Ugly Hedgehog - Photography Forum
Where do these brown streaks come from?
Oct 17, 2020 12:20:55   #
Verryl
 
I am building a very large (50x75 foot) two level HO model railroad. After almost 6 years we are to the scenery stage, and we have done the easiest scenery first, the upper level that is open and easier to access. The lower level has the upper level hanging over it, and there is about 20 inches between them. The lower level is lighted with fluorescent light fixtures mounted on the underneath side of the upper level.

I took a few photos yesterday, mostly of the ABS (automatic block signals), and one was of tracks and a single signal on a mast (pole) back at the dark end of one bench. The fluorescent lights were on to light the scene. I used a Panasonic Lumix G9 mirrorless 4/3 digital camera set on auto exposure mode (iA) and set to use the G9's semiautomatic focus stacking mode, which they call "post focus." Post focus works by taking a 1 second video at about 30 frames/sec while the focus is shifted over the whole depth of field. Then as a manual step I can assemble the 30 frames into a single image using the in-focus parts of the individual frames.

I can often take such a 1 sec video hand held if I can capture the scene without body contortions, because the camera and 12-60mm Leica lens have excellent image stabilization, but for this shot I would have had to lean over the lower bench to shoot along the bench. So I set the camera on a sturdy mini-tripod pointing along the bench under the fluorescent lamps with the lone signal, the object of interest for the shot, near the center of the frame. Here is a result, one of many tries at various angles. All showed the two brown "interference bands" horizontally across each exposure. I took about 20 shots varying the distance, focal length,and angle to the length of bench. All were with the iA auto exposure setting which is required for the "post focus" video setting, an exposure burst setting. All the photos lighted by fluorescents showed the bands, both in the 1 sec originating video and the assembled JPEG. The band intensity varied in the various shots depending on angle or focal length, but could not be avoided. The second attached photo is of the over hanging fluorescent lights, and it also shows bands. The camera location for the first photo is about midway between the bench and the overhanging lights, and about under the end of the first of the 4 tubes running along the bench length. The view of the camera was from the aisle about 3 feet from the center of the bench, but the camera was pointed roughly along the lower bench.

I finally gave up lighting the scene with the bench overhead fluorescent lights, and I turned them off and lighted the scene from the side with two LED lights with diffusers over each lamp. That is the third photo, again a focus stacked one, but with no brown bands, so it seems sure that the brown bands were from the fluorescents. When I got back to my computer, I looked at previous images I had taken of scenes on the lower level. They were all taken more or less perpendicular to the lower bench, and on close inspection they showed two very faint brown bands, which were so faint that I had not noticed them before. Therefore, aligning the camera with the bench has the effect of intensifying the bands, but shooting perpendicular to a lower bench seems to always cause bands, although they are more faint. They probably are caused by the closely overhanging lights, but I am not sure. But what else could it be?

I am convinced that the brown bands are an artifact of the fluorescent lighting only 12 to 16 inches above the camera (on the tripod), so LED lighting is a cure. With this post I am just seeking an explanation of how fluorescents close to the scene cause the banding. It must be due to the close proximity of the lights to the camera and the bench, and the video frame rate is 30 frames/sec, and I imagine the natural flicker frequency of the fluorescent lights is 60 cycles/sec.

I don't get banding on the upper level, where the room lighting is fluorescent also, but those lights hang several feet over the upper level benches, and there is a lot more indirect lighting from more distant fixtures and from reflections from the room ceiling and walls. Such reflected light must be diffuse light bouncing all around the room.

Note: I have not bothered to do post processing on the banded images.







| Reply
Oct 17, 2020 12:22:31   #
rmalarz Loc: Tempe, Arizona
 
The lighting combined with shutter speed.
--Bob
Verryl wrote:
I am building a very large (50x75 foot) two level HO model railroad. After almost 6 years we are to the scenery stage, and we have done the easiest scenery first, the upper level that is open and easier to access. The lower level has the upper level hanging over it, and there is about 20 inches between them. The lower level is lighted with fluorescent light fixtures mounted on the underneath side of the upper level.

I took a few photos yesterday, mostly of the ABS (automatic block signals), and one was of tracks and a single signal on a mast (pole) back at the dark end of one bench. The fluorescent lights were on to light the scene. I used a Panasonic Lumix G9 mirrorless 4/3 digital camera set on auto exposure mode (iA) and set to use the G9's semiautomatic focus stacking mode, which they call "post focus." Post focus works by taking a 1 second video at about 30 frames/sec while the focus is shifted over the whole depth of field. Then as a manual step I can assemble the 30 frames into a single image using the in-focus parts of the individual frames.

I can often take such a 1 sec video hand held if I can capture the scene without body contortions, because the camera and 12-60mm Leica lens have excellent image stabilization, but for this shot I would have had to lean over the lower bench to shoot along the bench. So I set the camera on a sturdy mini-tripod pointing along the bench under the fluorescent lamps with the lone signal, the object of interest for the shot, near the center of the frame. Here is a result, one of many tries at various angles. All showed the two brown "interference bands" horizontally across each exposure. I took about 20 shots varying the distance, focal length,and angle to the length of bench. All were with the iA auto exposure setting which is required for the "post focus" video setting, an exposure burst setting. All the photos lighted by fluorescents showed the bands, both in the 1 sec originating video and the assembled JPEG. The band intensity varied in the various shots depending on angle or focal length, but could not be avoided. The second attached photo is of the over hanging fluorescent lights, and it also shows bands. The camera location for the first photo is about midway between the bench and the overhanging lights, and about under the end of the first of the 4 tubes running along the bench length. The view of the camera was from the aisle about 3 feet from the center of the bench, but the camera was pointed roughly along the lower bench.

I finally gave up lighting the scene with the bench overhead fluorescent lights, and I turned them off and lighted the scene from the side with two LED lights with diffusers over each lamp. That is the third photo, again a focus stacked one, but with no brown bands, so it seems sure that the brown bands were from the fluorescents. When I got back to my computer, I looked at previous images I had taken of scenes on the lower level. They were all taken more or less perpendicular to the lower bench, and on close inspection they showed two very faint brown bands, which were so faint that I had not noticed them before. Therefore, aligning the camera with the bench has the effect of intensifying the bands, but shooting perpendicular to a lower bench seems to always cause bands, although they are more faint. They probably are caused by the closely overhanging lights, but I am not sure. But what else could it be?

I am convinced that the brown bands are an artifact of the fluorescent lighting only 12 to 16 inches above the camera (on the tripod), so LED lighting is a cure. With this post I am just seeking an explanation of how fluorescents close to the scene cause the banding. It must be due to the close proximity of the lights to the camera and the bench, and the video frame rate is 30 frames/sec, and I imagine the natural flicker frequency of the fluorescent lights is 60 cycles/sec.

I don't get banding on the upper level, where the room lighting is fluorescent also, but those lights hang several feet over the upper level benches, and there is a lot more indirect lighting from more distant fixtures and from reflections from the room ceiling and walls. Such reflected light must be diffuse light bouncing all around the room.

Note: I have not bothered to do post processing on the banded images.
I am building a very large (50x75 foot) two level ... (show quote)

| Reply
Oct 17, 2020 12:24:37   #
MSW
 
lens flare?

| Reply
 
 
Oct 17, 2020 12:42:48   #
Longshadow Loc: Audubon, PA
 
Fluorescent flicker v. shutter speed?

| Reply
Oct 17, 2020 12:43:12   #
larryepage
 
Verryl wrote:
I am building a very large (50x75 foot) two level HO model railroad. After almost 6 years we are to the scenery stage, and we have done the easiest scenery first, the upper level that is open and easier to access. The lower level has the upper level hanging over it, and there is about 20 inches between them. The lower level is lighted with fluorescent light fixtures mounted on the underneath side of the upper level.

I took a few photos yesterday, mostly of the ABS (automatic block signals), and one was of tracks and a single signal on a mast (pole) back at the dark end of one bench. The fluorescent lights were on to light the scene. I used a Panasonic Lumix G9 mirrorless 4/3 digital camera set on auto exposure mode (iA) and set to use the G9's semiautomatic focus stacking mode, which they call "post focus." Post focus works by taking a 1 second video at about 30 frames/sec while the focus is shifted over the whole depth of field. Then as a manual step I can assemble the 30 frames into a single image using the in-focus parts of the individual frames.

I can often take such a 1 sec video hand held if I can capture the scene without body contortions, because the camera and 12-60mm Leica lens have excellent image stabilization, but for this shot I would have had to lean over the lower bench to shoot along the bench. So I set the camera on a sturdy mini-tripod pointing along the bench under the fluorescent lamps with the lone signal, the object of interest for the shot, near the center of the frame. Here is a result, one of many tries at various angles. All showed the two brown "interference bands" horizontally across each exposure. I took about 20 shots varying the distance, focal length,and angle to the length of bench. All were with the iA auto exposure setting which is required for the "post focus" video setting, an exposure burst setting. All the photos lighted by fluorescents showed the bands, both in the 1 sec originating video and the assembled JPEG. The band intensity varied in the various shots depending on angle or focal length, but could not be avoided. The second attached photo is of the over hanging fluorescent lights, and it also shows bands. The camera location for the first photo is about midway between the bench and the overhanging lights, and about under the end of the first of the 4 tubes running along the bench length. The view of the camera was from the aisle about 3 feet from the center of the bench, but the camera was pointed roughly along the lower bench.

I finally gave up lighting the scene with the bench overhead fluorescent lights, and I turned them off and lighted the scene from the side with two LED lights with diffusers over each lamp. That is the third photo, again a focus stacked one, but with no brown bands, so it seems sure that the brown bands were from the fluorescents. When I got back to my computer, I looked at previous images I had taken of scenes on the lower level. They were all taken more or less perpendicular to the lower bench, and on close inspection they showed two very faint brown bands, which were so faint that I had not noticed them before. Therefore, aligning the camera with the bench has the effect of intensifying the bands, but shooting perpendicular to a lower bench seems to always cause bands, although they are more faint. They probably are caused by the closely overhanging lights, but I am not sure. But what else could it be?

I am convinced that the brown bands are an artifact of the fluorescent lighting only 12 to 16 inches above the camera (on the tripod), so LED lighting is a cure. With this post I am just seeking an explanation of how fluorescents close to the scene cause the banding. It must be due to the close proximity of the lights to the camera and the bench, and the video frame rate is 30 frames/sec, and I imagine the natural flicker frequency of the fluorescent lights is 60 cycles/sec.

I don't get banding on the upper level, where the room lighting is fluorescent also, but those lights hang several feet over the upper level benches, and there is a lot more indirect lighting from more distant fixtures and from reflections from the room ceiling and walls. Such reflected light must be diffuse light bouncing all around the room.

Note: I have not bothered to do post processing on the banded images.
I am building a very large (50x75 foot) two level ... (show quote)


Fluorescent lamps "strobe" 120 times per second...they do not provide continuous light. The dark band represents an interval that the lamps were dark. Some newer cameras have a function which synchronizes the shutter open time with the repetition of the lighting, eliminating the bands. If this function is not available. the best option is to make sure to use a long shutter speed (1/30 second or longer), which will at least mostly average out the lighting across the frame.

When using the focus stacking feature, the repetition rate of the exposures synchs with the repetition rate of the flicker, almost insuring that these "flicker bands" get accentuated in your final exposure. Unless you camera has a 'Flicker Reduction' feature which works with your focus stacking function, it may be very difficult to eliminate the bands. If you can change the rate of capturing the focus stack images to somethign other than a multiple of 30, you might be able to reduce the effect.

So...some manual research and some experimentation may be necessary here.

Tell us a little bit about your railroad. Prototype? Freelance? What road? What period? If I ever get mine built, it will represent the ATSF in eastern Arizona in the summer of 1970.

| Reply
Oct 17, 2020 13:50:54   #
SalvageDiver Loc: Huntington Beach CA
 
Verryl wrote:
I am building a very large (50x75 foot) two level HO model railroad. After almost 6 years we are to the scenery stage, and we have done the easiest scenery first, the upper level that is open and easier to access. The lower level has the upper level hanging over it, and there is about 20 inches between them. The lower level is lighted with fluorescent light fixtures mounted on the underneath side of the upper level.

I took a few photos yesterday, mostly of the ABS (automatic block signals), and one was of tracks and a single signal on a mast (pole) back at the dark end of one bench. The fluorescent lights were on to light the scene. I used a Panasonic Lumix G9 mirrorless 4/3 digital camera set on auto exposure mode (iA) and set to use the G9's semiautomatic focus stacking mode, which they call "post focus." Post focus works by taking a 1 second video at about 30 frames/sec while the focus is shifted over the whole depth of field. Then as a manual step I can assemble the 30 frames into a single image using the in-focus parts of the individual frames.

I can often take such a 1 sec video hand held if I can capture the scene without body contortions, because the camera and 12-60mm Leica lens have excellent image stabilization, but for this shot I would have had to lean over the lower bench to shoot along the bench. So I set the camera on a sturdy mini-tripod pointing along the bench under the fluorescent lamps with the lone signal, the object of interest for the shot, near the center of the frame. Here is a result, one of many tries at various angles. All showed the two brown "interference bands" horizontally across each exposure. I took about 20 shots varying the distance, focal length,and angle to the length of bench. All were with the iA auto exposure setting which is required for the "post focus" video setting, an exposure burst setting. All the photos lighted by fluorescents showed the bands, both in the 1 sec originating video and the assembled JPEG. The band intensity varied in the various shots depending on angle or focal length, but could not be avoided. The second attached photo is of the over hanging fluorescent lights, and it also shows bands. The camera location for the first photo is about midway between the bench and the overhanging lights, and about under the end of the first of the 4 tubes running along the bench length. The view of the camera was from the aisle about 3 feet from the center of the bench, but the camera was pointed roughly along the lower bench.

I finally gave up lighting the scene with the bench overhead fluorescent lights, and I turned them off and lighted the scene from the side with two LED lights with diffusers over each lamp. That is the third photo, again a focus stacked one, but with no brown bands, so it seems sure that the brown bands were from the fluorescents. When I got back to my computer, I looked at previous images I had taken of scenes on the lower level. They were all taken more or less perpendicular to the lower bench, and on close inspection they showed two very faint brown bands, which were so faint that I had not noticed them before. Therefore, aligning the camera with the bench has the effect of intensifying the bands, but shooting perpendicular to a lower bench seems to always cause bands, although they are more faint. They probably are caused by the closely overhanging lights, but I am not sure. But what else could it be?

I am convinced that the brown bands are an artifact of the fluorescent lighting only 12 to 16 inches above the camera (on the tripod), so LED lighting is a cure. With this post I am just seeking an explanation of how fluorescents close to the scene cause the banding. It must be due to the close proximity of the lights to the camera and the bench, and the video frame rate is 30 frames/sec, and I imagine the natural flicker frequency of the fluorescent lights is 60 cycles/sec.

I don't get banding on the upper level, where the room lighting is fluorescent also, but those lights hang several feet over the upper level benches, and there is a lot more indirect lighting from more distant fixtures and from reflections from the room ceiling and walls. Such reflected light must be diffuse light bouncing all around the room.

Note: I have not bothered to do post processing on the banded images.
I am building a very large (50x75 foot) two level ... (show quote)


I would agree with Bob that shutter speed is likely the problem with the following explanations.

1. The flicker frequency of a fluorescent bulb is 120Hz. So you need a shutter speed of 1/120s or slower to get 1 full cycle from the bulb. However, if the shutter speed is above the flash sync shutter speed (~1/250s) the two shutter curtains are only partially open as they sweeps over your sensor, creating a banding similar to what you see in your images. The banding width varies with shutter speed. Here is a video showing how this works.

https://www.youtube.com/watch?v=fyqbIuTzRVI

2. The Post Focus mode takes 30 frames/s in either 4k or 6k video. The frame rate is not the same as shutter speed. The 30 frames/s does not mean the shutter speed was set at 1/30s. While it takes 30 frames /sec, the exposure time can be anything faster than 1/30s. Since you used 'automatic exposure' the shutter speed could be anything, depending on the scene brightness (eV), iso and fn.

If you post your images with 'store original' selected others can view your exif data and confirm that shutter speed with the fluorescent lights is your problem.

| Reply
Oct 17, 2020 13:54:29   #
Verryl
 
I call it the Wyoming Division (of the Union Pacific) from Cheyenne to Ogden. So it has two major grades, Sherman Hill just west of Cheyenne and the Wasatch Mountains just east of Ogden. I model the 1957 era, so I have Challengers (~10 ea), Big Boys (5), 9000's (2), and FEF (5) plus smaller steam locos for locals and yard work; also early diesels, and all three gas electric turbines.

It is built specifically for operations, which normally (sans Covid) we have one per month plus a 3 day meet with 2 all day sessions for up to 50 operators. The monthly sessions draw up to 25 in winter, 18 or so in summers. I live in Sedona, AZ, and the layout is 17 miles away in a little horse community of Cornville, AZ.

It has double track mains each about 1006 feet long, plus a 186 foot long bidirectional Oregon Short Line to "Portland," a subsidiary staging yard of 9 tracks plus a through track to a turning loop for the City of Portland passenger train. Main staging is 12 tracks including a through track and loop for Cities of LA, SF, and St. Louis. These 4 City trains plus other local mixed trains run per a time table. Freights are up to 17 cars long plus engine, tinder, and caboose. Main staging is in 4 quarters, each section can hold a full train. There is also an adjacent 5 track stub staging we use for Omaha, Nebraska. There is also Track #3, the Harriman Cutoff, from the wye at Speer just SW of Cheyenne to Dale Junction as an alternate bidirectional route up the eastern slope of Sherman Hill. One third of #3 has an 18 foot double track section. The mains, #1 and #2, and #3 all have ABS. Operations is per Rule 251D with ABS, so the Dispatcher is a busy, but easy job. We have fixed phones, 8 on each of the two mushroom levels to the Dispatcher, and 10 FRS for DS to passenger superintendent, some YM's and other key positions so the Dispatcher can send orders to these "Temporary Tower Operators."

I have developed a unique 4-card car forwarding system (single move car cards, block cards, loco cards, and a one sheet train order) that is simple, easy to learn and follow, and more prototypical and random and flexible than repetitive 4-cycle car cards. We have through trains, some with switching, some without; switching locals; coal drags; LCL work, and 4 busy yards, 2 with a YM and Classification Foreman (Laramie and Green River); 2 refineries, 6 coal mines (3 each at Hanna and Rock Springs); a switching puzzle--13 industries at Downtown Ogden; 906 spots to be switched.

One day op sessions are from 9 to about 4:30, but the layout is big enough that only Laramie, Green River, and staging are necessary jobs; other operators can come and go as they need to. And operators can switch to more than one job in the one day and two day sessions.

See more at http://www.wyomingdivision.org/

Verryl

| Reply
 
 
Oct 17, 2020 14:00:07   #
Verryl
 
Thanks Mike.

| Reply
Oct 17, 2020 14:23:18   #
Ourspolair
 
Yup - it is the frame rate combined with the fluorescent strobing. Also exacerbated by the video mode using an electronic shutter (hence the irregular patterns on the images. If you can change the frame rate, you may be able to reduce some of the problem. The best thing to do would be to replace your fluorescents with L.E.D tubes - the ones which are not direct replacements for the fluorescents do not use the ballast as their supply, so they don't strobe.

| Reply
Oct 17, 2020 14:40:05   #
PHRubin Loc: Nashville TN USA
 
The second photo shows a patchwork of images stitched together, with a disconnect as circled below. Each of them shows the effects of the 120 Hz flicker rate of florescent lighting, which is much worse than that of incandescent lighting.

The 1st and 3rd seem to be simple flare. Then again, the curve on the upper left of the aberration in the 2nd shot appears to be similar in shape to that in the "flare" in the other two. All seem to be from digital manipulation.



| Reply
Oct 17, 2020 14:47:47   #
Verryl
 
Larry,

Thanks for the insight. I do not recall anything like "Flicker rate" in the camera manual. But the more simple thing to do would be to just use LED lighting, which is more constant.

Thanks again,
Verryl

| Reply
 
 
Oct 17, 2020 14:54:51   #
Verryl
 
The circled part suggesting flare, is a discontinuous change in the blue backdrop from curved inward at the top (near the camera) to straight up and down to allow a simple curve around to the right at the end of the bench.

I have just ordered a pair of medium sized bi-color LED panels from Amazon, Neewer 660 model. With my earlier pair of single globe LED (~4 inch diameter diffuser) lights I should have plenty of light control.

Thanks again,
Verryl

| Reply
Oct 17, 2020 15:09:31   #
PHRubin Loc: Nashville TN USA
 
Verryl wrote:
The circled part suggesting flare, is a discontinuous change in the blue backdrop from curved inward at the top (near the camera) to straight up and down to allow a simple curve around to the right at the end of the bench.

I have just ordered a pair of medium sized bi-color LED panels from Amazon, Neewer 660 model. With my earlier pair of single globe LED (~4 inch diameter diffuser) lights I should have plenty of light control.

Thanks again,
Verryl


Now I see that. I guess that is what I thought was flare in the other shots.

| Reply
If you want to reply, then register here. Registration is free and your account is created instantly, so you can post right away.
UglyHedgehog.com - Forum
Copyright 2011-2021 Ugly Hedgehog, Inc.