Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
02-19-2012, 08:21 AM   #91
Inactive Account




Join Date: Dec 2008
Location: Ames, Iowa, USA
Photos: Albums
Posts: 2,965
QuoteOriginally posted by Stone G. Quote
...
But this time, I have made some handheld shots in Manual Mode (with all the extra uncertainties that this will involve). Images to the left are the full, un-cropped frames without and with TC. Images to the right are 200% and 100% crops.ots.

By and large I think this confirms that each and every TC may not be that bad with all lenses and in all situations after all. One might also note that if I were to use extension tubes, I would have had to get much closer and couldn't have used the built-in flash.
Thanks for your quantitative work; You are sounding a lot like me in concluding "Sometimes TCs aren't too awful on a lens."

It seems to me that I'm thinking mathematically about resolution at the 10% level (ie. high frequency limit), while making experimental decisions - deciding which photo looks better - at about the 50% brightness level - far from diffraction limits. Maybe there are cases in this range of perception where the TC logically beats the enlarged no TC case?

02-19-2012, 10:24 AM   #92
Veteran Member




Join Date: Mar 2010
Location: NewYork
Posts: 899
I originally didn't post these as it seems that fotki compressed images for viewing which seemed to wipe out the difference. I just realized that there is a download original option that seems to show them right. just view a single image and then look for the little down point arrow icon above the picture to view the original. Also, sorry for the under exposure (I didn't recall them being so bad but I may have been more concerned about the compression). I had to work double shifts the last 2 days so I didn't have much time to do something with it. I may have time to take better pictures tonight. Regardless of the crappy photos, I think they still clearly show that even cheap glass looks better than cropping on a low resolution photo (the issue with the crop defiantly seems to be a resolution issue to me). I took the pictures in raw, converted to best quality jpeg in pentax laboratory, and cropped in microsoft paint (not the best procedures but the same was done to both).
If any one disagrees about the teleconverter looking better, let me know why (like I said, I have had little time to look at it). Maybe I could try an outdoor shot today if there is enough sunlight (overcast winter weather).
http://public.fotki.com/ripit2/teleconverter-vs-cr/
02-19-2012, 10:40 AM   #93
Pentaxian
Greyser's Avatar

Join Date: Oct 2011
Location: Los Angeles, California
Photos: Gallery | Albums
Posts: 1,072
QuoteOriginally posted by Stone G. Quote
By and large I think this confirms that each and every TC may not be that bad with all lenses and in all situations after all. One might also note that if I were to use extension tubes, I would have had to get much closer and couldn't have used the built-in flash.
Not bad at all. I may give it a try: to use TC with shorter lenses

QuoteOriginally posted by Stone G. Quote
Images to the left are the full, un-cropped frames without and with TC. Images to the right are 200% and 100% crops.
I'm not sure what 200% crop means
02-19-2012, 11:15 AM   #94
Site Supporter
Stone G.'s Avatar

Join Date: Jun 2010
Location: North Zealand, Denmark
Photos: Gallery | Albums
Posts: 1,508
QuoteOriginally posted by Greyser Quote
.......

I'm not sure what 200% crop means

Thanks Greyser,

It's a 100% crop (in this case 200 x 133 pixels) resampled by a linear factor of 2 (to 400 x 266 pixels). Thus, this provides the same field of view (when viewed on screen) as one gets with the TC for the same area of the object.

02-19-2012, 11:24 AM   #95
Veteran Member




Join Date: Mar 2010
Location: NewYork
Posts: 899
I discovered that when viewing on a computer when not logged in, I couldn't download them. Here are 2 with a ricoh 50mm 2.0 just shot out door and a vivitar 4 element teleconverter. shot at f5.6 and focus on the front wheel. Warning, these are the original raw files so they are big. Do what you like with them.

Wikisend: free file sharing service
Wikisend: free file sharing service

These were shot at f2.0 I have no idea what happened (flair?) but the teleconverter shots are terrible. Contrast looks blown bad and focus looks bad but I took 3 shots refocusing each time, and this was the best. It didn't have these problems indoor. I may be able to re shoot later after the wife gets home (gotta watch the kids right now).
Wikisend: free file sharing service
Wikisend: free file sharing service

If nothing else, the f5.6 shots seem to indicate a teleconverter shot can look better on a lower MP camera.
02-19-2012, 11:26 AM   #96
Pentaxian
Greyser's Avatar

Join Date: Oct 2011
Location: Los Angeles, California
Photos: Gallery | Albums
Posts: 1,072
QuoteOriginally posted by Stone G. Quote
Thanks Greyser,

It's a 100% crop (in this case 200 x 133 pixels) resampled by a linear factor of 2 (to 400 x 266 pixels). Thus, this provides the same field of view (when viewed on screen) as one gets with the TC for the same area of the object.
Are you talking about 2x up-conversion? what is the meaning for that?
02-19-2012, 11:30 AM   #97
Site Supporter
Stone G.'s Avatar

Join Date: Jun 2010
Location: North Zealand, Denmark
Photos: Gallery | Albums
Posts: 1,508
QuoteOriginally posted by Greyser Quote
Are you talking about 2x up-conversion? what is the meaning for that?
To get identical fields of view ("same image size") for comparison with the TC images.

See my previous post where both resized and non-resized non-TC images are shown side by side with the TC images.
02-19-2012, 11:32 AM   #98
Inactive Account




Join Date: Dec 2008
Location: Ames, Iowa, USA
Photos: Albums
Posts: 2,965
QuoteOriginally posted by Stone G. Quote
Thanks Greyser,

It's a 100% crop (in this case 200 x 133 pixels) resampled by a linear factor of 2 (to 400 x 266 pixels). Thus, this provides the same field of view (when viewed on screen) as one gets with the TC for the same area of the object.
Do you mean that the 100% crop was 200 sensor pixels x 133 sensor pixels wide??

That is, for your 100% display, each display pixel corresponds to one sensor pixel?

Dave

02-19-2012, 11:40 AM   #99
Site Supporter
Stone G.'s Avatar

Join Date: Jun 2010
Location: North Zealand, Denmark
Photos: Gallery | Albums
Posts: 1,508
QuoteOriginally posted by newarts Quote
Do you mean that the 100% crop was 200 sensor pixels x 133 sensor pixels wide??

That is, for your 100% display, each display pixel corresponds to one sensor pixel?

Dave
We are talking about the crops from the flower (crocus image) right?

The crop of the image taken with TC was 400 x 266 sensor pixels - and shown here pixel-to-pixel on screen, (first image below)

The crop on the non-TC image was 200 x 133 sensor pixels, (second image below)

That crop was resized in linear scale by a factor of 2 for comparison with the TC image to 400 x 266 pixels, which is what you see in the post (third image below).
Attached Images
View Picture EXIF
PENTAX K-5  Photo 
View Picture EXIF
PENTAX K-5  Photo 
View Picture EXIF
PENTAX K-5  Photo 
02-19-2012, 12:02 PM   #100
Veteran Member
demp10's Avatar

Join Date: Jun 2011
Location: Atlanta
Photos: Albums
Posts: 602
Here is another attempt to put things in perspective using simple math.
  • We have a camera with a 3,000 x 2,000 sensor (6MP)
  • We have a display with 1,500 x 1,000 pixels (1.5 MP)
  • We want a subject of interest to fill the vertical dimension of the screen so we need 1,000 pixels for the subject.
To display a full frame from this camera to this display we must reduce the image by a factor of 2 to fit on the screen. Essentially we take a 2x2 matrix from the original image and generate 1 pixel to be displayed.

We use the longest telephoto we have to photograph the subject from the closes distance we can get to it so we can get the largest possible image on the sensor.

Case 1: The subject registers with more than 1,000 vertical pixels on the sensor (let’s assume 1,500 pixels). Since the goal was to display the subject and fill the screen, we can crop the image by a factor of 1.33 and get a resulting image of 2,250 x 1,500. This image is still larger than the actual screen and needs to be reduced further for display purposes by a factor of 1.5. We are still using original pixels from the frame (as opposed to “manufactured” by digital up-sampling). Technically we are removing information from the original image (by down-sampling it for the display) but we are not introducing artifacts from generating new pixels in software.

This will hold true as long as the subjects’ size on the sensor is at least 1,000 pixels using this display and camera combinations. If we use a camera with more megapixels, we can crop proportionally more and still get the 1,000 pixels for the display. On the other hand, if we use a higher resolution display then we will need to crop less to get the vertical pixels required to fill the new display.

Case 2: The subject registers less than 1,000 vertical pixels on the sensor (let’s assume 500 pixels). Since the goal was to display the subject and fill the screen, we must crop the image by a factor of 4 and get a resulting image of 750 x 500. This image is smaller than the actual screen and needs to be enlarged for display purposes by a factor of 2. Essentially for every real image pixel we will manufacture a matrix of 2x2 new pixels for the display. No matter what re-sampling algorithm we will use, there will always be artifacts that will degrade the image quality to some degree.

Now, if we add a 2X TC to the lens, we will be able to magnify the image optically and get (in the above example) 1,000 vertical pixels for the subject. We still need to crop the image but the resulting crop will be 1,500 x 1,000 requiring no further enlargement for the display.

The real question now is which is better: a digitally enlarged original image or an optically enlarged one? Keeping in mind that we are still using the same lens, the images to start with are the same; we simply enlarge one digitally and the other optically. Unless the TC is of terrible quality, an optical enlargement will be better most of the times.

To put it another way, if after the crop we can end up with an image with sufficient resolution to fulfill our display (or print) needs then no TC is required. If the resulting image is much smaller, then a good TC will probably give better results.
02-19-2012, 12:21 PM   #101
Veteran Member




Join Date: Mar 2010
Location: NewYork
Posts: 899
It seems to me after seeing others examples (in particular with equipment a little too pricey for me to try myself) as well as my own very limited experience with rear mount teleconverters, and greater experience with from mount ones, asking if teleconverters are worth using is like saying are lenses worth using. Lenses can good or bad or somewhere inbetween. How well the lens performs may also be impacted by how well they suit the job. Even further, a lens might perform great stopped down a little (teleconverters seem to have a bit of that but so do some lenses), so how well they do might also be effected by available light. There are endless variables for instance, you might say the teleconverter shots will always be inferior at least by a little bit. Zoom lenses in general are considered at least a little inferior to primes overall, but use of zooms is widely accepted (and improves the circumstances you can shoot with limited equipment or budget just like a teleconverter can). How many primes does it take to replace a 55-300mm, how expensive would they be, and how big of a pack would it take to carry them all? How big and heavy are the lenses that would render teleconverters useless and how expensive are they? Seems all the pluses and minuses with teleconverters are many of the same pluses and minuses with lenses in general. The only exception that comes to mind, is that a pentax mount lens will generally preform up to it abilities on a pentax camera. Teleconverters seem to be a lot more finicky. A pentax mount teleconverter will not perform up to its abilities on all lenses, but only on some or perhaps even a few. How well the teleconverter matches the lens seems to be the only major limitation specific to teleconverters over other lenses.
02-19-2012, 01:09 PM   #102
JVi
Senior Member




Join Date: Mar 2011
Location: Ontario
Posts: 105
I enjoy using a Promaster Spectrum 7 1.7x (Z-AF) with my DA 300/4 wide open.
This was taken today from 30 feet on a Kx.
Attached Images
 
02-19-2012, 03:29 PM   #103
Inactive Account




Join Date: Dec 2008
Location: Ames, Iowa, USA
Photos: Albums
Posts: 2,965
QuoteOriginally posted by Stone G. Quote
We are talking about the crops from the flower (crocus image) right?

The crop of the image taken with TC was 400 x 266 sensor pixels - and shown here pixel-to-pixel on screen, (first image below)

The crop on the non-TC image was 200 x 133 sensor pixels, (second image below)

That crop was resized in linear scale by a factor of 2 for comparison with the TC image to 400 x 266 pixels, which is what you see in the post (third image below).
Thank you Stone. That's what I thought. Doesn't it seem unfair to push the without TC image beyond 100%?

Are the results relatively the same if you don't crop the TC version at all, just enlarge the the w/o TC by a factor of 2?

It seems that in your case the TC consistently outperforms the nonTC/enlarged image but this might be because you are pushing the nonTC beyond 100%.

Dave
02-19-2012, 03:32 PM   #104
Inactive Account




Join Date: Dec 2008
Location: Ames, Iowa, USA
Photos: Albums
Posts: 2,965
QuoteOriginally posted by Stone G. Quote
We are talking about the crops from the flower (crocus image) right?

The crop of the image taken with TC was 400 x 266 sensor pixels - and shown here pixel-to-pixel on screen, (first image below)

The crop on the non-TC image was 200 x 133 sensor pixels, (second image below)

That crop was resized in linear scale by a factor of 2 for comparison with the TC image to 400 x 266 pixels, which is what you see in the post (third image below).
Thank you Stone. That's what I thought. Doesn't it seem unfair to push the without TC image beyond 100%?

Are the results relatively the same if you don't crop the TC version at all, just enlarge the w/o TC by a factor of 2?

It seems that in your case the TC consistently outperforms the nonTC/enlarged image but this might be because you are pushing the nonTC beyond 100%.

Dave
02-19-2012, 05:38 PM   #105
Site Supporter
Stone G.'s Avatar

Join Date: Jun 2010
Location: North Zealand, Denmark
Photos: Gallery | Albums
Posts: 1,508
QuoteOriginally posted by newarts Quote
Thank you Stone. That's what I thought. Doesn't it seem unfair to push the without TC image beyond 100%?

Are the results relatively the same if you don't crop the TC version at all, just enlarge the w/o TC by a factor of 2?

It seems that in your case the TC consistently outperforms the nonTC/enlarged image but this might be because you are pushing the nonTC beyond 100%.

Dave
Thanks for comments Dave,

Surely, the "fairness" of my tests can be subject to discussion. But in my opinion the cropping part of the discussion boils down to what one wants to show on screen / print - reference also demp10's post above.

If we take the flower (or stamp for that matter), one thing is if I want to show the flower (stamp) as a whole or wheter I want to focus on the finest detail I can capture such as the stamen of the crocus or the seeping of the printer's ink on the stamp. What I have tried to illustrate is that fine detail may well be captured better by optical enlargement than by digital (again ref demp10 above) of the image captured by a sensor that is very far from being unlimited in resolving power/sampling. I think there is little doubt that I can see finer detail with TC than without in the specific setup's and situations that I have described.

Another very relevant (for me) example is astrophotography. Take planet Jupiter who's diameter is a mere 49 arc seconds. With my small 1000mm FL telescope that will give me an image some 40 pixels wide on my K200D sensor and some 50 pixels on my K-5 sensor. Not an awful lot to show the cloud belts on Jupiter! If I add my 2.5X apochromatic Barlow Lens (which is effectively the same as a TC) the situation improves and the Jupiter image will be 100 pixels / 125 pixels in diameter. Now that means that the images will occupy 10,000 pixels on my K200D sensor and 15,625 pixels on my K-5 sensor ----- and the remaining 10 million / 16 million dark and dull pixels I shall have to throw away if I want to to show you a reasonably decent image of Jupiter:


100% crop of a 10 million pixel image. Image taken with K200D
on a Vixen Polaris R-100L Newtonian at an effective focal
length of 2,500 mm

I dont want to draw any square conclusions about the ever-glorious benefits of TCs, but with my tests and my description of my setups and the situations in which the TC was used, I hope I can help others in their own judgement: "Are teleconverters worth using?"

Last edited by Stone G.; 02-19-2012 at 05:44 PM.
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
auto, k-mount, pentax lens, reviews, slr lens
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
Pentax SDM lenses, how much they are really worth or are they worth it? Pentaxor Pentax SLR Lens Discussion 20 01-17-2015 11:32 PM
2x teleconverters slr Pentax SLR Lens Discussion 5 08-20-2011 11:31 AM
teleconverters ACHEBONE Ask B&H Photo! 10 10-30-2010 05:46 AM
Teleconverters GingeM Pentax SLR Lens Discussion 7 03-29-2010 04:20 AM



All times are GMT -7. The time now is 07:07 PM. | See also: NikonForums.com, part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top