Viewing image on TV/monitor/phone etc

[Chefgage]

Let's say after finishing processing an image it's resolution is 4704 x 3360 and then I transfer it to my server to be viewed on my TV. My TV is 1920 x 1080 so when I view the image it is full screen but I assume is now downsampled to 1920 x 1080.  As I understand it there are various methods of downsampling, some better than others.

So my questions are, is it better to just resize my image during processing to 1920 x 1080 so that no downsampling takes place and possibly effects the image?

Also is it better to resize the image or bin the image?. If I need a specific size I would have to resize but which is best in terms of picture quality?

 

 

[fifeskies]

I would think it is better to use the power of the processing software to produce the lower resolution image as this will give predictable results.

I usually save my images (18Mb) as the full image and also a 1920 x 1080 (which is about 2Mb) for convenient (ie TV) viewing generated from the final result.

Also often a "thumbnail" size of around 1000 wide which is fine for email type sharing in messages.

[Chefgage]

1 hour ago, fifeskies said:

I would think it is better to use the power of the processing software to produce the lower resolution image as this will give predictable results.

I usually save my images (18Mb) as the full image and also a 1920 x 1080 (which is about 2Mb) for convenient (ie TV) viewing generated from the final result.

Also often a "thumbnail" size of around 1000 wide which is fine for email type sharing in messages.

Thanks for that. I think I will give that a try later and compare the processed reduced size image with the image that the TV downsamples.

Obviously to get the 1920 x 1080 I will just resize the image in pixInsight. But still wondering about binning and how that differs from just resizing?

Edited June 25, 2021 by Chefgage

[Martin Meredith]

Binning is one way to resize, but is generally suboptimal compared to the interpolation used in most resizing algorithms. Binning is convenient for sensors as it doesn't involve anything very complicated.

In technical terms, resizing downwards (downsampling) involves losing resolution i.e. high spatial frequencies. If you just bin, then the energy represented by those high spatial frequencies is still present in the image, but is spread around as noise. Proper downsizing ensures that the high spatial frequencies are attenuated first, so the image is just a little bit less noisy. Most likely it would be hard to spot much difference except in some edge cases, but since most resizing software already takes care of this, I wouldn't recommended binning.

cheers

Martin

[Chefgage]

4 hours ago, Martin Meredith said:

Binning is one way to resize, but is generally suboptimal compared to the interpolation used in most resizing algorithms. Binning is convenient for sensors as it doesn't involve anything very complicated.

In technical terms, resizing downwards (downsampling) involves losing resolution i.e. high spatial frequencies. If you just bin, then the energy represented by those high spatial frequencies is still present in the image, but is spread around as noise. Proper downsizing ensures that the high spatial frequencies are attenuated first, so the image is just a little bit less noisy. Most likely it would be hard to spot much difference except in some edge cases, but since most resizing software already takes care of this, I wouldn't recommended binning.

cheers

Martin

Thank you for that.