Background - I think I've got HDR down, and I've definitely got panoramas down. I use PhotoShop CS3 for the HDRs and I use AutoPano Pro to create panoramas. I shoot RAW and open all the photographs with the same settings for the panorama. I then export them as 16-bit TIF files from PhotoShop's Camera Raw algorithm, making sure that they are ALL set to the same temperature, tint, contrast, etc. so that they match up as well as possible in the panorama software. Problem - However, for any of you that do panoramas, especially ones that are 160+° across, you probably know that a single exposure setting can be a nightmare, especially if there's sky, because of the variation in illumination across the scene. I try to set my exposure so that the meter is never completely on one side or the other across the scene, but the camera often indicates there's as much as a 4-stop difference that I should be taking into account. Solution? - Apple just released new computers yesterday and I've ordered me a nice 8-core 2.8 GHz machine that I'll be loading with 6 GB RAM and 1.32 TB of hard drive space and a nice 30" monitor, so I will finally have hardware that will allow me to do what I propose: Is it feasible to do an HDR panorama? Besides a boat-load more time and processing power, there are some practical issues that I see due to two different possible workflows: First, you could do an HDR shot for EACH image section of the panorama. This means as you take the panorama, for each "location" within the field of view, you take your 2, 3, 4, etc. exposures with the different shutter speeds before moving onto the next frame of the panorama. Then during processing, you create the HDR of each frame and save it before importing it into the panorama software. The problem with this is ensuring the the HDR is processed the SAME way for EACH frame. Which I think would be very difficult. Second, you could do your panorama and either go through the entire field with one setting, change it, and go through the field again, or take the shots as I said for the first method ... it doesn't matter here, though doing it the way as in the first method would probably be more consistent. You then stitch the panorama normally, creating N panoramas were N is equal to the number of shutter speeds for the HDR. Then, you merge those N panoramas into an HDR image. The problem here is ensuring that the panorama projection is the SAME for EACH panorama, otherwise there will be projection offsets that aren't uniform across the field. Question - So ... has anyone attempted to do this? Can it be done? Can you think of a better way than either of my proposals, or think of solutions to the problems stated, or think of other potential problems I haven't thought of?