• Global community
    • Language:
      • Deutsch
      • English
      • Español
      • Français
      • Português
  • 日本語コミュニティ
    Dedicated community for Japanese speakers
  • 한국 커뮤니티
    Dedicated community for Korean speakers
Exit
4

Downscaling: TV processor vs. Photoshop

Engaged ,
Jan 31, 2024 Jan 31, 2024

Copy link to clipboard

Copied

  • Background: A different client will be using Sony A95L displays to present images taken by DSLR.
    The DSLR has a resolution of 30.1MP: 6720 x 4480.  This is 3.6x larger than the display’s 4k resolution: 3840 x 2160 = 8.2MP.  Therefore, the display’s processor will downscale the images.

 

Question #1:
Can anyone describe the specific issues that will be created by the TV’s processor downscaling?

From the articles the images can have:

  • Aliasing (jagged diagonal edges, termed "jaggies."  Anti-aliasing smooths the edges)
  • Edge halos,
  • Blurring

The images will appear “Muddy” (scientific term) as a result of a loss of sharpness.

Will the images suffer a loss of details?

Will shadows and highlights degrade?

What other items would you add?

 

 

Question #2: Depends on the Image:
I know many of you will say: "It depends on each image”.  “If resizing is artifact-free, you may not need to change a thing; photographic workflows can become complicated enough as is. Many photos do not have detail which is susceptible to moiré: regardless of the interpolation.”.  However, if you were processing 250,000 images, how would you approach the issue of which algorithm to use?

 

  • Photoshop Workflow:
    To prevent the TV manufacturer’s processor from downscaling, I can resize the images in Photoshop using image/image size and change the height and width to 3840 x 2160 (to match the display).
    Then chose Photoshop’s Bicubic Sharper (reduction) Algorithm

 

Question #3: Other articles recommended using Photoshop’s “Preserve Details 2.0

Does anyone have experience using one versus the other?
The goal is to minimize aliasing artifacts: moiré from downscaling or reducing the images.

 

Question #4:

  • Sharpening: After resizing, I assume you would recommend sharpening the images.
  • Denoise: Is any denoise required to offset the downscaling?
    Some online resources said this was not necessary because SNR was improved as a result of downscaling.  However, that does not make sense to me.
  • Other: What other adjustments would you recommend making?
  • Goal: The goal is to have an artifact-free sharpness

 

The following are the key points from the two articles:

 

  • Three Undesirable Artifacts: Edge halos, Blurring, and Aliasing (jagged diagonal edges, termed "jaggies.". Anti-aliasing smooths the edges)
  • “Even the most advanced non-adaptive interpolators always have to increase or decrease one of the above artifacts at the expense of the other two: therefore at least one will be visible.”
  • Loss of Data: The 4k dimensions result in a loss of 72% of the original data (8.2MP vs. 30.1MP).  “It is impossible to show detail which previously had a resolution of just a single pixel. If any detail is shown, this is not real and must be an artifact of the interpolator.”   “An image will always lose some quality each time interpolation is performed.” “The more you know about the surrounding pixels, the better the interpolation will become. Therefore, results quickly deteriorate the more you stretch an image, and interpolation can never add detail to your image which is not already present.”
  • Textures: and moiré: Images with a lot of texture can cause problems from moiré. “Images with fine geometric patterns are at the highest risk; these include roof tiles, distant brick and woodwork, wire mesh fences”.  “Interpolators which show detail outside the limit are adding patterns to the image which are not actually there (moiré).”
  • Softening: Loss of Sharpness: “Interpolation algorithms which preserve the best sharpness are more susceptible to moiré, whereas those which avoid moiré typically produce a softer result. This is unfortunately an unavoidable trade-off in resizing.”
  • Summary of Different Algorithms:
    “Sinc and lanczos algorithms produce the best results; they are able to resolve detail all the way to the theoretical maximum, while still maintaining the fewest artifacts beyond. Photoshop bicubic comes in second, as it has visible moiré patterns far outside the box. Furthermore, note how bicubic also does not show as much detail and contrast just inside the red box.
    “If your image has moiré, (Photoshop’s) sharper setting will amplify and the smoother setting will reduce it (relative to default). This highlights a key divide: some interpolation algorithms are much better at increasing than decreasing image size, and vice versa.
  • Ideal solution: is to use a sinc or lanczos algorithm to avoid moiré artifacts in the downsized image, then follow-up with a very small radius (0.2-0.3) unsharp mask to correct for any interpolation-induced softening. On the other hand, the sinc algorithm is not widely supported and software which uses it is often not as user-friendly.”

 

Thank you for your time.

 

TOPICS
Windows

Views

90

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Adobe
Community Expert ,
Jan 31, 2024 Jan 31, 2024

Copy link to clipboard

Copied

If you're after a definite answer, test with multiple sample images on the device and viewing distance/conditions in final use.

 

If you're happy with "general" suggestions, then I personally would use Bicubic Smoother, then an appropriate post downsampling sharpening of your choice (there are so many methods and even more subjective opintions).

 

If you suspect that there is image content that will alias and show moiré or other weird results such as fine repeating patterns, then a 1-2px pre-blur before the resize can help avoid or limit issues in such areas. Sometimes you will need to recreate such elements in the resized version if they are critical and have suffered greatly from the resize. Also, one can experiment with different resizing techniques which may have better results than others for such elements.

 

Also adding some minor selective noise (generally not in the shadows, sometimes in the highlights) to areas that have "lost life" in the resampling can bring back the appearance of minor detail.

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Engaged ,
Jan 31, 2024 Jan 31, 2024

Copy link to clipboard

Copied

Stephen:

Thank you for your quick response.  This is exactly the kind of “real-world experience” I am looking for.  I am eager to hear from others in the community so I can further refine the workflow.  At the end of the day, I will arrange with a local retailer (such as Best Buy) to permit me to load a series of pictures into 4 of their displays side by side to make the final assessment.  However, I need to have processed a good number of images with the key options to make that time investment worthwhile.

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
LEGEND ,
Feb 01, 2024 Feb 01, 2024

Copy link to clipboard

Copied

LATEST

You are unlikely to see any difference from downsizing. Saving as JPEG vs a non-lossy format might make a small difference.

The biggest advantage to downsizing before you send them to the TV is file size, you can fit more small files on your media, they will transfer across a network faster, and they will load faster on the TV. Just batch process and downsize as your last step before display.

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines