Camera Labs

Bonjour,

I was surprised to realise that some of my pictures hadn't a truly horizontal horizon, despite being very careful at holding the camera and using the AF frame on the viewfinder.

I've found similar point reported on many forums, called twisted pictures on K10D: see for example http://forums.dpreview.com/forums/read.asp?forum=1036&message=23887059 but there are many more in French.

When shooting a horizontal line (or a vertical one), the line goes one degree down (or right) and picture needs a one degree rotating back.

As shown below, I did experience that point with sea landscapes but I considered my fault not to hold the camera correctly.

1/500 f8.0 ISO200 73mm

http://www.flickr.com/photos/rei_vilo/1231786169/

Detail of the horizon

http://www.flickr.com/photos/rei_vilo/1231786387&size=l

Have you experienced that problem:?:

_________________
--- rei_vilo
Pentax K-5 + BG4 + DA* 16-50 + DA* 50-135 + DA* 60-250 + AF-540FGZ
reivilophotography.weebly.com

It sure can be hard getting perfectly straight horizon lines, especially when it comes to the Sea or coastlines.

When it comes to the Sea, remember it isn't a straight line when viewed with a sufficiently wide field, and your lens may also not be geometrically perfect.

Coastlines around the Sea or lakes can also be very tricky as there may be a slight natural diagonal from your viewpoint, but it ends up looking squint on a photo. Sometimes even when the spirit level in your tripod tells you it's technically right, the photo looks better if you tilt the camera so what looks like a horizon line really is straight across the image.

This is why I love grids, whether electronic or optical, so much!

I guess if the optical viewfinder weren't 100% square you could also have a problem, where even when it looks correct, the image will be a little squint. I'm not sure if any cameras suffer from this more than others, but it's another advantage for having a Live View feed direct from the sensor...

Gordon

Gordon,

Thanks for the numerous technical explanations. The mentionned forum was pretty serious about the issue.

From the top of the cliffs of Normandy, you can see the earth is round! With a little luck, a very good weather and a pint of imagination, British coast is just in front!

Best regards from France,

_________________
--- rei_vilo
Pentax K-5 + BG4 + DA* 16-50 + DA* 50-135 + DA* 60-250 + AF-540FGZ
reivilophotography.weebly.com

Yep, that pesky round Earth sure can be a problem when all you want is a nice straight line!

On another note, your photo really reminded me of one I took on the UK Dorset Coast of 'Durdle Door', here:

[img]http:///www.cameralabs.com/images/forumpics/Durdledoor.jpg[/img]

PS - I shamelessly cloned out the tourists on the beach!

Bonjour,

I reported before that I noted some of my pictures were turning right. The example I provided wasn't maybe the more explicit , but more pictures, especially for architecture, suffered from the same problem .

So I decided to investigate: here is what I've found so far.

Many forums talk about the problem

dpreview
pentaxuser.co.uk
pentaxforums
aficionados K10D (in French)
K10D forums (in French)

Here's my test

The K10D and an object (here, a level-meter) are placed on a table, at 120 cm. Both are centered.


The picture taken by the K10D is perfectly right and horizontal. I used the remote + 3 seconds, and the shake-reduction system is off.


However, the viewfinder shows something different: the image is twisted. That means that, if I rely on the viewfinder to check my picture is horizontal, I may have a twisted picture.


I performed the same test with a DA 18-55mm at 35mm and 55mm, and with a DA* 50-135mm at 50mm. Same results .

So I plan to send my K10D to be repaired

Are you experiencing something similar with your K10D

_________________
--- rei_vilo
Pentax K-5 + BG4 + DA* 16-50 + DA* 50-135 + DA* 60-250 + AF-540FGZ
reivilophotography.weebly.com

Hi Rei, nice tests...

I have experienced this before on some other cameras and it illustrates how we are at the mercy of a perfectly aligned and fitted viewfinder.

It also illustrates the value of Live View, as that really is showing what the sensor will record...

Bonsoir Gordon,

I agree with you, but this is a never-ending game

Personally, IMHO, I rather prefer simplicity and quality rather than new features, that may bring their own drawbacks as well as uncorrect usage by the photograph (like the shake reduction switched on while on a tripod )... And how to check a 1° spin on a 2" screen?

Back to my K10D, I went to Pentax France this afternoon, and let my K10D to be repaired. After a quick review and thanks to the pictures above, the after-sales technician said the focusing screen on the viewfinder may have moved a little. Fixing it would take two weeks. That means two weeks without taking picture

Stay tuned for the come-back!

_________________
--- rei_vilo
Pentax K-5 + BG4 + DA* 16-50 + DA* 50-135 + DA* 60-250 + AF-540FGZ
reivilophotography.weebly.com

Bonjour,

I received today a message from Pentax saying my K10D is fixed and sent back to me.

I hope to review it on Saturday and post some news.

Stay tuned

_________________
--- rei_vilo
Pentax K-5 + BG4 + DA* 16-50 + DA* 50-135 + DA* 60-250 + AF-540FGZ
reivilophotography.weebly.com

Bonjour,

I received today my fixed K10D and rushed to perform tests.

The repair slip details 4 operations: CCD parallax adjustment 0,3°, image tests, computing programming, general check.

Here are the results

I used the same protocol as before, described here. I used the remote + 3 seconds, and the shake-reduction system is off.


First batch of tests



Surprise, the picture is no longer horizontal while the K10D is horizontal It is the opposite problem of the one I had before!



Good news , the viewfinder shows exactly what is capture by the image. I could rely on the viewfinder to check my picture is horizontal.

Bad news , the viewfinder and the image sensor are not longer aligned with the body of the K10D.


Second batch of tests

So I tried a second series of tests. Here, the K10D is aligned so the picture on the viewfinder is perfectly horizontal. To obtain that, I simply put a CD under the right part of the K10D body.



Here what I see on the viewfinder.



No surprise, the viewfinder and the picture are the same, and both are horizontal .




For better readability, full-size pictures are available here.


Conclusion

So I have mixed feelings:

• I could trust the viewfinder now, I get what I see,
• but I could no longer rely on the alignment given by the K10D body.

For me, this seems really strange

Shall I send my K10D back again and take no picture for one more month

_________________
--- rei_vilo
Pentax K-5 + BG4 + DA* 16-50 + DA* 50-135 + DA* 60-250 + AF-540FGZ
reivilophotography.weebly.com

Bonjour Gordon,


I better understand now the reason for Live View, but I still doubt how I could identifiy a 1° shift on a 2.5" screen

Most important, I get what I seen through the viewfinder but I'm not a 100% happy

Have you experienced similar problem with other bodies

Best regards,

_________________
--- rei_vilo
Pentax K-5 + BG4 + DA* 16-50 + DA* 50-135 + DA* 60-250 + AF-540FGZ
reivilophotography.weebly.com

Hi Rei, thanks for your detailed testing - I really feel for you!

I'm pleased your viewfinder and sensor are now aligned, but that is annoying that they're not aligned with the body!

Sadly I suspect many DSLRs suffer from this sort of error and most people just put it down to human error.

Maybe there is some warp in your wooden sheet? I would prefer to see a test where the camera is mounted on a tripod with a spirit level, so that you know it is absolutely level. Then again this is also prone to error as the base plate and tripod mount may not be perfectly square. You could alternatively try a hotshoe mounted spirit level, but again the hotshoe may not be square either. It's a tricky one to test.

Maybe some others may wish to check this on their cameras too?

At least you can now see what you're going to get anyway!

Bonsoir Gordon,

Thank you for your nice comments, I really appreciate.

Actually, I rather prefer to get what I see on the viewfinder that having a picture aligned with the body, since the camera is mostly on my hands and a dSLR is based on the viewfinder. That's why I'm happy with the fix.

My main disappointment is originated by the fact my K10D comes from the very Pentax repair shop with a new flaw and no explanations! I would try to get some answers from Pentax. Maybe other K10D users are experiencing the same troubles

But, most important, I could now take pictures consistent with what my eyes see

Best regards,

_________________
--- rei_vilo
Pentax K-5 + BG4 + DA* 16-50 + DA* 50-135 + DA* 60-250 + AF-540FGZ
reivilophotography.weebly.com

Bonjour,

The answer from the Pentax technical department is that having the sensor, the viewfinder and the body aligned all together is very difficult.

In such a case, they give priority #1 to the alignment of the sensor and the viewfinder over other combination.

I'll try my K10D the week-end in real conditions and see if the misalignment of the sensor with the body is prejudicial.

I'm offered to re-send my K10D to the maintenance service. I'll think it twice before sending it again, since, if the repair takes another month, X-mas pictures'll be at jeopardy .

My question remains: are other dSLR concerned

Best regards,

_________________
--- rei_vilo
Pentax K-5 + BG4 + DA* 16-50 + DA* 50-135 + DA* 60-250 + AF-540FGZ
reivilophotography.weebly.com

Hi rei,

I'm glad you've got the viewfinder correctly aligned with the sensor now as that's a pretty fundamental requirement. If I've got my sums right, assuming the sensor assembly is attached to the body with mounting points 20mm apart a 1° misalignment could be the result of a tolerance of 175 microns at each mounting point. I'm not qualified to know whether that is good or bad with today's assembly methods but I suspect that a lot of cameras might have this problem as the chances of ever noticing it in real life situations, as opposed to the test bench, are slim.

Bob.

_________________
OM-D E-M5 + ED 12-50mm, H-PS14042E, H-F007014E, L-RS014150E, M.ZUIKO DIGITAL 45mm 1:1.8.
Gitzo GT1541T, Arca-Swiss Z1 DP ball-head.
Astrophotography: TEC 140 'scope, FLI ML16803 camera, ASA DDM60 Pro mount.

Bonjour bob,

Thank you for your message and your calculation (23,5 mm x sin(1°) / 2 close to 0,2 mm = 200 µm)

I fully agree with you: having the viewfinder correctly aligned with the sensor is the msot important.

However, what I don't understand is why the Pentax maintenance department moved the sensor (instead of the viewfinder) to correct the misalignment of the viewfinder with both the sensor and the body (the sensor and the body were aligned).

I keep asking the maintenance departement...

_________________
--- rei_vilo
Pentax K-5 + BG4 + DA* 16-50 + DA* 50-135 + DA* 60-250 + AF-540FGZ
reivilophotography.weebly.com