Creating hdri evo maps

CASINC

I'm attempting to create my own hdri environmental maps for Daz using a theta Z1 360 camera. Long story short, they just aren't coming out normal hdri quality like all the other hdr's I use. Workflow: -Have been taking 8 different exposure levels in raw format -Combine them in lightroom classic as a hdr. -Stitch them with the theta stitcher. - Open in photoshop, edit out stand and what not. - As 32 bit, save as Radiance. At this point I apply it to my scene in daz and it looks like an 8 bit jpg file with no real lighting info. Any suggestions or help would be great. Thanks!

markusmatern

How many F stops are your exposures apart?

CASINC

markusmatern said:

How many F stops are your exposures apart?

It's set at 2.1 not sure why, think it was a setting learned on a random tutorial.

markusmatern

I meant what is the exposure difference expressed in F stops between each of the bracketed 8 shots you took?

For some more information you can take a look at the FAQ from HDRI Haven

How many stops/EVs do you capture for each HDRI?

How do you measure the dynamic range (EVs)?

How do you reach 24+ EVs?

CASINC

markusmatern said:

I meant what is the exposure difference expressed in F stops between each of the bracketed 8 shots you took?

For some more information you can take a look at the FAQ from HDRI Haven

How many stops/EVs do you capture for each HDRI?

How do you measure the dynamic range (EVs)?

How do you reach 24+ EVs?

I have 8 shots at 1/2, 1/8, 1/30, 1/125, 1/500, 1/2000, 1/8000, 1/25000. But as I look into it more, turns out the theta Z1 is only has a 7k resolution...I may have wasted $1200.

Fishtales

CASINC said:

markusmatern said:

I meant what is the exposure difference expressed in F stops between each of the bracketed 8 shots you took?

For some more information you can take a look at the FAQ from HDRI Haven

How many stops/EVs do you capture for each HDRI?

How do you measure the dynamic range (EVs)?

How do you reach 24+ EVs?

I have 8 shots at 1/2, 1/8, 1/30, 1/125, 1/500, 1/2000, 1/8000, 1/25000. But as I look into it more, turns out the theta Z1 is only has a 7k resolution...I may have wasted $1200.

That is shutter speeds not exposure. Looking at the specs it does exposure compensation from -2.0 to +2.0 EV in 1/3 EV steps. And "A smartphone is required to change modes or configure manual settings"

CASINC

Fishtales said:

CASINC said:

markusmatern said:

I meant what is the exposure difference expressed in F stops between each of the bracketed 8 shots you took?

For some more information you can take a look at the FAQ from HDRI Haven

How many stops/EVs do you capture for each HDRI?

How do you measure the dynamic range (EVs)?

How do you reach 24+ EVs?

I have 8 shots at 1/2, 1/8, 1/30, 1/125, 1/500, 1/2000, 1/8000, 1/25000. But as I look into it more, turns out the theta Z1 is only has a 7k resolution...I may have wasted $1200.

That is shutter speeds not exposure. Looking at the specs it does exposure compensation from -2.0 to +2.0 EV in 1/3 EV steps. And "A smartphone is required to change modes or configure manual settings"

Not sure, the only settings it shows and I can change in the app are aperture (f2.1), shutter speed, iso (200), WB (4800).

Fishtales

CASINC said:

Not sure, the only settings it shows and I can change in the app are aperture (f2.1), shutter speed, iso (200), WB (4800).

 The Aperture is the exposure setting and that is what you need to change. You should be able to set it manually in the app. by -2.0 to +2.0 in increaments of 1/3. The 2.1 setting will probably only allow you to go up, then set it to the next setting which is 3.5 and go down a 1/3 at a time to -2.0 then up from 3.5 to +2.0, and then the next at 4.6 and do the same. By my calculations that should be 33 images ( probably wrong but you get the idea) :)

CASINC

Fishtales said:

CASINC said:

Not sure, the only settings it shows and I can change in the app are aperture (f2.1), shutter speed, iso (200), WB (4800).

 The Aperture is the exposure setting and that is what you need to change. You should be able to set it manually in the app. by -2.0 to +2.0 in increaments of 1/3. The 2.1 setting will probably only allow you to go up, then set it to the next setting which is 3.5 and go down a 1/3 at a time to -2.0 then up from 3.5 to +2.0, and then the next at 4.6 and do the same. By my calculations that should be 33 images ( probably wrong but you get the idea) :)

Ok, the only options are 2.1, 3.5, and 5.6. So basically are you saying, for instance, if I'm taking 8 pics at different shutter speeds I also need to take those pics at the different aperture settings too (giving me 24 pics total)?

markusmatern

In panoramic photography you usually don't change the aperture between bracketed shots, because this also changes the depth of field for the shots. As a first experiment you could set the aperture and ISO both to a fixed value and only change the time. There is a special shooting mode in the Theta App on your phone. Also for a first experiment don't take too many images, but further apart. Try it with -3 EV, 0 EV, +3 EV by adjusting only the time https://photographylife.com/exposure-value

CASINC

I guess I should better specify. The problem isn't really figuring out how to get a high dynamic range, but rather that it's just producing a low quality photo. When I export ("save as") the final pic out of photoshop it's only producing a 80mb file, when most 16k hdri's are around 300-600mb. I've used Dreamlight's tutorial, here at daz, on the subject. The 360 camera used in that tutorial is of a lower quilty then the theta Z1, and yet they get high quality hdri's. So I'm stuck not knowing what the problem is, camera, settings, processing, a combo of them all...

markusmatern

The size you described seems OK to me. The 8K versions of HDRIs from HDRI Haven are about 90MB Here is an example https://hdrihaven.com/hdri/?h=sunflowers Maybe you could pick an 8K image from HDRI haven with similar lighting to help as a sort of reference what it looks like in Photoshop?

CASINC

markusmatern said:

The size you described seems OK to me. The 8K versions of HDRIs from HDRI Haven are about 90MB Here is an example https://hdrihaven.com/hdri/?h=sunflowers Maybe you could pick an 8K image from HDRI haven with similar lighting to help as a sort of reference what it looks like in Photoshop?

I have plenty of their hdr's, no 8k though. I'll have to compare like you said. But ya, basically they have a 8ish K quality. Still, going off the dreamlight tutorial it seems like it should be producing at least 16k quality.

Fishtales

Lets step back to the beginning then.

The combining them in Lightroom to and HDR. Is that an HDR photo or an HDRI lighting file? The two are different.

https://vrender.com/what-is-hdri/

https://blog.hdrihaven.com/how-to-create-high-quality-hdri/

CASINC

Fishtales said:

Lets step back to the beginning then.

The combining them in Lightroom to and HDR. Is that an HDR photo or an HDRI lighting file? The two are different.

https://vrender.com/what-is-hdri/

https://blog.hdrihaven.com/how-to-create-high-quality-hdri/

The pics come off the camera as raw/dng files, usually around 40mb each, then combine them to hdr in lightroom. The final product comes out at about 80mb making a 6.5k map, but it looks about like 2k in daz with zero dynamic range.

Fishtales

CASINC said:

Fishtales said:

Lets step back to the beginning then.

The combining them in Lightroom to and HDR. Is that an HDR photo or an HDRI lighting file? The two are different.

https://vrender.com/what-is-hdri/

https://blog.hdrihaven.com/how-to-create-high-quality-hdri/

The pics come off the camera as raw/dng files, usually around 40mb each, then combine them to hdr in lightroom. The final product comes out at about 80mb making a 6.5k map, but it looks about like 2k in daz with zero dynamic range.

 

Did you read the HDR Haven article. In it he says that Lightroom changes the images as it combines them so he stopped using it. I have used the .dng converter once on one of my .raw files  and it completely ruind it so I wouldn't recommend using .dng files :)

CASINC

Fishtales said:

CASINC said:

Fishtales said:

Lets step back to the beginning then.

The combining them in Lightroom to and HDR. Is that an HDR photo or an HDRI lighting file? The two are different.

https://vrender.com/what-is-hdri/

https://blog.hdrihaven.com/how-to-create-high-quality-hdri/

The pics come off the camera as raw/dng files, usually around 40mb each, then combine them to hdr in lightroom. The final product comes out at about 80mb making a 6.5k map, but it looks about like 2k in daz with zero dynamic range.

 

Did you read the HDR Haven article. In it he says that Lightroom changes the images as it combines them so he stopped using it. I have used the .dng converter once on one of my .raw files  and it completely ruind it so I wouldn't recommend using .dng files :)

Actually just read it. There is one auto tone mapping setting when combining to hdri in lightroom classic, if not selected it does produce a much more natural pic, and a noticeably better end product. I just tested it. But in the end I'm afraid the camera just won't produce a high enough quality of at least 16k.

Fishtales

Greg Zal uses a Canon 600 D which is 18 MP, your Theta Z1 is 23 MP so the images from your camera are actually bigger than his  :)

CASINC

Fishtales said:

Greg Zal uses a Canon 600 D which is 18 MP, your Theta Z1 is 23 MP so the images from your camera are actually bigger than his  :)

I might sound ignorant but I'm not sure what that number matters given that his still can produce at least 16k resolution hdri's, while mine only does 7k.

Fishtales

CASINC said:

Fishtales said:

Greg Zal uses a Canon 600 D which is 18 MP, your Theta Z1 is 23 MP so the images from your camera are actually bigger than his  :)

I might sound ignorant but I'm not sure what that number matters given that his still can produce at least 16k resolution hdri's, while mine only does 7k.

 

I don't know if this will explain his method as he goes through the camera and set up.

https://blog.hdrihaven.com/camera-gear-for-hdris/

markusmatern

I think the problem may be what to expect from the resolutions of the two cameras.

From the Thata homepage: Records natural 360° images using approx. 23MP(6720 x 3360, 7K) still image shooting and highly accurate image stitching.

The sensor size is a 1.0-inch back illuminated CMOS image sensor.

So the above resolution of the final stitched equirectangular image is roughly 23MP and is based on this sensor size of 1".

 

HDRI Haven on the other hand uses an APS-C sensor based DSLR camera with a rectilinear lens (maybe a 10.5 mm) that has a vertical field of view of 74°. This means to shoot a full sphere a lot of images will be shot and then stitched. (There are calulators available online to do this http://www.hdrlabs.com/tools/panocalc.html)

From the blog:

As an example, my camera has a vertical (shortest side) resolution of 3456, and my lens has a vertical field of view of 73.4°. Using the formula above, my maximum HDRI resolution is:

3456 / (73.4 / 360) = 16950

Thus the maximum HDRI resolution for me is about 16k.

This basically means two things: The sensor is larger than on the Theta Z1. And the lens

This means the final equirectangular image has a maximum resolution of 143MP (16950 x 8475, 16K)

This is a much larger resolution!

Also on HDRI Haven in the FAQ you find:

Why don't you use a 360 camera?

There are two main reasons: Resolution, and highlight clipping...

360 cameras generally have a very low resolution, simply due to the nature of their size and inherent quality of lenses that size. The highest resolution dual camera 360 cam at the time of writing this (October 2019) can produce a 7k panorama, which is a quarter of the total pixels needed at a bare minimum for HDRIs in my opinion (which is 14k). There is one 360 cam I know of that uses a few dozen wide angle lenses and can produce 16k HDRIs, but it's sold as a service charging you per panorama and doesn't seem to give you much control over the output, and still has the clipping issue...

Preventing highlight clipping is absolutely paramount when making high quality HDRIs that accurately capture light from the real world. 360 cameras use multiple fisheye lenses that have a bulbus front element, which makes it impossible to fit an ND filter on the lenses, which makes it impossible to avoid clipping.

So, 360 cams might be a useful tool for your own personal use, or for TDs on a film set that need to be fast and are willing to sacrifice on quality, but as a tool for producing the highest quality HDRIs for public consumption, they're not a good option at all.

 

 

CASINC

markusmatern said:

I think the problem may be what to expect from the resolutions of the two cameras.

From the Thata homepage: Records natural 360° images using approx. 23MP(6720 x 3360, 7K) still image shooting and highly accurate image stitching.

The sensor size is a 1.0-inch back illuminated CMOS image sensor.

So the above resolution of the final stitched equirectangular image is roughly 23MP and is based on this sensor size of 1".

 

HDRI Haven on the other hand uses an APS-C sensor based DSLR camera with a rectilinear lens (maybe a 10.5 mm) that has a vertical field of view of 74°. This means to shoot a full sphere a lot of images will be shot and then stitched. (There are calulators available online to do this http://www.hdrlabs.com/tools/panocalc.html)

From the blog:

As an example, my camera has a vertical (shortest side) resolution of 3456, and my lens has a vertical field of view of 73.4°. Using the formula above, my maximum HDRI resolution is:

3456 / (73.4 / 360) = 16950

Thus the maximum HDRI resolution for me is about 16k.

This basically means two things: The sensor is larger than on the Theta Z1. And the lens

This means the final equirectangular image has a maximum resolution of 143MP (16950 x 8475, 16K)

This is a much larger resolution!

Also on HDRI Haven in the FAQ you find:

Why don't you use a 360 camera?

There are two main reasons: Resolution, and highlight clipping...

360 cameras generally have a very low resolution, simply due to the nature of their size and inherent quality of lenses that size. The highest resolution dual camera 360 cam at the time of writing this (October 2019) can produce a 7k panorama, which is a quarter of the total pixels needed at a bare minimum for HDRIs in my opinion (which is 14k). There is one 360 cam I know of that uses a few dozen wide angle lenses and can produce 16k HDRIs, but it's sold as a service charging you per panorama and doesn't seem to give you much control over the output, and still has the clipping issue...

Preventing highlight clipping is absolutely paramount when making high quality HDRIs that accurately capture light from the real world. 360 cameras use multiple fisheye lenses that have a bulbus front element, which makes it impossible to fit an ND filter on the lenses, which makes it impossible to avoid clipping.

So, 360 cams might be a useful tool for your own personal use, or for TDs on a film set that need to be fast and are willing to sacrifice on quality, but as a tool for producing the highest quality HDRIs for public consumption, they're not a good option at all.

 

 

Thanks for the effort here. I think I'm back to my second conclusion, camera overall just isn't good enough. Ya there is one other 360 cam that shoots 12 or 16k at 108mb. But ya, it's only producing visually a 2-4k quilty hdri. Guess a should ask the forum at hdrihaven for a setup recommendation. Thanks anyways, think this thread is disappointingly resolved. Lol