Everything you want to know about amp Camera Technology is right here. How it works, how it saves data, and when it’s going to be available.
Contact us if you have additional questions that we can answer about amp Camera.
Everything you want to know about amp Camera Technology is right here. How it works, how it saves data, and when it’s going to be available.
Contact us if you have additional questions that we can answer about amp Camera.
Very cleverly. At the heart of the amp camera is a patented method for splitting up the light from a single camera lens onto three image sensors. All images are captured using the exact same exposure time and are spatially identical. The images are separated by a large number of stops, typically over 3.5 stops. Once captured, images are combined using proprietary merging algorithms. Video is output via HDMI in tone mapped stream.
If you’re still longing for information on how amp works, check our the Technical Paper that was accepted to SIGGRAPH. It is a detailed account of our prototype, algorithms and results.
We use a variant of Long Exposure (LE), Medium Exposure (ME), and Short Exposure (SE) terms typically used for describing HDR photography. This is done because each of our images is captured using the exact same exposure time, so these older terms are misleading.
What terminology is used with AMP?
So, use caution with comparing amp terms to traditional terms, specifically the LE term. The video titled “amp Melting Snow” gives a good overview of how amp captures raw HE, ME and LE frames to produce an HDR frame that can then be tonemapped any number of ways for display.
Yes. We currently have 2 issued, 2 pending, and 3 international patents for amp Technology. These are
owned by our parent company, Contrast Optical Design & Engineering, Inc.
Yes, amp was designed to be sensor independent and therefore can accommodate a variety of sensor sizes, resolutions, and frame rates.
We will always strive for the maximum stop separation between images for a given sensor, thus maximizing the dynamic range of the amp camera system. However, maximum stop separation is a function of the sensor characteristics. amp Gen I prototype had over 3.5 stops between HE/ME and ME/LE images, this is a typical range.
In order to maximize the dynamic range using the fewest sensors, we separate our images
by as many stops as the image sensors can handle. HDR photography uses closely spaced
exposures to ensure proper blending. One reason this is done is to minimize motion blur due to
the different exposure times for each images. amp cameras capture all three images with the
exact same exposure time, and at the same moment in time, ensuring that all three images are
spatially identical, including motion blur. Therefore, amp is the only technology with the freedom
to stretch the number of stops between images.
A stepped neutral density filter is illuminated using an integrating sphere and a very bright
light source. HE, ME, and LE images of this uniformly-lit, stepped-density filter were then
captured and merged. This image was then analyzed to determine the number of stops that are
identifiable.
amp cameras are capable of producing images with over 17 stops of distinct range.
Gen II amp cameras use a Nikon F-mount. Any manual (non-electronically operated) lenses
that fit this mount will work with our Gen II cameras. And our amp technology resides “behind the lens”
so we can build custom cameras that adapt to any lens mount.
amp has no FOV limit. Two-camera rigs not only have a very strict 45-degree HFOV limit, they
also suffer from change-of-angle effects for different FOV angles reflecting of their beamsplitter.
See our publication for a detailed explanation of why these methods can’t really be used for
FFOV angles much greater than about 20 degrees. SPOILER ALERT: It’s basically due to the
fact that two-camera rigs have the beamsplitter that is located in front of the camera lens.
With amp, you still have complete control over the exposure, in every way. The amp optical
engine splits the light up so it’s at different levels when it hits all 3 sensors, but the 3 sensors
all have the same settings, which you get to set yourself. Think of it like this: you set the camera how you want to catch the scene’s lowlights,
and there are always two extra cameras “built right in” to catch all the highlights of the scene.
Simple: it’s not a fair comparison, unless you compare footage from our amp camera with
footage shot using a camera with the exact same sensor we use. amp is a
sensor-independent technology, so it does not make sense to compare an amp prototype to a commercial camera that uses a different type of sensor — because
then you’d just be comparing the two different sensors to one another.
For the best direct comparison, see our video titled “Getting the Whole Picture”. This video does several
direct comparisons of our footage to that shot using a single SI-1920 cinema camera. Same
exact lens, same exact scene, same exact sensor. This video is a direct comparison, apples to
apples, and it really shows the extension of dynamic range in both directions — more detail and
less noise in the lowlights, more detail and less saturation in the highlights.
Absolutely! Lighting is the key to the look and feel of any great movie or shot. That will never change. Our camera is here to help give filmmakers and cinematographers even more flexibility with lighting and allow them to be able realize shots they never dreamed possible.
Videos are the key to showing the real power of true high-dynamic range imaging and our amp camera. However, this can be a tricky thing to convey. Our Gen2 cameras capture a very large dynamic range of light levels, and then produce a live, tone mapped output, which is designed to maximize the amount of range displayed on a typical screen. This live tone-mapped output looks very natural and life-like, and so there is none of the typical “dreamy HDR” look to this live output video. What you will notice is that there are no saturated or “toasted” areas of these videos, like you’d see if the scenes were shot with a typical camera. Where other cameras would just see white or would see distorted highlights (also called a toasted look), we see detail and accurate color all the way through the light range. Using our adaptive tone-mappers we can see accurate colors and details not only up into the highlights, but also down into the lowlights.
Gen II amp currently does not store RAW data, but provides a real-time
HDMI output which can be recorded using standard video recording equipment.
Currently our Gen II cameras only have a real time tone map operator (TMO) output via HDMI. We are incorporating saving of RAW, uncompressed data into next gen cameras.
amp Gen II cameras are currently in production, but they are being used for testing and
evaluation only. Contact us if you’re interested in helping us out by providing feedback.
The amp camera produces a real-time, tonemapped video stream. In the Gen II amp cameras, the
HE, ME, and LE images are combined using custom, on-board electronics, in real-time. The amp video
stream is then tone-mapped to 8-bits for display, then output in real-time via HDMI.
The amp goal is to provide the best, most natural images possible. Therefore, it’s expected that
amp will use a limited variety of very natural Tone Map Operators (TMO) for real-time output.
Currently, we have a variety of different real-time TMOs available on board amp. These include, 1 Pure Luminance (false color), 2 Kiser-Rienhard natural TMO, 3 Basic log TMO for Natural scenery and 4. custom local TMO used for extremely high-contrast scenes.
Our merging algorithm, ECHOblend, is a proprietary mix of merging, blending, and horsepower.
Since typically-available software could not be used to merge amp images, we hand-crafted our
own. It was essential that the algorithm be robust enough to handle exposure separations of up
to 5 stops or more if necessary. Blending is also key to gracefully handle any transition zones
in the combined image. Finally, our algorithm had to be adaptable to real-time implementation.
ECHOblend has all of these features and more.
See our publication for more information.
Nope. Due to the large separation between our images, most merging algorithms completely
fail, or add undesired artifacts to HDR images. This is not acceptable to us and was the main
driving force behind developing ECHOblend, our in-house merging/blending algorithm. RAW data saved from amp is stored as a blended, .hdr or .exr series of frames and may be batch tone mapped using commercially available software.
For more info on the mathematical theory behind ECHOblend, see our recent publication.
Not at this time. Contact us and let us know your interested in our test footage raw data and we
will consider your request.
The AMP Gen II camera sports a solid metal case, milled from a single billet of high-strength
aluminum. It measures 3.5” x 4” x 5” and weighs in at just under 5 lbs, without the lens.
amp cameras are controlled wirelessly via tablet or smartphone interface.
We are currently building custom amp cameras. Please contact us to learn more.
Not at this time, however, it’s possible that some of our Gen II amp cameras will be available
through these avenues. Contact us to be added to our list for consideration.