Page 3 of 3

PostPosted: Wed May 23, 2012 5:58 pm
by mediavets
jeradg wrote:I also am using manual WB, but have not found a particular one to be pleasing yet.

Read the camera manual and learn how to set a custom WB, rather than using one of the standard options.

PostPosted: Wed May 23, 2012 6:00 pm
by mediavets
jeradg wrote:Here is 1 attempt at an outdoor pano, 6 @ 60 degrees, f/11, 1/125 shutter.

What do you think - are you happer with that one?

PostPosted: Wed May 23, 2012 6:01 pm
by mediavets
jeradg wrote:Added "Force Single Row/Column" to the indoor shot.... Apparently that was the last thing necessary for a good stitching without any work.

Hurrah! Some progress at last. Congrats. :)

PostPosted: Wed May 23, 2012 6:16 pm
by jeradg
It was necessary for me to read the camera manual to even find all the settings. The WB is the latest thing I have been trying to manually set. I haven't yet attempted the fully manual "custom" though. I will tomorrow most likely in some of the rooms where I will be shooting.

The outdoor one stitched very well. I knew the uniqueness of each shot would really help. I did think that APP was better at determining which objects are unique (such as the trash can), but re-examining that pano now, the trash can is really the only unique thing in that location.

From the looks of the rendered locker shot, I should also continue to work on the focus and noise reduction...

Here is a link to the tour I created using fisheye, poor shutter/aperture selection, RAW directly to APP, manual CP additions, vertical line tool, HDR rendering and (poorly done) Photoshop CS5 color/lighting fixes.... http://www.law.utulsa.edu/virtualtour

I don't think it's embarrassingly bad for a first attempt, but all of your help on this thread will definitely help me improve upon it greatly. I think that if you look in the starting location and move horizontally you can see how the tables/couches are stretching on the bottom corners -- if someone can tell me what I am seeing there, please do; the stretching of the couches and tables is what I have been referring to as distortion.

PostPosted: Wed May 23, 2012 6:17 pm
by jeradg
mediavets wrote:
jeradg wrote:Added "Force Single Row/Column" to the indoor shot.... Apparently that was the last thing necessary for a good stitching without any work.

Hurrah! Some progress at last. Congrats. :)

Yes, thank you. :cool:

I will be back tomorrow for more learning. I am supposed to be redoing all of the shots at the end of next week.

PostPosted: Wed May 23, 2012 6:36 pm
by mediavets
jeradg wrote:I will be back tomorrow for more learning. I am supposed to be redoing all of the shots at the end of next week.

How are you deciding the manual exposure settings when shooting a pano scene?

PostPosted: Wed May 23, 2012 7:29 pm
by hankkarl
To add to what mediavets said:
Wood - real wood has unique grain.
When you have a thin veneer of real wood, the grain may look the same in two adjacent peices becasue it is a thin shaving from the same board.
Artificial wood-like formica or a plastic coating on real wood- has many repeated patterns.

The worst I ever saw was on vinyl siding-there were only a few different grain patterns but a lot of siding.

PostPosted: Wed May 23, 2012 7:33 pm
by hankkarl
jeradg wrote:I don't think it's embarrassingly bad for a first attempt, but all of your help on this thread will definitely help me improve upon it greatly. I think that if you look in the starting location and move horizontally you can see how the tables/couches are stretching on the bottom corners -- if someone can tell me what I am seeing there, please do; the stretching of the couches and tables is what I have been referring to as distortion.

As Klaus had said, edges of the image distort at 90 degrees, try zooming in on the initial view.

When you render, try different projections. spherical projection is probably the best for your lens, but YMMV.

PostPosted: Thu May 24, 2012 3:17 pm
by jeradg
mediavets wrote:
jeradg wrote:I will be back tomorrow for more learning. I am supposed to be redoing all of the shots at the end of next week.

How are you deciding the manual exposure settings when shooting a pano scene?

I was going off of an online article that said to set the aperture to as big as it would go, which I now know is not right, and then would just change shutter speed until I got something that looked well lit, and then take a couple shots with a higher and lower shutter speed to create an HDR. With the fully manual fisheye however, I may have had the aperture inadvertently set at some unknown stops without knowing it. I had ISO set for somewhere between 200-400 I believe. I know for a fact that the WB settings were off because I didn't even begin messing with it until last week.

Now I am testing out f/8, f/11 as you and Klaus have suggested. I hope to minimize that bleeding light (is there a more technical term for this effect?) if at all possible. I also have noticed that ISO affects the amount of noise/grain in the image, and I want to bring that down as well.

It appears as though someone else in the building had need of the camera for today. It will be a wasted one for me, as I can't continue to practice without having a camera handy.

PostPosted: Thu May 24, 2012 3:35 pm
by mediavets
jeradg wrote:I also have noticed that ISO affects the amount of noise/grain in the image, and I want to bring that down as well.

Then use the lowest ISO setting available to see what effect that has on noise. Of course that will mean a lower shutter speed. But....

Some recent DSLRs display less noise with high ISO.

Klaus will probably chip in with info about mirror lock up and other things, but you should probably be using a remote control to trigger the shutter to avoid camera shake - do you have remote control?

PostPosted: Thu May 24, 2012 3:48 pm
by klausesser
mediavets wrote:Klaus will probably chip in with info about mirror lock up and other things, but you should probably be using a remote control to trigger the shutter to avoid camera shake - do you have remote control?

Here i am . . :cool::)

Indeed mirror lock is essential with exposure times lower than 1/125 sec. The more hires you shoot the more critical it gets. The longer your lens is the more critical it gets too - tele-lenses need very short exp. times but nevertheless become also sensitive to mirror-shaking even with short times. It´s the same effect as is with binoculars: hard to hold them steady. Even breathing moves it.

The shorter the lens is the less sensitive it becomes. But indoors you most likely need some longer exp.time - better use mirrorlock even with a fisheye.

best, Klaus

PostPosted: Thu May 24, 2012 4:25 pm
by jeradg
mediavets wrote:
jeradg wrote:I also have noticed that ISO affects the amount of noise/grain in the image, and I want to bring that down as well.

Then use the lowest ISO setting available to see what effect that has on noise. Of course that will mean a lower shutter speed. But....

Some recent DSLRs display less noise with high ISO.

Klaus will probably chip in with info about mirror lock up and other things, but you should probably be using a remote control to trigger the shutter to avoid camera shake - do you have remote control?

The reason I even mention it is that I've noticed some (probably unnecessary) noise in my pictures.

No remote control that I know of. The department that we are "borrowing" the camera from for this might have one, but I doubt it. I know that the camera shakes when I push it though. I should probably look into what mirror lock is since this is the first I am hearing of it.

PostPosted: Thu May 24, 2012 4:43 pm
by klausesser
jeradg wrote:I should probably look into what mirror lock is since this is the first I am hearing of it.

A DSLR - which means DigitalSingleLensReflex - has a mirror which "routes" the lightrays coming through the lens to the viewfinder-glass where you can see them as a projection.
That means the lightrays dosn´t reach the sensor. To make them hitting the sensor two actions happen: 1) the mirror claps up and 2) the shutter - which holds any light away from the sensor - opens and the sensor registers the projecting lightrays.
Because of the mirror must have a certain size and a certain weight: when it claps up and covers the viewfinder-glass a shaking occurs because the mass of the moving mirror which hits the viewfinder-housing.
This shaking is very short - but relatively strong: you can feel it when you hold the camera tight. That means the camera´s body shakes a bit for the fraction of a second. Using an expoure-time wich lasts relatively long means blurred images: the sensor moves because the body moves relatively to the object.

Usually at speeds faster than 1/125 you most unlikely will see the effect - but using a tele lens ADDS the "binocular effect" to the mirror shaking. You need to use much shorter exp. times then.

best, Klaus

PostPosted: Thu May 24, 2012 4:58 pm
by jeradg
klausesser wrote:
jeradg wrote:I should probably look into what mirror lock is since this is the first I am hearing of it.

A DSLR - which means DigitalSingleLensReflex - has a mirror which "routes" the lightrays coming through the lens to the viewfinder-glass where you can see them as a projection.
That means the lightrays dosn´t reach the sensor. To make them hitting the sensor two actions happen: 1) the mirror claps up and 2) the shutter - which holds any light away from the sensor - opens and the sensor registers the projecting lightrays.
Because of the mirror must have a certain size and a certain weight: when it claps up and covers the viewfinder-glass a shaking occurs because the mass of the moving mirror which hits the viewfinder-housing.
This shaking is very short - but relatively strong: you can feel it when you hold the camera tight. That means the camera´s body shakes a bit for the fraction of a second. Using an expoure-time wich lasts relatively long means blurred images: the sensor moves because the body moves relatively to the object.

Usually at speeds faster than 1/125 you most unlikely will see the effect - but using a tele lens ADDS the "binocular effect" to the mirror shaking. You need to use much shorter exp. times then.

best, Klaus

Apparently the Nikon D5100 doesn't have a true MLU, and some people have even complained that using it in remote mode, delayed exposure, or live view has mirror movement causing issues (in different ways/times for each mode).

Btw, taking the 6 shots with the fisheye has created a spherical pano that looks much better in the tour, particularly with a ~60 FOV. With the sub-15000 width of the fisheye panos, I will need to continue to clean up the image quality.

Thanks for all the help everyone! :D

PostPosted: Thu May 24, 2012 5:10 pm
by mediavets
jeradg wrote:No remote control that I know of. The department that we are "borrowing" the camera from for this might have one, but I doubt it. I know that the camera shakes when I push it though. I should probably look into what mirror lock is since this is the first I am hearing of it.

The Nikon D5100 supports either a wired remote:

http://www.cameta.com/Nikon-MC-DC2-Wired-Remote-Shutter-Release-Cord-for-D4-D800-D3100-D5000-D5100-D7000-37751.cfm

Or a wireless IR remote:

http://www.cameta.com/Nikon-ML-L3-Wireless-Infrared-Shutter-Release-Remote-Control-12892.cfm

I am sure there are many Chinese clones of both of these available too.

PostPosted: Thu May 24, 2012 5:22 pm
by mediavets
jeradg wrote:Btw, taking the 6 shots with the fisheye has created a spherical pano that looks much better in the tour, particularly with a ~60 FOV. With the sub-15000 width of the fisheye panos, I will need to continue to clean up the image quality.

Thanks for all the help everyone! :D

I looked at your first tour:

http://www.law.utulsa.edu/virtualtour/

You seem to have chosen not to offer zoom in/out controls, is that correct and intentional?

In any case at the size your tour is displayed on the web page the D5100 with the Samyang-clone FE lens should give you a sufficient pano resolution.

What resolution are you getting for your 2:1 aspect ratio spherical stiched pano from APP/APG at 100% render?

I know that with a single row you are not covering a full 360x1980 degrees pano FOV but if you choose to force a full 360x180 in APP/APG you'll find it easier to handle the images in Panotour Pro, just use the Crop tool in the Hotspot editor panel to limit the field of view when the tour is viewed.

If you crop in APP/APG before rendering then you will have probably have to enter the pano image FOV values manually in PTP in order to avoid distortions. These values are to be found in a custom EXIF tag in the rendered pano image and is read automatically bt PTP but this EXIF tag is typically removed by most image processing software so is likely not to be available by the time you import the image into PTP.

PostPosted: Thu May 24, 2012 8:49 pm
by hankkarl
Couple of thoughts:
1. use the lowest natural ISO. Some Canons are better at ISO 100 than ISO 50 (ISO 50 is an "extended" ISO).
2. you can use the selftimer instead of the remote. 2 seconds is enough if you can select the timer time.
3. there is some discussion that MLU is not needed at very long exposures. But I always use MLU on indoor panos, and often use it on outdoor ones.

PostPosted: Tue May 29, 2012 4:26 pm
by jeradg
mediavets wrote:
jeradg wrote:Btw, taking the 6 shots with the fisheye has created a spherical pano that looks much better in the tour, particularly with a ~60 FOV. With the sub-15000 width of the fisheye panos, I will need to continue to clean up the image quality.

Thanks for all the help everyone! :D

I looked at your first tour:

http://www.law.utulsa.edu/virtualtour/

You seem to have chosen not to offer zoom in/out controls, is that correct and intentional?

In any case at the size your tour is displayed on the web page the D5100 with the Samyang-clone FE lens should give you a sufficient pano resolution.

What resolution are you getting for your 2:1 aspect ratio spherical stiched pano from APP/APG at 100% render?

I know that with a single row you are not covering a full 360x1980 degrees pano FOV but if you choose to force a full 360x180 in APP/APG you'll find it easier to handle the images in Panotour Pro, just use the Crop tool in the Hotspot editor panel to limit the field of view when the tour is viewed.

If you crop in APP/APG before rendering then you will have probably have to enter the pano image FOV values manually in PTP in order to avoid distortions. These values are to be found in a custom EXIF tag in the rendered pano image and is read automatically bt PTP but this EXIF tag is typically removed by most image processing software so is likely not to be available by the time you import the image into PTP.

We chose not to display the zoom controls, but after reshooting tomorrow, the new tour will (most likely) have some zoom ability. FOV was locked at 83 (set in PTP).

The shot of the lockers and the outdoor shot gave me ~11,600 x 4,835 (+/-10). I was previously rendering at 10,000 x 4,200 by keeping h/w ratio and setting width to 10,000.

As for the FOV setting being removed by most image processing: would checking the FOV setting in PTP before editing, and then manually adding that value back in after importing the edited image be a simple work-around? I never crop in APP, but I didn't have the force 360x180 setting checked before.

PostPosted: Tue May 29, 2012 4:34 pm
by jeradg
hankkarl wrote:Couple of thoughts:
1. use the lowest natural ISO. Some Canons are better at ISO 100 than ISO 50 (ISO 50 is an "extended" ISO).
2. you can use the selftimer instead of the remote. 2 seconds is enough if you can select the timer time.
3. there is some discussion that MLU is not needed at very long exposures. But I always use MLU on indoor panos, and often use it on outdoor ones.

Set to 125 last week it appeared to come out alright, but I will be experimenting with the 100 ISO today. I will also try to self-timer, but it appears as though it is a 10 second timer only.

PostPosted: Tue May 29, 2012 5:55 pm
by jeradg
klausesser wrote:
jeradg wrote:Edit: how can I eliminate the effect of the light bleeding onto the ceiling tiles? I have been doing it in photoshop after stitching.

You can´t. It´s the lens. Given you cleaned your lens properely the bleeding shows that the coating isn´t good.
This results in such effects - the border between dark and very bright isn´t as sharp accentuated as it should be
and a halo occurs.

You can minimize it a bit by using a small aperture - f:11/16/22 test them.

best, Klaus

P.S.: how did you process the HDR?

Now I have all my camera settings manual, and preset my WB via measure function.
I cleaned my FE lens, and then I tried some f:11/16/22 just now... It looks like the lights are still bleeding :mad: :(

I was really hoping to use the fisheye, but with bleeding lights I might have to resort to using the standard lens.

PostPosted: Tue May 29, 2012 6:28 pm
by jeradg
I took the shots in RAW + JPEG (fine)...

I definitely see the advantage of working with RAW format in Photoshop CS5 rather than allowing the camera to handle it. The "Clarity" slider directly deals with the bleeding lights (although it is imperfect because it just darkens the edges in a reverse sort of way).

So after shooting RAW and making adjustments, I am guessing that I should have some sort of preset to equalize the images for a given pano and apply the settings to each RAW before putting it in to APP. Please let me know if this is correct.

PostPosted: Tue Aug 14, 2012 10:21 am
by Ajithkumar Yadhavan
Using Autopano Pro & Autopano Giga:Thank You! It is a good idea to support our efforts in bringing you the best stitching software available on the market today. To do so, just go to ouronline store:The head uses a simple ascii serial protocol; the serial line uses standard settings: 9600 8N1 (9600 bauds, 8 bits data, no parity, 1 stop bit). The head waits for incoming commands, and send response to them. The 2 axis are identical, and uses the same commands. They both wait for incoming commands; depending of the first value (see below), the command will be sent to one or the other axis.....................