Pages: 1 [2]   Go Down

Author Topic: Panoramic head vs. shift - question  (Read 6934 times)

shadowblade

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2839
Re: Panoramic head vs. shift - question
« Reply #20 on: April 09, 2015, 10:07:48 am »

The same stretching occurs in the corners of a lens projected image on a flat sensor, and the optical quality at the edge of the image circle is much lower. The only difference is that the sensor will capture the stretched (and lower lens resolution!) pixels accurate, and the rotated pano will have to interpolate its (much higher quality) pixels very accurately. The interpolation is rather accurate, because it's only a small stretch if you do not exceed what would be in the image circle of the shifted lens. Using a longer focal length for stitching will close the 'gap' even more.

Not entirely correct. The difference lies in quantisation.

With a tilt-shift lens, the stretching is carried out before the light hits the sensor. And it's not really distorting the light in any case, since it's a retrofocus lens - the light actually hits the sensor at angles similar to what comes out of a 50mm lens, not at extreme oblique angles. Therefore, when you shift to each end and stitch on a 36MP sensor, you end up with the full 60MP of detail

With a rotational method, you capture the image first, quantising it into pixels, prior to any stretching. When you stretch it in post-processing, you lose a lot of these pixels; the wider the virtual lens, the more you lose.

For the equivalent angle of view of the TS-E 24L (equivalent horizontal angle of view to a single shot from a 14mm lens on full frame) this stretching is considerable. For the equivalent of the TS-E 17L (equivalent horizontal angle of view to a 10mm lens on full-frame when stitched) it's even greater.

The only way around this with a rotational panorama is to shoot using a significantly longer lens and start with a higher-resolution image. Then, once you stretch it, you still retain the same effective resolution (pixels per degree) in the corners.
Logged

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8913
Re: Panoramic head vs. shift - question
« Reply #21 on: April 09, 2015, 02:39:37 pm »

Not entirely correct. The difference lies in quantisation.

With a tilt-shift lens, the stretching is carried out before the light hits the sensor. And it's not really distorting the light in any case, since it's a retrofocus lens - the light actually hits the sensor at angles similar to what comes out of a 50mm lens, not at extreme oblique angles. Therefore, when you shift to each end and stitch on a 36MP sensor, you end up with the full 60MP of detail

That's including vignetting and edge of image circle deterioration. I wouldn't call that 'full detail', but it does sample the light (with all its flaws) as it hits the sensor. The angle of incidence may be not extreme, due to the retro-focus optical design, but the incoming rays (entrance pupil) do cover a wide angle. Just look at the examples I posted, the extreme corners ('only' 106-107 degrees diagonal coverage) are 'smeared' and stretched and relatively underexposed on an ETTR exposure setting.

Quote
With a rotational method, you capture the image first, quantising it into pixels, prior to any stretching. When you stretch it in post-processing, you lose a lot of these pixels; the wider the virtual lens, the more you lose.

As explained, the pixels start at a much higher quality (especially with a proportionally (portrait orientation) longer focal length that can be used for the same or wider FOV) before resampling.

Quote
For the equivalent angle of view of the TS-E 24L (equivalent horizontal angle of view to a single shot from a 14mm lens on full frame) this stretching is considerable. For the equivalent of the TS-E 17L (equivalent horizontal angle of view to a 10mm lens on full-frame when stitched) it's even greater.

Only if the 'wrong' focal length is used in a direct comparison.

Here is an example, at too short a focal length for proper comparison (it should not be 24mm but 35mm or more) of the un-shifted center of image circle image quality with only basic Raw conversion and minor Capture sharpening without CA correction (click on the image for a full resolution version, full size=9,826,078 bytes!):

The resampling to a corner tile of a pano stitch will only stretch the pixels for the geometrical projection of such a wide angle on a flat plane. Alternative projection methods, requiring even less geometrical 'distortion', can also be used if the image content allows.

Quote
The only way around this with a rotational panorama is to shoot using a significantly longer lens and start with a higher-resolution image. Then, once you stretch it, you still retain the same effective resolution (pixels per degree) in the corners.

That's what I already explained earlier. It would be useful for a discussion to illustrate it with an example. I'm a bit busy right now, or otherwise I'd have done it already. Maybe you have something concrete to discuss (although we do not disagree about the principle, but maybe on the effective practical consequences)?

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==

BernardLanguillier

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 13983
    • http://www.flickr.com/photos/bernardlanguillier/sets/
Re: Panoramic head vs. shift - question
« Reply #22 on: April 09, 2015, 06:39:02 pm »

Not entirely correct. The difference lies in quantisation.

With a tilt-shift lens, the stretching is carried out before the light hits the sensor. And it's not really distorting the light in any case, since it's a retrofocus lens - the light actually hits the sensor at angles similar to what comes out of a 50mm lens, not at extreme oblique angles. Therefore, when you shift to each end and stitch on a 36MP sensor, you end up with the full 60MP of detail

With a rotational method, you capture the image first, quantising it into pixels, prior to any stretching. When you stretch it in post-processing, you lose a lot of these pixels; the wider the virtual lens, the more you lose.

For the equivalent angle of view of the TS-E 24L (equivalent horizontal angle of view to a single shot from a 14mm lens on full frame) this stretching is considerable. For the equivalent of the TS-E 17L (equivalent horizontal angle of view to a 10mm lens on full-frame when stitched) it's even greater.

The only way around this with a rotational panorama is to shoot using a significantly longer lens and start with a higher-resolution image. Then, once you stretch it, you still retain the same effective resolution (pixels per degree) in the corners.

Just out of curiosity, have you tried spherical stitching?

I understand your theoretical concerns, but they simply have never resulted in what I find to be any real issue in corners, or at least not even close to the poor corner quality of T/S lenses.

To my eyes both techniques have potential and using T/S can be the best option when both shift and tilt must be used/time is limited and limited resolutions are sufficient, but leaving this case aside I find spherical stitching to be the superior approach overall in real world applications. The good news is that there is no reason to commit to one approach only.

Cheers,
Bernard
« Last Edit: April 09, 2015, 07:04:11 pm by BernardLanguillier »
Logged

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8913
Re: Panoramic head vs. shift - question
« Reply #23 on: April 10, 2015, 05:32:07 am »

Just out of curiosity, have you tried spherical stitching?

I understand your theoretical concerns, but they simply have never resulted in what I find to be any real issue in corners, or at least not even close to the poor corner quality of T/S lenses.

To my eyes both techniques have potential and using T/S can be the best option when both shift and tilt must be used/time is limited and limited resolutions are sufficient, but leaving this case aside I find spherical stitching to be the superior approach overall in real world applications. The good news is that there is no reason to commit to one approach only.

I agree with Bernard, who like me is an experienced Pano shooter/stitcher. In fact I stitch unshifted (thus optimal center of image circle) but tilted images (thus with more DOF control with less diffraction blur) with my TS-E lenses. So not only do I not commit to one approach, I use both at the same time ...

I also use a deeper lens hood on my TS-E 24mm II, to reduce the potential for veiling glare, and because there is no shift involved it can be much deeper without risk of vignetting.

As far as the TS-E 24mm II goes, I think it is a very good lens for this type (only using tilt) of stitching for relatively wide (or high) vistas. It's better than a single TS-E 17mm shot could ever produce, and it comes with the benefits of stitching (such as no FOV restrictions, the subject may dictate the optimal crop, not the camera). The image quality in the un-shifted corners is excellent, as demonstrated in the (see attached) right hand side bottom corner crop of the image I posted earlier in reply #21.

This time I used a more recent version (V8.2 instead of V4.81) of Capture One Pro to convert the Raw, and I only added some sharpening (no specific detail enhancements or Clarity yet) to bring out the level of detail that's available.

Cheers,
Bart
« Last Edit: April 10, 2015, 05:57:49 am by BartvanderWolf »
Logged
== If you do what you did, you'll get what you got. ==

BernardLanguillier

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 13983
    • http://www.flickr.com/photos/bernardlanguillier/sets/
Re: Panoramic head vs. shift - question
« Reply #24 on: April 10, 2015, 06:28:40 am »

In fact I stitch unshifted (thus optimal center of image circle) but tilted images (thus with more DOF control with less diffraction blur) with my TS-E lenses. So not only do I not commit to one approach, I use both at the same time ...

Yep, same here. The following image was shot 3 years ago with the Nikon 24mm T/S tilted and is made up of 3-4 images shot on a spherical RRS head. This can sometimes cause a bit of a headache to the pano software though... :)



Cheers,
Bernard
« Last Edit: April 10, 2015, 06:31:27 am by BernardLanguillier »
Logged

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8913
Re: Panoramic head vs. shift - question
« Reply #25 on: April 10, 2015, 07:37:13 am »

This can sometimes cause a bit of a headache to the pano software though... :)

Quite so, but not so much for e.g. PTGUI, if one uses the correct interventions. A bit of horizontal/vertical offset and similar, after disabling the other optimization parameters that were first used, can do miracles. The mostly automatic stitchers, like Photoshop, may choke on such a task.

Your image is a good example that when relatively little light is available, one can often still work with relatively wide apertures (but with additional DOF provided by tilt) and shorter exposure times, thus reducing the risk of wind induced motion spoiling parts of the image.

Cheers,
Bart
« Last Edit: April 10, 2015, 09:43:50 am by BartvanderWolf »
Logged
== If you do what you did, you'll get what you got. ==

shadowblade

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2839
Re: Panoramic head vs. shift - question
« Reply #26 on: April 10, 2015, 09:04:40 am »

Just out of curiosity, have you tried spherical stitching?

I understand your theoretical concerns, but they simply have never resulted in what I find to be any real issue in corners, or at least not even close to the poor corner quality of T/S lenses.

To my eyes both techniques have potential and using T/S can be the best option when both shift and tilt must be used/time is limited and limited resolutions are sufficient, but leaving this case aside I find spherical stitching to be the superior approach overall in real world applications. The good news is that there is no reason to commit to one approach only.

Cheers,
Bernard

I regularly stitch using both methods - spherical for narrower fields of view that are too narrow for the TS-E 24L, shifted for anything wide enough to use the TS-E 24L or the TS-E 17L. At narrower angles, the spherical method works fine, because the narrower field of view means that you're not dealing with much distortion at the edges anyway. At wider angles, I find that the spherical method loses too much resolution at the edges due to distortion from stretching to a rectilinear projection (unless you start with significantly more megapixels than a shifted panorama) and the direction of noise/fine patterns can look weird.

Remember, at 60 degrees from the middle (for a 120 degree angle of view) you're stretching the edge pixels to double their original dimension, and the corners even more.

First one - spherical stitch using Sigma 120-300
Second one - spherical stitch using Canon 70-200
Third one - shifted stitch using Canon TS-E 24L
Fourth one - shifted stitch using Canon TS-E 17L
Logged

Jim Kasson

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2370
    • The Last Word
Re: Panoramic head vs. shift - question
« Reply #27 on: April 10, 2015, 10:44:56 am »


First one - spherical stitch using Sigma 120-300
Second one - spherical stitch using Canon 70-200
Third one - shifted stitch using Canon TS-E 24L
Fourth one - shifted stitch using Canon TS-E 17L

Really nice images!

Jim

BernardLanguillier

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 13983
    • http://www.flickr.com/photos/bernardlanguillier/sets/
Re: Panoramic head vs. shift - question
« Reply #28 on: April 10, 2015, 06:54:52 pm »

Really nice images!

Jim

Indeed!

Cheers,
Bernard

Torbjörn Tapani

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 319
Re:
« Reply #29 on: April 11, 2015, 04:44:50 pm »

Really nice. All of them!
Logged
Pages: 1 [2]   Go Up