Jump to content

I have troubles with the sharpness of my images.


JOHANVAN

Recommended Posts

Posted

Dear Reader, excuse me for my English. (Dutch speaking)

I build a very fine static dome out of plexi glass (diameter 80 cm, radius 40 cm), which is a dream to work with. (60 holes)

I use a cold beam light to illuminate my objects (trilobites in this case), with a flexible fiber optic of 2 meter. I could not use LED's and software while it was to expensive (I am an old paleontologue of 66 years old) and know nothing about their electronical drives.

I will send plans and photographs to people interested in a cheap hemisphere. Taking the 60 pictures is a proces of 10 minutes and totally by hand; one by one !!!

I take an overall picture with four Balled Pro Led Lights in top of my hemisphere for a general view.

I work with a Nikon 7000D and different macro lenses so as a Nikon 60mm , a Sigma (50mm,105mm and 150mm). The flatter the object the better to use the 150 mm.

I use "Helicon Remote Control" as my screen on my laptop Mac Notebook during the photographing and  for taking the pictures with my mouse (cordless...a great advantage after all).

The trouble is: my images and despite the possibility to double my image on screen is not clear enough, to see if my images are sharp or not (there is also a : "find edges" possibility, but of no use, if one does not photograph in stacking mode).

I use also the DOF metering as an app on my ipad and normally I should have enough margin to set sharp.

Allthough I use a macrolens at a distance of 40 cm I can not become very fine enlargments, which is normal and not so bad, but in processing as .rti images or .ptm all of the images lack sharpness and have much noise.

To help these I use programs onto  my images  as PhotoZoom Pro 6 to sharpen in batch from .dng to .tif;  and use MacPhun Noiseless Pro to .tiff for unnoising the images (one by one....)

The .tiff images: I set them over to .jpg by a software called: ContentaConverter PREMIUM V 6.3 which do also a very fine job.

Of course I use RTI Builder and RTI Viewer and allthough everything I have good results after all; (I am working on trilobites which are not flat and go sometimes on macro and beyond !!!)

Can somebody help me in saying how I can get better and sharper images, please ?

Thanks for helping;

Johan

 

post-871-0-89658500-1439818604_thumb.jpg

post-871-0-33970400-1439818606_thumb.jpg

post-871-0-61474800-1439818608_thumb.jpg

Posted

Hi Johan,

 

A suggestion is to streamline your workflow to eliminate the sharpening and noise reduction during the conversion from .dng to .tif and then to .jpg.  Check the guide to RTIBuilder and you'll see that it recommends no sharpening, no noise reduction (also no electronic image stabilization, fyi, although you didn't mention using any), and no processing of individual images--all images must be batch processed exactly alike to get the best results.  You're using a number of software programs in your workflow that have the potential to alter the images and this could be affecting the quality of the final results.  The only batch adjustments you should need to make are white balance and exposure, nothing else, and then try converting directly from .dng to .jpg at the highest quality settings.  If you follow the steps in the RTIBuilder guide very closely, it could improve the RTIs you're getting. 

 

Another suggestion is to focus in Live View or use the camera's autofocus setting for a test image, and then switch to manual focus and don't touch the camera or focus ring once you've gotten a good sharp test image.  Generally, you'll get sharper images if you don't set the aperture too small (use lower f-stop numbers than f/11 for wider apertures).  As you noted, the longer focal length lenses (105 mm and 150 mm) work better with flatter objects because of limited DOF, but you still might be tempted to use f/11 or higher to improve the DOF, which adversely affects image sharpness because of diffraction. 

 

Finally, if you're manually changing the position of the light source in the dome, there could be some small vibrations causing the images to lose registration.  If this is the case, the normals won't calculate properly and you'll get blurry artifacts in the RTIs.  You might be able to correct the image registration using PTGui or perhaps Photoshop tools, but I haven't attempted this myself.

 

Thanks for the images of your dome and setup--it looks very interesting.  I hope some of this helps!

 

Best,

Taylor

Posted

Thanks Taylor for the great advice!

 

I want to echo what Taylor said about no sharpening and do not individually process images.  While the software will not complain, you may be introducing artifacts from the sharpening, not in the original subject.  Also the software assumes that all the images are captured and processed identically with the only thing changing being the light position.

 

In addition to recommending the Guide to Highlight Image Processing, I'm also going to recommend a thorough read of the Guilde to Highlight Image Capture.  The capture guide has detailed advice about getting images in focus.  For more sculptural subjects you may need to use a wider angle lens and/or move the camera farther away.  AS Taylor says, increasing depth of field through using a smaller aperture loses resolution, and will not give you sharp images.  The ideal aperture to use is somewhat dependent on your lens, but for many SLR lenses the sharpest aperture is between 7.1 and 8, and the going above f11 or f13 is not advised.

 

The guides can be found in the download area of our website: http://culturalheritageimaging.org/What_We_Offer/Downloads/

 

There are lots of places to read more about the affect of small apertures on sharpness.  Here's one I like:  

http://www.kenrockwell.com/tech/focus.htm

 

Good luck.

 

Carla

Posted

I have to thank both of you dear Taylor Bennett and Carla Schroer for all the good advices you have given to me.

 

There was one remark I especially appreciate about light- drift in capturing the images.

The constructor of the hemisphere took attention to drill the sayed holes towards the epicentre of the sphere but as said, changing the light position is done by hand.

The thickness of the sphere is only 6mm;  the time between opening and closing the shutter can be long...so unwanted movements can occur meanwhile .

I made a construction to hold the light fiber-optic in place as sturdy as possible; but now I have my doubts about the accuracity.

One can say this could happen also in taking photographs of greather objects, but not in the same way, while here; only details of "trilobites" are of interest. (so: having light drift on very small fields of interest will affect the end results seriously!)

About using the mentionned software; as you said, the images and use of software have to be streamlined all the way. I may be wrong but I had no really bad experiences while my scoop was to prove and not to define and all pictures were done in batch mode or with predifined options.

On the other hand I am convinced that you and Carla are right in saying to use only white balance and exposure. . I will try PTGui Pro v 10. to see if  the light position- drift would have  affected the results. If yes; no other software will be necessary.

Thank you Carla for the interesting articles about sharpening and apertures. I will read them with the most interest.

 

Best Regards,

Johan Vanhauwaert

 

 

Posted

Hi Johan,

 

Just to clarify, the problem with vibration when moving the light source isn't the accuracy of the light position; it's any slight movement between the camera and the object.  Every pixel of every image in the sequence must align precisely, which is referred to as "registration," otherwise the RTIBuilder software gets confused and doesn't calculate the normals correctly.  Since the pixels are very small, even a slight vibration will cause the images to mis-align, or to lose registration.  This mis-alignment of pixels in a stack of images is what PTGui, or possibly other tools such as later versions of Photoshop, is intended to correct.  As I said, I haven't tried using these tools myself. 

 

Moving the light source to different positions and securing it to the dome could transmit small vibrations to the camera and/or stage where you've placed your trilobyte for photographing, and these vibrations cause movement between the camera sensor and the object.  However, if the dome is physically isolated from the stand holding the camera and from the stage where the object is placed, the vibrations wouldn't be transmitted and it would save you a lot of processing work with PTGui.  PTGui is a tool to correct misalignment of pixels, but it's better not to have to use it.  I couldn't tell from your images if the dome is isolated.  If it's not possible to isolate the dome, you could try holding the fiber-optic without touching the dome and use the holes simply as a guide to positioning the light source, but it's hard not to make any contact while triggering the shutter and holding the fiber-optic.  This is why some domes use a controller like Arduino and LEDs to switch the light position.  Others have used simple incandescent lights and arranged switches to avoid touching the setup.  You can find some examples elsewhere in this "Dome Method" forum or the "Projects" forum.

 

I also haven't worked with a "cold beam" and fiber-optic light source, but a consideration is whether the light source is evenly distributed across the entire surface of the object.  If the beam is too narrow or has a center hot-spot, halo, or other uneven distribution, it will also affect the accuracy of the normals.  You might still get a good enough result for visualizing the surface texture, but the normals won't be as accurate as possible if the light beam is too concentrated or not centered on the object.  If it's possible to change the "spread" of the light beam produced by the fiber optic, you want it to be as evenly distributed as possible, and centered on the object.  This becomes harder to achieve at the low-angle "raking" light positions.  That's why CHI recommends setting speedlights at their widest-angle beam spread, if there's an adjustment.  I'm not sure how much control you have with the fiber optic source, but it's something to consider.  Best of luck with your project!

 

Taylor

  • 4 weeks later...
Posted

Late to the party, but - we make aligning the images in Photoshop part of our standard processing. The procedure we use is as follows (Photoshop CS6):

 

  • In Bridge select all the images
  • In Bridge, open the image files in Camera Raw and apply any bulk processing needed (don't crop)
  • Export them to Photoshop (Tools | Photoshop| Load files into Photoshop layers)
  • In Photoshop select all layers
  • In Photoshop Edit | Auto-Align Layers
  • Crop the resulting image (so edges will align when exporting)
  • Export the processed layers as JPEGs (File | Scripts | Export Layers To Files...), quality 12

This lets us deal with even the least stable supports as long as the essentials are in all the images (complete subject with a reasonable margin for the auto-align process, and the targets).

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...