What's all this image processing palaver about – anyhow?

11-06-2018 | By Paul Whytock

Camera or computer? The boundary lines separating the two are way out of focus these days as snapping-machines soak up increasing amounts of electronic technology.

Long gone are the days when photographers would have to be in charge of a camera's settings to produce the desired image; be they pin-sharp, soft-focused, moodily subdued, psychedelically vivid or whatever effect they were seeking. Today's cameras have powerful image processing chips that take care of all that.

Before them though avid snappers like myself and my friend and founder of Electropages Craig, would have to understand how shutter speeds and lens apertures where creatively intertwined. How different films with different ASA ratings would impact on the final image. And try explaining the implications of what depth-of-field focus is to a non-photographer. It’s a bit like trying to explain the offside rule to a disinterested soccer agnostic.

A classic camera still resides on one of my office shelves - the Nikkormat FTn (pictured). This beast use to be my weapon of choice back in the early 1970s. With a very heavy metal construction and a totally mechanical mechanism this was a camera that you could potentially club a lion to death with. Unsurprisingly its durability made it one of the favourites with photographers covering the Vietnam War. It was absolutely reliable and with an F1.4 NIPPON KOGAKU (the founding name of what became Nikon) lens on the front of it the images produced where spot-on.

The FT's exposure control system used a needle pointer that shows when looking through the camera viewfinder and an integral cadmium sulfide (CdS) light meter. The photographer would manually adjust the shutter speed to freeze or blur motion in conjunction the lens aperture f-stop to control depth of field focus until the needle was centered between + and - signs The picture was then ready to take providing the split-image screen rangefinder indicated the lens was in focus.

Nowadays all that palaver sounds like the ideal way of missing a potential picture, which makes you realise just how good some of those war photographers were.

But these days forget all that. Pick up a Nikon D7500 and you'll find it uses an internal image processing chip called the EXSPEED 5 that can take control of just about everything a good picture needs. Things like, Bayer filtering, de-mosaicing, image sensor corrections, dark-frame subtraction, image noise reduction, image sharpening, image scaling, gamma correction, image enhancement, active D-Lighting, colour space conversion, chroma sub-sampling, frame-rate conversion, lens distortion, chromatic aberration correction, image compression and JPEG encoding, video compression, digital image editing, face detection, audio processing/compression/encoding and computer data storage and data transmission.

Armed with a snapping machine like that how can you fail to take anything other than a great picture? Well technically that's true but the one thing an EXSPEED processor cannot do for you is generate creativity, imagination and opportunity - the key elements that can result in a truly unique photograph.

But what exactly is this EXSPEED gubbins? They are Socionext Milbeaut image and video processors fabricated using a multi-processor system-on-chip architecture and they are used in a lot of cameras and mobile phones.

For those not familiar with the name Socionext, this is a chip company formed out of the LSI businesses of Fujitsu and Panasonic. Here in Europe it is headquartered near Langen, Frankfurt which makes it a near neighbour of Fujitisu Europe which, of course, is no coincidence.

The processors are based on the Fujitsu Multiple FR-V processor cores, with each single processor-core able to calculate a load of instructions and operations in parallel that are key elements for image processing. Storage and display interfaces and a DSP help to maximise the number of simultaneous computations.

A 32bit Fujitsu FR microcontroller with dual-core ARM architecture governs all the processors, modules and interfaces and fundamentally is the main control unit of the camera.

Just like computer technology, camera technology and the way in which photographic images are produced and manipulated is a rapidly evolving business.

Only recently semiconductor company THine Electronics, a specialist in high-speed serial interfaces and image signal processing and California Eastern Laboratories unveiled the CDK, Camera Development Kit.

As already mentioned, advanced camera systems require the operational compatibility of multiple components and for the electronics engineer this means greater design challenges. And ISP firmware design is one of the greatest because it depends on CMOS image sensor and related camera module components such as the lens, image improvements according to customer requirements and preferences, and technical requirements such as video compression.

Typically ISP firmware development gobbles up thousands of script lines, which demands some pretty smart programming expertise. The idea behind the CDK is to help cut development time and costs.

To do this the CDK consists of hardware, firmware and THine’s GUI-based tuning tool. This provides real-time camera image feedback, which helps with image tuning.

And image tuning is important, as most politicians will testify. And there is no doubt that modern camera electronics do help to create technically superb photographs. This is of course great for us snappers as it just leaves us get on with the business of creating exactly what story we want our pictures to tell!

 

Read More: Biometric cards beefing up anti-fraud fight

paul-whytock.jpg

By Paul Whytock

Paul Whytock is Technology Correspondent for Electropages. He has reported extensively on the electronics industry in Europe, the United States and the Far East for over thirty years. Prior to entering journalism, he worked as a design engineer with Ford Motor Company at locations in England, Germany, Holland and Belgium.