Hi,
I consider this great news. I don't like add-on GPS receivers in general, as I laid out elsewhere. Because external loggers then provide a superior solution. But here, Pentax did it right and added the necessary innovation over an external logger feature:
- compass
- earth rotation tracking
- write back heading info into exif
The good news is that Pentax didn't stop to innovate and stupefy the competition. I hope this is going to continue to more photographically relevant innovations in the near future
I hope that a firmware update will write both level angles information as well. As the combined info would then identify the photographed region of sky.
Originally posted by Lowell Goudge gps don't need calibration, they know where north is all the time
not true
Originally posted by pxpaulx I guess if it can tell where magnetic north is, combined with the gyroscope and gps it can know its 3rd position, we'll have to wait and see.
The amazing thing is that almost all sensors are used in this solution:
- GPS position (3D)
- compass heading (1D)
- level meter angles (2D)
- focal length
which gives the full 7D beam parameters of a shot, enough to compute a list of all stationary photographed objects, be it sky or earth-bound
Google is working on a database which is fine-grain enough to return the names of photographed buildings if you upload the exif from such a pentax shot.
Somebody here in Germany actually filed a patent covering the case where a camera then would point you to an interesting photographic opportunity (suggesting direction and zoom). It was meant to be joke patent but now all of a sudden, could become reality
However, it won't replace a scope with mount and tracking motor. Typically 2 min for 200mm focal length won't compete with a scope able to track 2000mm for an hour ... OTOH, if you have to live on a budget, it may be good enough to stack images.
Originally posted by v5planet I'm a bit skeptical of that... it would need to track the stars in both right-ascension and declination -- you could simulate this if the sensor could both rotate and tilt, but I don't know if it can do the latter and in any event would stretch the image like in a tilt-shift lens. Hm.
not true.
rotate and shift are sufficient to compensate for the earth rotation's first order effects. And the K-5 sensor can both rotate and shift. Near the polar star, rotation is the dominant effect (small), near the horizon, shift is it (large).
Tilt would only be required to compensate for second order effects as required for stacking multiple photos together. So, this would have to be done in a separate post processing step but would still be feasible. But for a single ~2 min exposure, tilt is neglegible.
Originally posted by Arpe It only knows position. It calculates North when the unit is moved in some direction, but when still, it only knows its position and the direction of last movement, so if you turn it on its axis, it won't know where north is.
not true.
The PENTAX O-GPS1 features a compass with ±5° accuracy. The compass accuracy is actual key to the proper functioning of star tracking. If you rotate the sensor around an axis off 5° from the true axis then this will be the main source of error. It is rather straightforward to compute the remaining star trail sizes (in pixels) from this accuracy.
Originally posted by Raybo it's a 5 min max with a 300mm lens (think about it).
In this case the sensor would not need to "rotate".
Its more like 1.5 min for 300mm and it will need to rotate as well. Under 50mm and with 5 min, rotation may actually be the main effect you need to compensate for. It certainly is if the polar star is within the image.