker' plugin, without success:

https://www.compression.ru/video/def.../index_en.html

Could also try their 'Smart Brightness and Contrast' plugin:

https://www.compression.ru/video/sma.../index_en.html

I had a degree of success with it for smoothing out mild variances in brightness in some dive videos, but I doubt it would really help in this case - for one thing, to work well the brightness needs to fairly uniform across the frame.

Resolve (which I know the OP uses) has a 'Dehaze' FX filter, but it is only available (without a bold logo watermark( in the Studio version. 'Dehaze' filters can work quite well for treating/dampening static flare haze. The problem in this case is compounded by the fact that these irregular haze bursts are appearing over a complex background.

There again, the OP hasn't asked for a fix.

Nick Mirro · November 30th, 2020, 11:46 PM

I tried the dehaze filter but no benefit. The built-in deflicker effect seems to help quite a bit, best with the default settings - though not enough to make it usable.

I understand ND would not help. Also that my thoughts about the transmitter voltage made no sense at all :)

Quote:

Also if it were due to in-camera auto-exposure fluctuation you would expect to see marked fluctuation in the height (intensity) of the specular peak (at 255, when examined at 'Full' data level), and you don't - it stays pretty constant.

Not sure I understand. Wouldn't this be gone if I shot in manual? That would lock the aperture and gain.

If I display the qualifier, the points of intense reflection in the parade dance up and down.

I see the right being more affected. Shadows on the left, equally dark to right are less over exposed during pulses. How is that? Maybe (as I think you've been saying) the pulsing specular reflection (wind moving trees and gravity moving water) is the light source the sensor is responding to. Maybe we're blinded to the effect by looking away from the annoying intense glare (which there was). The poor camera sensor had to look.

Are such pulses there, but we can't make them out bare-eyed? The inverse square law throws much less light on the opposite bank (speculating). Maybe a good way to damage a sensor :-/

November 25th, 2020, 11:34 PM	#1
Nick Mirro Regular Crew Join Date: Jun 2008 Location: Dallas, Texas Posts: 119	Unexpected natural light pulsing Hi. I accidentally left the XC15 in aperture priority F8 and autofocus. The clip has 2 parts. The first 20 seconds (no audio) was shot nearly directly into the sun. A couple of steps left and direct intense sun reflection appears on the water. The autofocus swings wildly. The exposure is pulsing up and down. The last 12 seconds (seems normal) is shot facing the opposite direction, though with the same settings. Other points are... 1. The audio is bad in part 1 because one of the transmitters was accidentally left on 12V. 2. Both are shot in log with a Canon 1D LUT applied 3. No ND filter into the sun I'm trying to understand this camcorder. Is the intense light too much for it? Could the audio voltage issue have thrown the exposure and focus out of whack? Thanks much for any thoughts on this. YouTube (Short URL)

November 26th, 2020, 01:50 AM	#2
Paul R Johnson Inner Circle Join Date: Nov 2005 Location: Lowestoft - UK Posts: 4,030	Re: Unexpected natural light pulsing Just the mega bright reflections in the sensor auto area that are identified as the overall brightness changing, and anything auto would be confused. Autofocus and auto exposure is historically known for these issues, although in fairness they have been getting better. I assume you weren’t looking at the image when you were shooting? Every camera responds differently. One of mine jitters the brightness, a sort of wobble that looks odd when it’s adjusting slightly slower than the reflections and can’t keep up. It’s old and over the years I just know when it will play up. Sunlight on waves is where it’s worst. Those little pinprick hotspots are probably your danger area and it will be fine on everything else. I’m old, so autofocus has never worked for me. Always unreliable.

November 26th, 2020, 02:07 AM	#3
Bryan Worsley Major Player Join Date: Oct 2005 Location: Canada Posts: 439	Re: Unexpected natural light pulsing Plus, there's some camera movement. Edit: Yep, stepping through the downloaded video, I'd say it's camera movements catching flare and specular reflection, off the water, from the sunlight coming through trees on the right, just out of frame. Last edited by Bryan Worsley; November 26th, 2020 at 09:21 AM.

November 27th, 2020, 12:09 AM	#4
Nick Mirro Regular Crew Join Date: Jun 2008 Location: Dallas, Texas Posts: 119	Re: Unexpected natural light pulsing This is so helpful, thanks much! So in hindsight, manual exposure and focus. It is log and highs were not terribly blown out. Would ND have helped the first scene with the high contrast? Yes definitely on the shaking. How does that pick up flare? Also, are you saying the movement also magnifies the specular reflection? (which I just looked up :)

November 27th, 2020, 12:47 AM	#5
Pete Cofrancesco Inner Circle Join Date: May 2004 Location: USA Posts: 3,004	Re: Unexpected natural light pulsing The pulsing is caused by auto exposure, shooting in manual would solve that issue. ND filter wouldn't fix this or give you more dynamic range. ND filters are still useful because in bright conditions there are only 3 ways to reduce the light. The first is the aperture but usually its not possible to close the aperture enough and other times you might want the aperture to be wide open for a shallow dof. The second way is the shutter speed. Ideally the shutter speed should be no more than twice the frame rate. If it's set too high it can look odd. So that's why ND filters are used to control the exposure so the aperture and shutter can be left at their optimal settings. Shooting outdoors is a specialty unto itself.

November 27th, 2020, 07:41 PM	#6
Bryan Worsley Major Player Join Date: Oct 2005 Location: Canada Posts: 439	Re: Unexpected natural light pulsing Still think it's primarily flare/glare 'haze' coming from strong specular, and diffuse, reflections off the water on the far right side. The left side of the frame is far less affected. Actually, stepping through the video again, I'm less inclined to think it's camera movement, as there are also strong flashes when the camera is static. Could easily be due to the sunlight breaking through patchy cloud passing over though. There's also what could be a flare spot (light purple) on the tree, just up-right from center, unless it's dust or a dried water spot, which would be more conspicuous at higher aperture.

November 28th, 2020, 02:45 AM	#7
Paul R Johnson Inner Circle Join Date: Nov 2005 Location: Lowestoft - UK Posts: 4,030	Re: Unexpected natural light pulsing We’re all right. Those reflections of the sun are at almost sun intensity, so when they are in the cameras auto exposure zone they Indicate the exposure is too high and it closes the aperture, or sensitivity depending on setting. As the water moves, these triggers the. Vanish and the camera compensates. I have a small handicam that is terrible in its control. Far too quick and sharp and in some circumstances it wobbles or flickers. Fast moving clouds ramp the brightness up and down very noticeably. Another camera is very slow to compensate and much better. Churches cause me problems with windows with stained glass. As clouds and sun play outside, the changes inside set off the small camera, but annoyingly it’s manual control is terrible. This phenomenon is quite normal for some cameras, and you work around them.

November 28th, 2020, 04:25 PM	#9
Paul R Johnson Inner Circle Join Date: Nov 2005 Location: Lowestoft - UK Posts: 4,030	Re: Unexpected natural light pulsing If you cover up the right hand side of the frame, the exposure level of the entire image changes. Something is clearly influencing the camera's exposure. Traditonal point sources that cause flaring tend to only influence parts of the whole frame. This appears to be a frame wide artefact, hence my belief it's electronic in nature, rather than a physical iris movement. Should be easy to replicate in the studio with a grey card and a torch. Sorry Doug, we will have to differ in our opinions.

November 28th, 2020, 05:16 PM	#10
Bryan Worsley Major Player Join Date: Oct 2005 Location: Canada Posts: 439	Re: Unexpected natural light pulsing The left side of the image is far less effected though. Examining the histogram behavior (in Resolve) you can clearly see that the luminance 'pulsing' is much stronger in the cropped right third of the image compared to the left, and that the pulses are irregular. Also if it were due to in-camera auto-exposure fluctuation you would expect to see marked fluctuation in the height (intensity) of the specular peak (at 255, when examined at 'Full' data level), and you don't - it stays pretty constant. Last edited by Bryan Worsley; November 28th, 2020 at 07:17 PM.

November 28th, 2020, 05:29 PM	#11
Doug Jensen Vortex Media Join Date: Jan 2009 Location: Florida Posts: 3,423	Re: Unexpected natural light pulsing Bryan, you are exactly correct again. This is not an auto-exposure problem. If I could download the original file and run it through my waveform monitor it would be obvious. __________________ Vortex Media http://www.vortexmedia.com/ Sony FS7, F55, and XDCAM training videos, field guides, and other production tools Last edited by Doug Jensen; November 29th, 2020 at 06:32 AM.

November 28th, 2020, 07:16 PM	#12
Bryan Worsley Major Player Join Date: Oct 2005 Location: Canada Posts: 439	Re: Unexpected natural light pulsing Needless to say, nigh-on impossible to fix, successfully, in post. Last edited by Bryan Worsley; November 28th, 2020 at 07:54 PM.

November 29th, 2020, 04:00 AM	#13
Andrew Smith Equal Opportunity Offender Join Date: Feb 2009 Location: Brisbane, Australia Posts: 3,026	Re: Unexpected natural light pulsing I'm late to the discussion, but wanted to say that my first thoughts were that the wind was moving leaves/branches of a nearby tree and this was causing the rapidly blocking and then un-blocking of sunlight directly hitting the lens. You may find that the de-flicker filter from RE:Vision Effects fixes it, per the example at 28 seconds on the demo video. Might as well give it a try. DEFlicker Overview Reel - YouTube Andrew

November 29th, 2020, 11:03 AM	#14
Bryan Worsley Major Player Join Date: Oct 2005 Location: Canada Posts: 439	Re: Unexpected natural light pulsing I tried it with a VirtualDub 'deflicker' plugin, without success: https://www.compression.ru/video/def.../index_en.html Could also try their 'Smart Brightness and Contrast' plugin: https://www.compression.ru/video/sma.../index_en.html I had a degree of success with it for smoothing out mild variances in brightness in some dive videos, but I doubt it would really help in this case - for one thing, to work well the brightness needs to fairly uniform across the frame. Resolve (which I know the OP uses) has a 'Dehaze' FX filter, but it is only available (without a bold logo watermark( in the Studio version. 'Dehaze' filters can work quite well for treating/dampening static flare haze. The problem in this case is compounded by the fact that these irregular haze bursts are appearing over a complex background. There again, the OP hasn't asked for a fix. Last edited by Bryan Worsley; November 29th, 2020 at 05:58 PM.