Panorama Quality

[Newt59]

i am loving the new pano features for the P4P in the latest firmware from December 2017.

But one thing I notice is they are very washed out overall. I am running Auto Focus and Auto Exposure With the camera spinning towards and away from the sun and various angles, I see no other way to compensate. And since it is a pano, when viewing it in a “flat mode”, I’m not sure how much post I can do.

I am going to work some various settings and such, but curious as to how everyone else is doing with these new features.

BTW, used the Litchi app a few times and there is no comparison. DJI is the way to go.

Here is one of my pano pics Kuula

 

[ryantrax]

I have to agree that the capture process in DJI Go 4 seems much better and faster than Litchi, but there aren't as many customization options. Unfortunately once you start the pano in DJI Go there is no way to modify any settings per individual photo. When shooting into the sun or in sky areas with not much definition the camera may struggle a bit to focus, panos are best shot with the sun well above the horizon to avoid this.
You can manually take the photos, I usually do 3 rows of 16 and 2 looking straight down.

 

[With The Birds]

Dial in your exposure settings with the camera aimed at the brightest area of the proposed area to be imaged in manual mode then take it from there.

 

[eMario]

Hi
My name is Mario. I'm from Poland. I'm amateur pilot and hobbyist.
I have Phantom 3 Standard.
I personally do panoramas like this:
-three rows of photos (360)
-photo tabs 50%
-time such that the sun would be as high as possible

Look at my panorams (my small city - seaside town)

krpano - panorama4

 

[Helihover]

Dial in your exposure settings with the camera aimed at the brightest area of the proposed area to be imaged in manual mode then take it from there.

This what I’m going back and forth with in my head. I can’t try the Go app because they don’t offer this feature. I have been using Litchi.

Litchi shoots in manual. Go shoots auto.

I feel shooting in auto for a 360 is or could be better. In my experience shooting in manual, I get a line at the start/stop. This is because the EV is off where the stitching took place. If the camera was in auto, the EV would match on all the pics and you shouldn’t get the line.

I have since starting editing most of this line out in post, but if you watch this video you can see a line in almost every pano I shot. The sun stays pretty low here this time of year so that doesn’t help either.

 

[Helihover]

i am loving the new pano features for the P4P in the latest firmware from December 2017.

But one thing I notice is they are very washed out overall. I am running Auto Focus and Auto Exposure With the camera spinning towards and away from the sun and various angles, I see no other way to compensate. And since it is a pano, when viewing it in a “flat mode”, I’m not sure how much post I can do.

I am going to work some various settings and such, but curious as to how everyone else is doing with these new features.

BTW, used the Litchi app a few times and there is no comparison. DJI is the way to go.

Here is one of my pano pics Kuula

I edit all my Panos in Light Room.

 

[With The Birds]

This what I’m going back and forth with in my head. I can’t try the Go app because they don’t offer this feature. I have been using Litchi.

Litchi shoots in manual. Go shoots auto.

I feel shooting in auto for a 360 is or could be better. In my experience shooting in manual, I get a line at the start/stop. This is because the EV is off where the stitching took place. If the camera was in auto, the EV would match on all the pics and you shouldn’t get the line.

I have since starting editing most of this line out in post, but if you watch this video you can see a line in almost every pano I shot. The sun stays pretty low here this time of year so that doesn’t help either.

You have more chance of getting an obvious stitch line shooting in auto. This is simply because it is auto mode which might provide for the subject images to have different exposure values. If all shot on the same EV settings in manual this can’t occur.

 

[Helihover]

You have more chance of getting an obvious stitch line shooting in auto. This is simply because it is auto mode which might provide for the subject images to have different exposure values. If all shot on the same EV settings in manual this can’t occur.

When you shoot a pano, the EV changes every time the drone moves when shooting in manual. On some of the panos I’ve shot, I’ve seen as much as a 3 point difference in EV as the drone turns.

If you shoot in auto, the EV stays the same through out the process, which is 0. Shooting in manual, only your first picture is at an EV of 0. The rest of the pics will be at whatever the light dictates because your exposure is locked. If you set your first shot facing into the sun, your going to get some very under exposed pictures during the pano.

Do you have any panos shot in manual you could show?

Anyone else have panos shot in auto?

I really wish the Go app had the pano option for the P4 so I could just find out for myself. From what I’ve seen though, it looks as if the panos on Go are better. Probably because they are shot in auto

I’ve shot over 200 pano missions for drone Base. They do the stitching and recommend you shoot in manual. Out of the 200 plus panos I’ve shot for them, not one was in manual, all were done in auto.

 

[With The Birds]

When you shoot a pano, the EV changes every time the drone moves when shooting in manual. On some of the panos I’ve shot, I’ve seen as much as a 3 point difference in EV as the drone turns.

If you shoot in auto, the EV stays the same through out the process, which is 0. Shooting in manual, only your first picture is at an EV of 0. The rest of the pics will be at whatever the light dictates because your exposure is locked. If you set your first shot facing into the sun, your going to get some very under exposed pictures during the pano.

Do you have any panos shot in manual you could show?

Anyone else have panos shot in auto?

I really wish the Go app had the pano option for the P4 so I could just find out for myself. From what I’ve seen though, it looks as if the panos on Go are better. Probably because they are shot in auto

I’ve shot over 200 pano missions for drone Base. They do the stitching and recommend you shoot in manual. Out of the 200 plus panos I’ve shot for them, not one was in manual, all were done in auto.

While you will almost always see a change in the intensity of light entering the lens with re-aiming the camera Exposure Value (EV) does not vary. EV is simply a reference providing for a quick means to determine how camera settings (aperture, shutter and ISO) interact to alter exposure. For example if you open the aperture one fstop you would need to reduce shutter or ISO by one stop to maintain the same EV for a given scene brightness.

When shooting in auto EV does not stay the same, it will vary with respect to scene brigthess as the camera automatically adjusts settings to arrive at the same average exposure. Exposure is simply a function of EV and scene Illuminance.

The likely reason dronebase suggests the pano’s are shot in manual is that this method has demonstrated to give the most natural results.

My approach (and that I understand to be common practice) is to meter for the brightest area of the scene and take all images at the same EV (manual mode). What generally works best is to expose to the right of the histogram to ensure highlights are preserved and dynamic range is maximised. This will give the most natural presentation to the stitched image.

By shooting all frames in auto you are likely to get a less natural looking result as the lighter areas will appear darker and the darker lighter than that seen by the eye in the actual scene.

 

[Helihover]

Maybe I should make video to explain what I’m say because your defiantly not understanding what I’m typing.

The EV (exposure value) is displayed on the screen while you fly. This is adjustable via SS, ISO, and on a P4P, aperture. This value, EV, tells you how under or over exposed your picture will be. 0 being neutral, negative numbers under exposed (dark) and positive number being over exposed (too bright). Agree?

When you start a pano, in Litchi, on a P4, your first picture sets the EV at 0 via SS. If you have sun in the first picture and your SS is at say 1/1600, then that SS will remain the same throughout the pano. When the drone moves to a different shot with less light, the EV drops.

If the drone were in auto it would be constantly adjusting the SS each shot to keep the EV at 0.

At this point I’m not saying one is better than the other. I’m, thinking, that shooting auto might be better because the EV will remain the same throughout the shoot.

 

[With The Birds]

Maybe I should make video to explain what I’m say because your defiantly not understanding what I’m typing.

The EV (exposure value) is displayed on the screen while you fly. This is adjustable via SS, ISO, and on a P4P, aperture. This value, EV, tells you how under or over exposed your picture will be. 0 being neutral, negative numbers under exposed (dark) and positive number being over exposed (too bright). Agree?

When you start a pano, in Litchi, on a P4, your first picture sets the EV at 0 via SS. If you have sun in the first picture and your SS is at say 1/1600, then that SS will remain the same throughout the pano. When the drone moves to a different shot with less light, the EV drops.

If the drone were in auto it would be constantly adjusting the SS each shot to keep the EV at 0.

At this point I’m not saying one is better than the other. I’m, thinking, that shooting auto might be better because the EV will remain the same throughout the shoot.

I don’t think I have misunderstood. Imagine You have two cameras on your phantom. One has a lens with a focal length that gives it the same field of view as the resultant image from a panorama stitched from three rows of three images from the second camera. Both cameras are set to auto exposure metering. The imaged area has the sun in the frame (not centred). You can see where I’m going here. Camera one will likely produce an image with blown highlights. The stitched image will preserve more highlight detail however the brighter and darker areas will be pulled closer to the average scene brightness. The stitched image will lack contrast. To get the same effect in the single shot you would need to underexpose the brighter areas and overexpose the dark areas. This isn’t possible of course in a single shot but it is precisiely what you are doing with camera two.

 

[Helihover]

I don’t think I have misunderstood. Imagine You have two cameras on your phantom. One has a lens with a focal length that gives it the same field of view as the resultant image from a panorama stitched from three rows of three images from the second camera. Both cameras are set to auto exposure metering. The imaged area has the sun in the frame (not centred). You can see where I’m going here. Camera one will likely produce an image with blown highlights. The stitched image will preserve more highlight detail however the brighter and darker areas will be pulled closer to the average scene brightness. The stitched image will lack contrast. To get the same effect in the single shot you would need to underexpose the brighter areas and overexpose the dark areas. This isn’t possible of course in a single shot but it is precisiely what you are doing with camera two.

I’m not sure what your trying to speculate here.

You stated that the EV does not change during a pano shot in manual. This is not the case. In fact the EV does change with every shot. When you shoot in auto, the EV stays the same because the SS is constantly being adjusted to keep the EV at 0. I’m very surprised with the amount of knowledge you have that you can’t see this. You have shot panos right? Do you use Litchi or Go?

Watch this video. This shows a gent shooting a pano, in auto, with Go. Watch how the SS changes with every shot and how the EV stays at 0.

If you like, I can make a video showing how a pano works in manual and how the EV changes with every shot.

 

[With The Birds]

I’m not sure what your trying to speculate here.

You stated that the EV does not change during a pano shot in manual. This is not the case. In fact the EV does change with every shot. When you shoot in auto, the EV stays the same because the SS is constantly being adjusted to keep the EV at 0. I’m very surprised with the amount of knowledge you have that you can’t see this. You have shot panos right? Do you use Litchi or Go?

Watch this video. This shows a gent shooting a pano, in auto, with Go. Watch how the SS changes with every shot and how the EV stays at 0.

If you like, I can make a video showing how a pano works in manual and how the EV changes with every shot.

I thought I had mentioned earlier, I shoot panos in manual. I haven’t updated GO and haven’t used the auto pano feature in litchi either.

The EV does not change when you shoot in manual. EV-3, for example, simply means the optimum exposure as determined by the camera light meter is three stops more than the current AV, TV and ISO settings will produce. When your shoot with a DSLR and switch from weighted average to spot metering you can see EV +/- change, often significantly, for different areas of the frame. EV0 is the average of all these readings (it actually won’t be as every SLR I own and have used are weighted towards a centre average on the assumption the most important area of the frame is towards the centre). Think of the 9 images shot on auto in the panorama as being the spot meter readings.

 

[bsartist]

I thought I had mentioned earlier, I shoot panos in manual. I haven’t updated GO and haven’t used the auto pano feature in litchi either.

The EV does not change when you shoot in manual. EV-3, for example, simply means the optimum exposure as determined by the camera light meter is three stops more than the current AV, TV and ISO settings will produce. When your shoot with a DSLR and switch from weighted average to spot metering you can see EV +/- change, often significantly, for different areas of the frame. EV0 is the average of all these readings (it actually won’t be as every SLR I own and have used are weighted towards a centre average on the assumption the most important area of the frame is towards the centre). Think of the 9 images shot on auto in the panorama as being the spot meter readings.

Jumping in on this thread. As far as I know, if you are in manual and leave your settings locked and take multiple pics across the landscape to make a pano, then the EV will change if there is more or less light in the various pictures you take. Let’s say you take your exposure reading with the camera framed where The Sun is not in the frame and you lock your settings, then when you pan over and take a shot where The Sun is in the frame then the EV measured will be higher. More light coming in but you haven’t changed any settings to compensate. Auto will remeter for each shot and will by definition try to keep the EV at zero.

When I do panos with my DSLR I’ve always used manual and exposed for the part of the scene that is most important and don’t change settings. Why do I do this? As alluded to earlier, if you shoot in auto then the settings will change on you to compensate for brighter parts of the sky. This may sound like a good thing. But sometimes it’s not when pictures taken with different settings are stitched. When one picture is forced to expose darker than the one next to it, you can end up with unnatural transitions across the sky.

If you had a camera that could take one wide pano image with one press of the shutter you only get one meter reading - one set of settings. Shooting many based one one exposure measurement and set of settings, in essence, accomplishes the same thing.

I suppose in theory one could say, “but exposing for each shot let’s you get more dynamic range by dropping highlights and lifting shadows in the different frames.” But AFAIU this has been discouraged because of trouble caused to the stitching process. But maybe stitching algorithms are getting better to allow it.

 

[bsartist]

Some good reading here:

Which parameters to lock between images for a panorama?

By the way, using a polarizer for wide panos with sky can mess you up. Some sky will be dark while some will be light.

 

[Helihover]

Jumping in on this thread. As far as I know, if you are in manual and leave your settings locked and take multiple pics across the landscape to make a pano, then the EV will change if there is more or less light in the various pictures you take. Let’s say you take your exposure reading with the camera framed where The Sun is not in the frame and you lock your settings, then when you pan over and take a shot where The Sun is in the frame then the EV measured will be higher. More light coming in but you haven’t changed any settings to compensate. Auto will remeter for each shot and will by definition try to keep the EV at zero.

When I do panos with my DSLR I’ve always used manual and exposed for the part of the scene that is most important and don’t change settings. Why do I do this? As alluded to earlier, if you shoot in auto then the settings will change on you to compensate for brighter parts of the sky. This may sound like a good thing. But sometimes it’s not when pictures taken with different settings are stitched. When one picture is forced to expose darker than the one next to it, you can end up with unnatural transitions across the sky.

If you had a camera that could take one wide pano image with one press of the shutter you only get one meter reading - one set of settings. Shooting many based one one exposure measurement and set of settings, in essence, accomplishes the same thing.

I suppose in theory one could say, “but exposing for each shot let’s you get more dynamic range by dropping highlights and lifting shadows in the different frames.” But AFAIU this has been discouraged because of trouble caused to the stitching process. But maybe stitching algorithms are getting better to allow it.

I’m one to think exposing each shot would make for a better pano, but as stated earlier, the GO app doesn’t support Panos on the P4. Hopefully it comes soon and then I can do my own back to test. I’m not saying one is better than the other, just thinking that exposing each shot would make for a better stitch. Wouldn’t, in theory, two pictures that had the same EV be easier to stitch than two pictures that had an EV of 0 and one that was -2.3? The pic with an EV of -2.3 would almost be black.... How is that supposed to stitch better than something that matches?

Anyway, anyone who’s used a camera for more than 5 minutes knows that the EV changes every time you move the lens. Some just love to argue about anything. Even when you show them first hand with proof. I find it funny, but hey, I guess some have fun in different ways than others, or?

 

[bsartist]

I’m one to think exposing each shot would make for a better pano, but as stated earlier, the GO app doesn’t support Panos on the P4. Hopefully it comes soon and then I can do my own back to test. I’m not saying one is better than the other, just thinking that exposing each shot would make for a better stitch. Wouldn’t, in theory, two pictures that had the same EV be easier to stitch than two pictures that had an EV of 0 and one that was -2.3? The pic with an EV of -2.3 would almost be black.... How is that supposed to stitch better than something that matches?

Anyway, anyone who’s used a camera for more than 5 minutes knows that the EV changes every time you move the lens. Some just love to argue about anything. Even when you show them first hand with proof. I find it funny, but hey, I guess some have fun in different ways than others, or?

I think for sure there are situations where two photos with the same EV may not stitch together well. As I pointed out above, let’s say one frame contains blue sky, the one next to it contains blue sky and the sun or other bright subject. The latter will drop the exposure to keep the same overall EV but the blue sky will be much darker than in the other picture because the sun is causing the exposure to be pulled down to keep an overall EV of zero. So where they meet will be a problem.

 

[bsartist]

Also, we can’t always blindly follow the meter anyway. It wants to expose for 18% grey. What if a picture was predominantly white clouds? The meter doesn’t know that. So it’ll underexpose to try and get 18% grey. That’s why we add exposure compensation. Like shooting in snow. This is probably a bit of an oversimplification with more modern matrix metering modes etc. but you get the idea.

One other thing, I think the exposure compensation in the DJI Go app lets you define your compensation in + or - EV. This makes EV “look like” a setting you can change. That sometimes confuses people and terminology.

 

[Helihover]

Yep, we can sit here and jibber jabish all day long

Here’s what I get when I shoot a pano in manual.

You can see the stitch line in every one of these. You can see the line because the exposure is off where they come together. This is pretty obvious in the second pano. These panos are also edited with the shadows and highlights at their max. They also all have a radius gradiant filter added too, bringing down the exposure, in the sky, over 2 points! If not, I get a blown out sky and ground that’s too dark to see anything. If these were shot in auto, the sky wouldn’t be so bright and the ground wouldn’t be so dark.

I wish someone could post some 360 panos shot in auto.....

 

[bsartist]

Yep, we can sit here and jibber jabish all day long

Here’s what I get when I shoot a pano in manual.

View attachment 94265

View attachment 94266 View attachment 94267

You can see the stitch line in every one of these. You can see the line because the exposure is off where they come together. This is pretty obvious in the second pano. These panos are also edited with the shadows and highlights at their max. They also all have a radius gradiant filter added too, bringing down the exposure, in the sky, over 2 points! If not, I get a blown out sky and ground that’s too dark to see anything. If these were shot in auto, the sky wouldn’t be so bright and the ground wouldn’t be so dark.

I wish someone could post some 360 panos shot in auto.....

I wasn’t thinking 360 tiny planet panos. Not sure I would have said anything different but I hadn’t really thought in those terms specifically. I’m mainly going from my experience with landscape panos from a DSLR. Yes we can jibber jabber but I’m not convinced that talking about some underlying theory is totally without merit. But, yes, practical examples to advance the discussion. I do think, at the end of the day, it might depend on the scene and even the stitching software used. I know I’ve had trouble with panos shot in auto. And you are having trouble with some shot in manual. So I think a valuable lesson for me is that there is no hard and fast rule and I’ll either try to adapt based on the theory and see what I can learn. Or try them both ways and see what I can learn. And as I’ve alluded to, some of my theory may come from vestigial beliefs based on older software and/or metering systems.