Duckware
 You are here: Duckware » Technology Blog » 1080p vs 2K, 4K, 5K, 8K   Contact us  
1080p vs 2K, 4K, 5K, 8K
Expected quality differences between video resolutions

1. Video Resolutions
2. Ring Floodlight Pro Case Study
3. Ring Floodlight Pro Interactive Test
4. Video Quality Calculator / Bits per Pixel

 1. Video Resolutions

Background: The most common HD video resolutions today are:
  • 720p refers to a one megapixel 1280x720 1280 horizontal, 720 vertical HD video.
  • "Full HD" / FHD / 1080p (wikipedia.org) refers to a 1920×1080, two megapixel video.
  • 2K video has around 2× the pixels as FHD This results in a horizontal resolution ≈1.4× that of FHD. For example, 2560×1440, and 2688×1520 4 megapixels, 2× FHD.
  • 4K video (wikipedia.org) has a horizontal resolution ≈4000 pixels 2× FHD. For example, 3840×2160 8 megapixels, 4× FHD, or 2160p.
  • 5K video (wikipedia.org) has a horizontal resolution ≈5000 pixels 2.7× FHD. For example, 5120×2880 15 megapixels, 7× FHD.
  • 8K video (wikipedia.org) has a horizontal resolution ≈8000 pixels 4× FHD. For example, 7680×4320 33 megapixels, 16× FHD.

Relative video frame sizes (16:9 ratio)

Video/Photo resolutions are written as "H×V", or (a) the number of horizontal H pixels, (b) the multiplication sign, and (c) the number of vertical V pixels. For example, 1920×1080 means 1920 horizontal pixels and 1080 vertical pixels.

Why the "p": Why the 'i' in 1080i and why the 'p' in 720p and 1080p? 'i' means interlaced every frame refreshes only half the rows and 'p' means progressive every frame refreshes all rows. Interlaced video is now an historical artifact and everything produced for online streaming is progressive.

Resolution vs Megapixels: When resolution doubles in both horizontal and vertical directions, megapixels quadruples.
If you have a "H×V" resolution video/photo and increase both horizontal 'H' and vertical 'V' resolutions by a factor 'F', the total number of pixels increases by the square of 'F' -- as the new number of pixels is: (H×F)×(V×F), or H×V×F²

An interesting observation about resolution numbers -- they are always evenly divisible by '2' quite a few times. For example, the 4K '3840' horizontal resolution is 2×2×2×2×2×2×2×2×3×5.
Expected video quality: So the obvious question is: How much better is 2K, 4K, 5K, and 8K over Full HD? Well, the key factors that determine video quality are and improving any factor, or not, is all about $$$:
  1. resolution how many pixels are there horizontal and vertical
  2. compression what image compression quality setting was used
  3. optics how good/bad are the optics
  4. scene complexity how complex is the scene being videoed + how much motion
  5. focal length how wide or zoomed-in
  6. codec AVC1 H.264 vs HEVC H.265
1) "Resolution": See for yourself the impact that resolution has on quality. There is a very noticeable increase in quality every time there is an increase in resolution:
"Resolutions"
1080p 2K 4K 8K
1080p 2K 4K 8K
NOTE: This illustration intentionally EXCLUDES the impact that video compression and other factors, like optics design will have on video quality. The comparison is designed to demonstrate the 'best case' outcome resulting ONLY from changes in resolution.
Just because some company claims to have a higher 'K' video camera does NOT necessarily make that camera 'better' than lower 'K' cameras. It may, but it may not. Resolution is just one of many factors that ultimately determines video quality!
For example, I feel that too many security cameras are compressing the video way too much -- to save on storage. And all that does is reduce video quality, which some people really want/need.
2) "Compression": MP4 videos effectively use JPEG compression for 'key frames' and the impact of 'compression' on JPEG photos is well documented (panohelp.com) -- high compression can often result in photos/video that (a) looks 'fuzzy', 'blurry', or 'grainy' license plate below right, (b) have lost fine detail, and (c) contain visible 'blocking' gray fender left of license plate.

"Compression"
Interactive: Flip the compression photos above to better 'see' the changes.
3) "Optics": It is virtually impossible for the average person to independently verify optics quality, but high-quality optics vs bad optics can make a huge difference in the resulting video quality. Often times, just 'pixel peep' and use own eyes to decide.
This seems like a replay of the 'megapixels race' in digital cameras, where a 1.5 megapixel camera could counterintuitively take MUCH higher quality photos than a 4 megapixel camera -- due to much better optics in the 1.5 MP camera. By 'pixel peeping', the optics quality difference becomes very obvious crisp vs blurry:

"Optics"
Good 1.5-MP camera Bad 4-MP camera
Source: The Megapixels Myth (panohelp.com)
4) "Scene Complexity": How complex is the scene? How much motion is there? What is actually being videoed is a huge wildcard, as it will be different for every person testing. Scene complexity directly impacts how well compression works or not, which in turn impacts video quality. A scene of low complexity will appear like it is higher quality some areas don't need as many bits than a scene of higher complexity not enough bits to render fine detail all over.

5) "Focal Length": How 'wide' or 'zoomed-in' a video camera is can have a major perceived impact on video quality. It all comes down to how many pixels there are per unit of horizonal distance In general, the larger that number, the higher the perceived quality will be.
Replacing a wide angle floodlight camera with a narrow angle floodlight camera will result in higher quality images a license plate that is 30 pixels under a 'wide angle' camera and barely readable may be 45 pixels under a 'narrow angle' camera and more readable. The tradeoff for that clearer picture is less field of view.
5) "codec": Which 'codec' is used can have a major impact on video quality. Every device supports H.264 (AVC1), but the newer H.265 (HEVC) can produce the same video quality at around half the bitrate. So if bitrate is kept the same and the codec is switched from AVC1 to HEVC, video quality will improve.


 2. Ring Floodlight Pro '1080p' vs '2K' case study

Ring Floodlight Pro
Background: In January 2025, Ring upgraded (ring.com) their Floodlight Pro (amazon.com) outdoor floodlight camera from 1080p to 2K via a firmware update. When Ring released this camera in 2021, it used a 2K sensor, but only provided Full HD video. Ring is now allowing 2K video.

How to change 1080p to 2K: In the Ring app, select 1080p/2K at "Device Settings > Video Settings > Video Quality".

Pixel Peep: To compare 1080p to 2K, let's perform some extreme 'pixel peeping' blue arrow below in video frames, looking at a license plate:
Actual Test Results: Below are actual results when I selected '1080p' and '2K' for 'video quality' within two minutes of each other for this camera in the Ring App:
Ring Floodlight Pro
1080p 2K
Ring Floodlight Pro 1080p Ring Floodlight Pro 2K
2K Jaggies
Ring Floodlight Pro 2K, 2.6 Mbps
Bad 'jaggies' on person walking
Test Files: The Ring Floodlight Pro 'pixel peeping' captures immediately above were pulled from the following full-sized video frames if YOU want to pixel peep. These are PNG files (wikipedia.org), so that no pixels are lost/changed in what I saw on my screen vs what you will see: Sharpness: Frankly, in the Ring Floodlight Pro, it is very hard to notice any significant difference in video quality between '1080p' and '2K'. There IS an improvement in 'sharpness', but it is not obvious that 2K reveals any more detail, which the comparison between video resolutions far above shows that there should be.
Is 2K sometimes actually worse to 1080p? Take a look at the jaggies on the person walking photo, right.
Conclusion: On the Ring Floodlight Pro, do NOT expect an unreadable license plate under 1080p to become readable under 2K. And with 2K, there is an increase in 'sharpness' in the overall image, but expect more 'jaggies' on moving objects (see example right).

And I have directly tested this on a Ring Floodlight Pro by turning on 1080p, running a liveview, and driving the car away until the license plate can just barely no longer be read. And then switching the Ring Floodlight Pro to 2K and starting another liveview. The license plate should look better, right? But it doesn't -- it is still unreadable. But it should be readable due to the 1080p to 2K resolution improvement. This confirms that there is something else on the Ring Floodlight Pro other than resolution contributing to that unreadability compression, optics, or encoder settings, something else.

Reproducible: I then replicated this case study at a different house, in a different State, using a different Internet provider, using a different Ring Floodlight Pro camera -- and obtained the exact same results. Refer to interactive test immediately below.


 3. Ring Floodlight Pro Interactive Test

An 'interactive' Ring Floodlight Pro '1080p' vs '2K' comparison test: Using a different Ring Floodlight Pro than the case study above, comparing 1080p to 2K, the snaphots below are 1:1 PNG pixels grabs from the Ring videos so no pixels lost from from I saw vs what you are seeing below, taken within one minute of each other at night, with floodlight lights ON, and confirm the case study results.

See for yourself by using this interactive test. Order is randomized, and can be randomized again by clicking on the 'Randomize' link in the table. Then, the 'Show Labels' link will identify the photos:
Randomize -- Ring Floodlight Pro -- Show Labels
View thermometer used (on amazon.com)

TIP: Right click on the images above and select 'Open image in new tab' to see and play around with the raw PNG. Open as three tabs in one browser window and use ctrl-tab to flip between the PNG.

There was a noticable brightness difference between 1080p and 2K which I equalized so that you could run the interactive test above multiple times -- otherwise, after the first try, the brightness, and not quality, would reveal which was 1080p and which was 2K.
Oversharpen
Oversharpening?
Is Ring's "2K" overly sharpened? Pixel peeping at the Ring Floodlight Pro 2K video showed classic signs of oversharpening see right; a whiter than white background 'halo' around black lettering. So I added a separate "sharpened 1080p" image to the interactive test agove with similar sharpening applied, which now makes 'sharpened 1080p' very hard to tell apart from 2K.

Test setup: Tests were conducted with the Ring Floodlight Pro camera running firmware 18.0.13 visible under camera settings by signing into ring.com. There was no movement in video frame at all so a completely static scene. The camera's 1080p video was 1.5 Mbps at 1920x1080. The camera's 2K video was 2.5 Mbps at 2688x1520. Both videos were 24 fps, with a 48 frame or 2 second GOP, and H.265 HEVC encoded.


 4. Video Quality Calculator / Bits Per Pixel

MP4 videos use lossy compression to dramatically reduce file size. The higher the compression ratio permitted, the smaller the resulting MP4 files size is, and the fewer bits of information there are to re-create the original video. So video quality decreases the more a video file is compressed, as more and more data is 'thown away'.

Objective: One of the most basic ways to objectively compare a video to itself, but with different encoder settings, is to compute "Bits Per Pixel", or on average how many bits of information are used to represent a single pixel in a video.
But a warning, a tiny jump in BPP technically has 'better' video quality, but it may be very hard, if not impossible, to visibly see tiny quality differences until there is a very significant 'jump' in BPP.
Bits Per Pixel Formula: The BPP formula is just the "encoder_bits_per_sec" divided by total pixels in one second, which is "frame_width × frame_height × frames_per_sec":
Bits per Pixel
However, BPP is only one tool for evaluating video quality. Another major factor is GOP size.

GOP: A GOP (wikipedia.org) is a 'group of pictures', or a collection of frames, and is often two seconds long. A GOP starts with an I-frame a whole frame snapshot, and all other frames in the GOP are p-frames, or b-frames changes since other frames. I-frames are dramatically larger than 'p' or 'b' frames.

The larger a GOP is with all other encoder factors the same:
  • The fewer larger I-frames there are -- replaced by smaller p/b frames.
  • The better video quality can be because fewer bits are wasted on I-frames.
  • But the slightly longer time it takes for random seeking during video playback on client devices, as there is only an I-frame at the start of each GOP. However, with the incredible speed of modern client devices, GOP size is now much less of a concern than it was even just five years ago.
Interactive Video Quality Calculator: Selecting a 'good' BPP and GOP size is all about balancing opposing factors.
Video Quality Calculator
Encoder Bitrate: Mbps [0-20]
Resolution: x (H×V)
Frames: per second [1-100]
GOP size: seconds [1-10]
I-factor: % [10-90]
 



In general, the higher the 'BPP', the better video quality will be. AND, the higher the 'GOP bytes', the better video quality will be. But this needs to be balanced against the desire to have a small GOP size to improve client device video playback performance.

Also, the higher the 'frames per second' with smoother the video will appear, but this comes at the expense of a larger video size.

Doubling 'GOP time' Illustration: Below is an illustration of a one second GOP at 20 fps, where the GOP is changed to 2 seconds. One of the I-frames is replaced by a p-frame, and the extra bytes hilighted in red must then be redistributed to all other other frames in the new two second GOP, slightly increasing the size and quality of all other frames.
It is worth noting that even though the average bitrate for the entire video remains the same, that the byte size for ALL frames increases, except the red I-frames converted to p-frames.


Copyright © 2025 Duckware