Photography

HDRinstant creates high dynamic range stills from frames of video

HDRinstant creates high dynamic range stills from frames of video
An HDR still made from successive frames of video using HDRinstant
An HDR still made from successive frames of video using HDRinstant
View 3 Images
An HDR still made from successive frames of video using HDRinstant
1/3
An HDR still made from successive frames of video using HDRinstant
The main requirement is that the footage was shot with the exposure set for the highlights
2/3
The main requirement is that the footage was shot with the exposure set for the highlights
The software "morphs" the individual stills so that their composition is identical
3/3
The software "morphs" the individual stills so that their composition is identical
View gallery - 3 images

Ordinarily, cameras either expose for the dark areas of a scene by leaving everything else over-exposed, or for the brightest parts of a scene by leaving everything else under-exposed. Thanks to the advent of high dynamic range (HDR) photography, however, it's now possible to produce single images in which everything is exposed properly. Although HDR photos are typically captured with still cameras, HDRinstant software allows them to be created from frames of video.

On an HDR-capable still camera, a series of differently-exposed images are automatically merged into one composite photo, in which it's possible to see both what's in the shadows and in the sky – as an example. HDR shots can also be created manually by bracketing the exposure over a series of photos (using a tripod to keep the framing consistent), which are then merged on a computer.

In the case of HDRinstant, though, users start by choosing the frame of recorded video that they would like to turn into an HDR still – the main requirement is that the footage was shot with the exposure set for the highlights. The program then selects a series of frames adjacent to that one, and incrementally enhances their exposure to reveal what's hiding in the shadows. Those altered frames of video are next digitally "stacked" together, to form one still in which all of the exposure levels are represented.

The main requirement is that the footage was shot with the exposure set for the highlights
The main requirement is that the footage was shot with the exposure set for the highlights

Because there was likely to have been some camera movement as the video was being shot, however, the frames won't perfectly line up with one another, potentially resulting in a blurred composite image. To get around that problem, the software "morphs" the individual stills so that their composition is identical – no tripod is required while shooting the footage, although obviously the steadier the shot, the better.

HDRinstant works on Macs or PCs, and is available either as a stand-alone program or a plug-in for Adobe Photoshop Lightroom. You can currently get a beta version for free, although watermarks will be visible on the images. The full stand-alone can be pre-ordered for US$49, with the plug-in priced at $29.

The system is demonstrated in the video below.

Source: HDRinstant via sUAS News

View gallery - 3 images
3 comments
3 comments
Nairda
Just a thought. With most modern still cameras taking 1080p at 60 frames, would it not be conceivable to apply this HDR tech to the entire video.
ie - what is the technical limitation for the software to process highlights of every second frame, then recombine the video in HDR with an effective 30fps output.
Facebook User
Hi Nairda,
Thank you for your very pertinent comment.
For the time being, HDRinstant is only for creating HDR images from video or burst mode shots.
We are currently working towards applying this technology to video.
Thank you.
HDRlog SARL


Astro Rosaire
These pieces of software are incredibly simple (conceptually). Most image reconstruction can be performed using a class of techniques called super resolution reconstruction in which many images of low quality are combined into high resolution images. It's quite fascinating and incredibly powerful for imaging. One could take garbled video and run reconstructions that make the result clearer than the original video could ever be! I'm currently using it in my PhD. Fun stuff :)