Loading...
> >

Homemade HDRI panoramas

northallertonnorthallerton Posts: 15
in Technical Help (nuts n bolts)

So I use my camera with the tripod and pano head, record multiple exposures in RAW, stitch in PTGUI Pro and save as a .hdr file. Then I load this into DAZ but the file still can't get the nice dark shadows, especially when compared to the HDRIs on sale such as Cake One or Mec4d. 

Yes I can edit the .hdr pano in Photoshop and bring up the exposure of just the sun area but what am I doing wrong initially? As I have good bracketed exposures, is the problem with the way I'm using PTGUI Pro - any experts out there who can shed some 'light' on the subject?

Comments

  • algovincianalgovincian Posts: 2,636

    What is that dynamic range of the exposures you are taking?

    - Greg

  • northallertonnorthallerton Posts: 15

    I tried 6ev and then 8ev range, ie. 4 above normal exposure and 4 below using shutter speeds. It worked best when I opened the hdr in photoshop and overexposed the sun area by a factor of 4. Just wondering how the vendors were doing it to get such good results. 

  • algovincianalgovincian Posts: 2,636

    There’s your answer. It’s not your processing of bracketed images that is stopping you from getting nice shadows from the sun. You need a much higher dynamic range than +/- 4 ev.

    For example, check out Greg Zall’s HDRI (https://hdrihaven.com/hdris/) - he mentions up to 26 EVs.

    Hope this helps.

    - Greg

     

  • northallertonnorthallerton Posts: 15

    That's where I was getting confused - I thought I was doing well to range from 1/4000th sec down to 1/15th sec, obviously I need to add higher and lower speeds. I wonder what the effect would be of just taking 3 exposures to include only the highest eg. 1/8000th then normal then 1sec? Thanks, I'll keep experimenting! 

  • algovincianalgovincian Posts: 2,636

    When HDRI are created from bracketed exposures, there essentially is some form of interpolation going on in order to create a continuous representation. Since there is a loss of information on the bright end (clipping in the individual 8-bit exposures), whatever algorithm is being used must guess at what is going on in between exposures.

    Obviously, the smaller the gap between exposures, the easier it is for the algorithm. What you can get away with as far as gap size is going to be a function of the environment being captured, as well as what kind of results will satisfy you.

    Experiment away!

    - Greg

     

Sign In or Register to comment.