On 23.07.2008 12:31, kid2k4 at email.it wrote: > To Abel: Thank! > We have thought about the scanner's depth of field after some test with > A4 size. > I must tell you the truth, we really like the effect nuanced we get, we > are also working on this.
this proves that I have hardly an idea of artistic thinking ;) > A technician told us that the sensitivity of > ccd can be adjusted, although there is as yet unclear. Right, you can add an additional amplifier to the signal path. But I think you will most likely need additional lamps too. And you should be aware that you will get more noise into the signal, similar to the noise you can see in images taken with digital cameras with "higher speeds", like ISO 800 or ISO 1600. But this might again be too technical thinking ;) > We also think about a Frankenstein-scanner made of various parties other > scanners (A4 course). We need to do so because we not have a budget large. > We have also seen this site <http://www.artila.com/>, there are some > controller and engine, the controller contains a small linux and may be > (as we understood) used to control engine or electrical appliances. We > plan to use it to manage the engine scanner. > you think we can use for our project? A small industry computer is the right choice as the machine controller, I think. But I am not sure if the processing speed and the amount of RAM of such a device is sufficient to do the post-processing, like stitching the separate images together. There is another caveat: I am not sure if a small computer can "digest" the data sent simultaneously by four of five scanners. Again, that is easy to test, if you connect four of five scanners to this machine and try to run "scanimage" simultaneously for all scanners. > Is possible to build a mega-scanner with part of A4 scanner? Well, I am not yet convinced that you will get a, let's say, "sufficiently recognizable" image when you use regular scanner optics for a distance of 0.5 or 1 meter between the scanned object and the scan head. But that it easy to test. Regarding scanners, you should look for a model where the Sane backend controls much of the "basic work", like controlling the motor that moves the scan head. Most more expensive scanners have a built-in microcontroller which receives relatively simple commands from the host computer, like "set the scan size and the scan resolution to values x,y,z" or "start a scan" or "read 4 kB of image data". The microcontroller then automatically ensures that the lamp is switched on, often calibrates the CCD sensor, and controls the motor of the scan head. These scanners would be the wrong choice, because their microcontrollers will probably become confused, when you try to aquire an image from a largely disassembled scanner. For example, they might expect a signal from a mechanical sensor/switch that tells "scan head reached the end of the scan area". I never had a closer look into the backends for these cheaper devices, so please take the following with a grain of salt, but from a very quick view into the source code I'd say that scanners for the gt68xx or the genesys backend might be useful. But the backend authors can definitely give a far more qualified comment about reasonable scanner models. > There's an engineer among yourselves? Do you mean Sane developers? Well, I am not 100% sure -- we generally do not introduce each other by profession ;) But I think especially the people who wrote the backends for the low-level backends have enough knowledge not only about software development but also about electronics ;) (BTW, I am a physicist by education but work as a software developer.) Abel
