[photoptimist]

I'm a long ways from a gyro expert, but don't they keep moving (whether that motion is going round and round or vibrating back and forth) in the same plane (same direction), even when stuff around them is moving? Then it shouldn't matter what direction the camera is pointed relative to the earth's axis of rotation, just that the direction the camera is pointed (along any of the axises measured by the gyroscopes) has changed.

If you attempt to rotate a spinning gyroscope, it exerts a directional force that can be measured and used to estimate the axis and rate of the imposed rotation. The gyro sensors in the camera can't hold the camera still but they can provide data to tells the SR system which way to pan or rotate the sensor to correct for the measured rotation of the camera body.

The camera direction really does matter alot when it comes to how the Earth's rotation affects the motion of the image.

For someone in the Northern hemisphere, pointing the camera due East, the gyro sensor will tell the SR system to pan upwards and to the right at an angle from the zenith related to the latitude and time of year.

For someone in the Northern hemisphere, pointing the camera due West, the gyro sensor will tell the SR system to pan downwards and to the left at an angle from the nadir related to the latitude and time of year.

If the shooter does not wait for a GPS lock, the camera does not know that it is in the Northern hemisphere or its latitude. If the shooter does not recalibrate the compass often enough, the camera can't be sure whether it is pointing East or West (or North or South).

If the shooter did always wait for a GPS lock and did always carefully recalibrate the compass, then that information could be used to subtract the known spin of the Earth from the measured spin of the camera to avoid a star-trail effect on Earth-bound scenery.

---

Last edited by photoptimist; 01-06-2020 at 03:55 PM. Reason: error correction (time of year effects)

[Lowell Goudge]

The reason this is a problem is that to correct for the effects of motion of the planet, the camera must know the pointing direction of the camera. Even if both the camera and subject are rotating together once per 24 hours, the camera still needs to know pointing direction of the camera to properly subtract that once-per-day spin from the data camera's gyroscope sensor. The correction for a North-facing camera is the opposite of the correction for a South-facing camera.

Knowing the camera's pointing direction requires an accurate magnetic compass reading and that requires frequent calibration to correct for changes in the magnetic signature of both the location and the camera (such as after a lens change). People's challenges with getting astrotracer to work well show how hard this can be. The Earth's magnetic field is really very weak so it's signal is easily distorted by nearby metals and magnets.

If this is the issue, then a built in gps plus the correction table for difference between magnetic and true north, and a 3 D magnetic field meter, the same as in all phones, could easily sort this out

But is the issue more the jitter in the rotation And not the rotation itself

[WorksAsIntended]

The jitter is really not that high, is it?

[Lowell Goudge]

The jitter is really not that high, is it?

I know it is an issue with astronomy thatS one reason why they have adaptive optics. And it was discussed as part of the justification / explanation for how they deduced the direction the MH370 went before it was lost.

---------- Post added 01-06-20 at 06:53 PM ----------

Neither have I! Though it seems to spin more after three or four scotches!

If you are working out ballistics for a bullet traveling out a mile, yea, I can see the issue. But working with the speed of light... Not a good rational from my sense.

The issue with bullets is not acceleration but differential velocity of shooter with respect to target when shooting north to south or south to north. People at the equator are going faster than opeople off the equator people at higher elevation are going faster than people at lower elevations

---------- Post added 01-06-20 at 06:53 PM ----------

Neither have I! Though it seems to spin more after three or four scotches!

If you are working out ballistics for a bullet traveling out a mile, yea, I can see the issue. But working with the speed of light... Not a good rational from my sense.

The issue with bullets is not acceleration but differential velocity of shooter with respect to target when shooting north to south or south to north. People at the equator are going faster than opeople off the equator people at higher elevation are going faster than people at lower elevations

[WorksAsIntended]

Jitter in earth rotation causing different directions of flights?

Jitter in transmitter respond times maybe, but the earth rotation? That is far from reasonable in my opinion.


The only jitter in astro physics I know is those of lighter planets when searching new ones by doppler method. I mean, I never really worked in this area but I am also not that far off. Can you provide me a link with more information?

---------- Post added 01-06-20 at 05:05 PM ----------

If this is the issue, then a built in gps plus the correction table for difference between magnetic and true north, and a 3 D magnetic field meter, the same as in all phones, could easily sort this out

Actually thinking about this I think you dont need to calculate it, you can just measure it as long as you got an characteristic value range, at least if the motion is stable in time it should be easy to do so and easy to do more precisly than by compas and gps.


That of course does not stay true if your jitter theory is right, but I actually cannot see a reason for that jitter in earth rotation. Thinking about it I have seen way to precise measurements without active compensation for it being all to high.

---------- Post added 01-06-20 at 05:08 PM ----------

Neither have I! Though it seems to spin more after three or four scotches!

If you are working out ballistics for a bullet traveling out a mile, yea, I can see the issue. But working with the speed of light... Not a good rational from my sense.

The speed of light is of complete irrelevance to this, as you stated correctly it is very fast, too fast to matter in that case. The relevant factor is the ratio of rotational speed and shutter speeds. If you correct by pure data, you "overcorrect" the earthmovement too. The object you are taking a photograph of is not so the sensor slowly shifts (relativ to the photographed object) in the opposite direction of earth movement.

[photoptimist]

If this is the issue, then a built in gps plus the correction table for difference between magnetic and true north, and a 3 D magnetic field meter, the same as in all phones, could easily sort this out

But is the issue more the jitter in the rotation And not the rotation itself

Quite true and it seems that Pentax cameras do contain all that. But the problem is the need for recalibration any time something changes such as changing lenses.

The Earth's magnetic field is really weak (about 25-65 microtesla). That signal is easily warped or overwhelmed by nearby iron/steel objects and magnets. I've seen the compass on my iPad get totally decalibrated -- thinking it was always pointing South plus or minus a few degrees regardless of orientation. And I've seen similar anomalies in the K-1 and GPS watches.

That means that the 3 D magnetic field meter needs to be recalibrated before using it for SR. Pentax Astrotracer has this same issue which is why it takes frequent recalibration to get decent results with Astrotracer.

[WorksAsIntended]

I still keep thinking about the jitter. I know there is a periodic change in speed because of the mode the earth is "wobbling" in. But this again is very easy to get out of the data as the frequency is easy to identify and very characteristic.

So there must be a less predictible change to the rotational speed for the jitter to be relevant, or even to call it a jitter in the first place.

Which is also a point against the jitter theory is that the achieved stabilisation time is exactly the one you would predict from the average earth movement without any jitter considered.

[Lowell Goudge]

Jitter in earth rotation causing different directions of flights?

Jitter in transmitter respond times maybe, but the earth rotation? That is far from reasonable in my opinion.


The only jitter in astro physics I know is those of lighter planets when searching new ones by doppler method. I mean, I never really worked in this area but I am also not that far off. Can you provide me a link with more information?



That of course does not stay true if your jitter theory is right, but I actually cannot see a reason for that jitter in earth rotation. Thinking about it I have seen way to precise measurements without active compensation for it being all to high.

The issue with the satellite transmission and earths rotational jitter, comes about because the transmissions from the plane to the satellite suffer from the difference in rotational jitter Between an object in synchronous orbit and one more firmly tied to the earths rotation (the plane). At least that is how it was explained, I never really bought it because a satellite in orbit at the equator would see the same impact of jitter on the plane in Either direction north or south.

[WorksAsIntended]

The issue with the satellite transmission and earths rotational jitter, comes about because the transmissions from the plane to the satellite suffer from the difference in rotational jitter Between an object in synchronous orbit and one more firmly tied to the earths rotation (the plane). At least that is how it was explained, I never really bought it because a satellite in orbit at the equator would see the same impact of jitter on the plane in Either direction north or south.

I highly doubt this too. jitter in this case is usually created by electronics using their own clock and its quantisation as well as simpy electronical noise in recievers and transrecievers.

Edit: https://www.atsb.gov.au/media/5187038/mh370_burst_timing_offset_dec2014.pdf
This pretty much states exactly my thoughts about it.

[swanlefitte]

I have to wonder if this is based on average hand held user expectations. A very stable user who can get a shot at 0.5 seconds can get a shot at 6.5 stops slower is the same as an unstable shooter who can get a shot at only 1/180 6.5 stops slower? Would a really stable person who can get a 5 second photo be able to get a 3 minute photo in theory or is that limit based on the human sweet spot?

[photoptimist]

I have to wonder if this is based on average hand held user expectations. A very stable user who can get a shot at 0.5 seconds can get a shot at 6.5 stops slower is the same as an unstable shooter who can get a shot at only 1/180 6.5 stops slower? Would a really stable person who can get a 5 second photo be able to get a 3 minute photo in theory or is that limit based on the human sweet spot?

The 6.5 stop limit seems to be with respect to format and resolution-appropriate variants of the old 1/F rule.

The Earth's rotation (if not correctly handled) causes the SR system to add up to a 1 pixel blur at shutter speeds of about 150/F on the K-1 and about 100/F on a 20MPix M4/3 camera. Thus, even the most stable shooter would start to see Earth-motion blur in their images at those slow shutter speeds.

Shaky people might seem much better (or worse) numbers of stops of stabilization (relative to their personal limits on shutter speed) depending on what kind of shake that have. Someone with fast little back-and-forth twitches but no drift might see huge improvements. Someone with no twitches but lots of drift might see less improvement. It all depends on how long it takes the unstable person's hold to wander out of the motion-limits of the SR system.

[DWS1]

Hmmm... I get where you're coming from, and I wholeheartedly agree with the general concept of learning better technique rather than relying on technology to make up for deficiencies therein; BUT...

What about someone like me who knows and applies good technique but is simply no longer as steady as he used to be? At just 50 years old, my days of accurate pistol shooting (even two-handed) and rifle shooting (unless prone or using a rest) are long gone. With photography, I'm similarly challenged. Where once I could stand in the field with no tripod, not even a tree or fence post to brace against, and achieve good results at lower shutter speeds, now I generally need to brace against something - unless, that is, I'm using image stabilisation, which gets me right back to where I used to be and better still.

So, whilst I agree that image stabilisation shouldn't be a substitute for good technique, it can be an extremely useful feature when coupled with it.

I believe you and I are in agreement here. I'm am 60+ and can no longer hold a camera as steady as I once did, so I have modified my technique to account for this, just as you modified your rifle shooting technique to account for your changes in physical capabilities. Yes, image stabilization is very helpful for hand held shots. I use it, and I'm thankful for it. I just think if one needs an increase in shutter speed of 6.5 times or more to get a blur free image caused by camera movement, then one should also be looking at other factors to get it like a tripod, a rock, a bench, a table, the ground, increase ISO, use a flash, etc. instead of just relying on technology to make it happen. I guess relying upon it totally goes hand in hand with the crowd that believes it doesn't matter how they compose and capture an image, they will just fix it all in post, which I also don't agree with.

[clickclick]

Does it make any difference if you jump?....

[StiffLegged]

Indeed it does. There’s a YouTube video somewhere of experiments measuring how big a jump will defeat the IBIS in Sony A7 cameras and how much better the Panasonic cameras cope with landing on a trampoline instead of an air bag. The experiments also clearly show how bokeh is ruined in these cases. Quite fascinating and I just wish I could find the link again.

[biz-engineer]

Earth's rotation or not, I'd like to see them get 6.5 stops of image stabilisation with my unsteadiness...

You can get 7 stops of stabilisation with a Pentax K1 if the initial shutter speed with SR disabled is 1/1000th of a sec. Enable SR and reduce the shutter speed by 7 stops gives 1/8th, that's fast enough for a K1 to take a sharp image. Now if you start with the max shutter speed (1/8000), you can get more than 7 stops.

---------- Post added 07-01-20 at 08:56 ----------

Does it make any difference if you jump?....

No, if you jump but keep your hands fixed at the exact same position, the camera and SR performance won't be affected. Basically, you are allow to move all parts of your body by as many stops as you wish, as long as you keep your hands fixed.

---------- Post added 07-01-20 at 09:00 ----------

The 6.5 stop limit seems to be with respect to format and resolution-appropriate variants of the old 1/F rule.

The resolution should also be specified. If the camera sensor had 1 single large pixel, stabilisation wouldn't be needed at all.