At first glance, it seems discreet and unassuming. But when used, the digital camera proves itself to be more than just a servant to the user. It is full of good ideas and suggestions -- and it isn't shy.

This cheekiness makes the R927 typical of an entire generation. To the naked eye, it may look like a perfectly normal digital camera. But that's just a disguise. Indeed, look deep inside and you'll find a powerful computer -- the light it captures with its lens is simply one of the raw materials it uses to create often complex image montages.

The R927 is the newest digital offspring of US computer manufacturer Hewlett Packard. Its lens and its photo sensor are nothing special. But its true power becomes evident when images reach the camera's electronic brain.

It is here that the R927 contains an entire battery of effects, ready to be called up at the press of a button. These effects, ironically dubbed "Real Life Technologies," are capable of transforming photos into artificial photo graphics. Some of the camera's features involve combining three, four or even more images into a single panorama, giving the new image a yellowed, "retro style" look or converting it into what the manufacturer calls a "cartoon or pop art masterpiece à la Andy Warhol."

"You press the button, we take care of the rest." That was one of Kodak's advertising slogans more than a century ago when it embarked on a campaign to revolutionize the photography market. Some of the latest digital cameras are now latching on to Kodak's claim to the point of absurdity.

With the new cameras, one would think that observation is the photographer's only remaining responsibility. But these image machines don't even hesitate to cross that barrier. By selecting the menu item "Instant Advice" on the R927, the photographer can learn whether the camera feels that a given photo is successful. The device also isn't bashful about suggesting ways to improve a shot.

This über-clever camera is by no means an isolated oddity. In fact, it is typical of many of the devices on display at the Photokina Convention in Cologne that opened on Tuesday. With the photography industry in a fundamental reshuffling process, every manufacturer wants a piece of the new pie. The new gimmicks are the bait.

The names of the exhibitors in Cologne -- including mobile phone manufacturers like Nokia and computer companies like Apple -- highlight the extent to which once-separate industries are now merging. Some photo mobile phones already offer more than 3 million pixels, a flash and lenses made by famed camera lens manufacturer Carl Zeiss. Indeed, some mobiles can already compete with conventional entry-level cameras.

At the same time, sales of compact digital cameras are suffering. Digital single-lens-reflex (SLR) cameras have hit their stride with their high resolution, excellent image quality and superb light sensitivity. Even better the trigger delay has become so minimal that it's often barely measurable. This has prompted more and more customers to turn to the bulkier but higher-quality conventional SLR technology, leading to even more cutthroat competition among manufacturers of small, amateur cameras.

To keep up, camera makers are coming up with more and more outlandish computer acrobatics:

• In Fuji's Finepix S6500fd, the camera's electronic brain automatically recognizes faces and sharpens their contours.

• When the Casio EX-Z1000 has already been focused on a person and that person moves, the device automatically follows the subject and uses its so-called "Auto Framing" feature to adjust the focus.

• Some of HP's cameras include a "slimming function" called "Pixel Diet," which makes heavyset people appear thinner, as if they were standing in front of a fun-house mirror.

• Using Casio's "Revive Shot" feature, photographers can photograph old pictures and allow the camera to remove distortion and adjust faded colors.

• Some cameras include a feature that corrects "plunging lines" in wide-angle shots of buildings that would normally make the subject appear to be tilting backward.

• Some cameras, such as the Kodak Easyshare C875, automatically select menu programs for difficult lighting conditions, such as at the beach, on the slopes or during fireworks displays.

• Some Casio cameras feature a "pre-shot" function that allows the user to place their subjects in front of any background.

• To correct distortion in poor lighting conditions, some cameras take two pictures. The first is sharp but too dark, and the second is distorted but the lighting is correct. One image is digitally superimposed on the other with features such as Samsung's "Advanced Shake Reduction" (ASR).

The list of examples could continue ad infinitum. Much of the arduous work photographers used to perform on the computer using software like Photoshop is now done automatically.

But this isn't always a good thing. Upon closer inspection, some images contain strange artifacts, mysterious edges or phantom-like double exposures, almost as if they had been created in an experimental photography course or by someone on psychedelic drugs.

For the viewer, the pictures that come out of these cameras are becoming increasingly obscure. Hardly anyone can figure how exactly -- that is, with which computer tricks -- a photograph was created. Even photographers aren't always sure what their cameras will do when they switch them on. In fact, some of the new tricks are little more than cover-ups for technical defects, say experts like Klaus Peeck of c't, a computer magazine. Some of the much-touted anti-shake programs, for example, are merely used to offset a camera's lighting challenges. This highlights a dilemma the industry faces: Customers want tiny cameras, but they also want larger and larger pixel counts. If more than 10 million sensors are packed onto a small surface, each pixel is barely able to capture light, leading to distorted and shaky images.

As the camera's optical characteristics deteriorate, the magic of software plays an increasingly important role in cosmetically fine-tuning its tiny, underexposed images -- as if a painter with night blindness attempted to compensate for his visual impairment by using brighter colors.

Camera makers are using computing power in an attempt to solve this dilemma -- with varying success. A magic marker-like function redraws watery contours while a function akin to a watercolor brush smoothes out rough surfaces. But as a result, sometimes the more delicate structures in an image, like the grid of a fence, are simply computed away, or completely new patterns and artifacts are created -- the hallucinations of the computer's electronic brain.

More and more computer scientists are now devoting their efforts to finding ways to touch up photos taken under poor lighting conditions. At DXO, a French company headquartered in Boulogne, a city southwest of Paris, 15 employees spend their days in the laboratory carefully measuring the defects of current cameras and then adjusting blurred edges, color faults and blurriness. The technology being explored at DXO, which has been around for four years now, was in fact developed for the field of "machine vision," which involves the attempt to give robots eyes. DXO expects to integrate its software into cameras by the end of the year, in yet another example of how the boundaries between photography and computer graphics are blurring.

And this is just the beginning. Even digital cosmetic surgery is already being performed automatically. In August, Israeli computer scientist Tommer Leyvand unveiled "face beautification" software that employs a few basic principles to beautify portrait photos. Under Leyvand's simple formula, the eyes become bigger, the lips fuller and the corners of the mouth are raised. The digital "surgery" takes a few seconds and is barely detectable after the fact. Leyvand says that he has already been contacted by many companies interested in installing his software into their cameras.

But who is really the author of these new computer photos? Do they represent some form of collaboration between the photographer and software developers? When a photojournalist published grossly distorted images of the bombing of Beirut in August through the Reuters news agency, the computer-manipulated images were easy to recognize and prove. But who would have been responsible for an image of one airplane circling over Beirut accidentally turning into two aircraft, because the camera's anti-shake program failed to correctly superimpose two separate images?

This sort of error would be difficult to prove, because hardly any camera stores the unaltered raw versions of the images, and because processing is not fully documented. To date, no one has come up with something on the order of "fake reduction" to accompany systems like "Shake Reduction."

Meanwhile, the race to come out with the latest in photo-computers continues, with new functions constantly appearing on the market. Take "Tourist Remover," for example. The new software offers a ready-made solution to the age-old and annoying problem of tourists constantly walking across a photographer's frame. The program essentially removes bothersome strangers from the picture, almost as if someone had dropped a neutron bomb on the scene, sparing only the intended subject.