Amazing and almost creepy 4k video on vimeo

[hawaiims]

Even if you don't have a 4k monitor you will definitely see the difference. Downscaled 4k still looks a lot better than a regular 1080p video. And this isn't your mother's 4k (by that I mean the crappy 4k on recent smartphones), this was shot with a Canon 1-DC (10000$ camera) 

 

It's really worth watching, and if you want the completely uncompressed version, you can download the 935 Megabyte 2-minute file (yup uncompressed 4k takes up that much space) on that same vimeo page. Even the compressed version still looks stunning though

 

 

Be sure to press the HD button!

[hawaiims]

Oh and here's a video from the same famous filmmaker Phillip Bloom using the Sony FS700 (aka the linuscam)

 

[TheNinjaNextDor]

That's cool, but why did he film that? 

[Zeal]

That was an amazing macro test footage! If only we could see such clarity up close!

And 4k on Vimeo>>>>4k on Youtube

[iruixos]

That's cool, but why did he film that? 

 

Exactly what I was thinking lol

 

Of all the things to film...

[hawaiims]

Exactly what I was thinking lol

 

Of all the things to film...

 

That's cool, but why did he film that? 

Here's a more conventional and IMO the best 4k test footage I've seen so far (at least on youtube)

 

Some of the shots look absolutely gorgeous

 

[TheNinjaNextDor]

Here's a more conventional and IMO the best 4k test footage I've seen so far (at least on youtube)

 

Some of the shots look absolutely gorgeous

 

 

Even on my Chromebook that looks AMAZING.

[Tea1337]

I need a 4k monitor now. Even at 1080p on this laptop, these videos look pretty damn good.

[LAwLz]

1) You're just experiencing the placebo effect. On a 1920x1080 monitor the 4K and 1080p version looks almost identical.
2) The video is not uncompressed. If it was truly uncompressed then it would be far more than just 1GB. something like 100GB would be more reasonable.
 
Here are some comparisons in case you don't think it's mostly placebo (when you hover over the image with your mouse it becomes the 1080p version, hold the mouse away and it's the 4K version):
Comparison 1
Comparison 2
Comparison 3
 
For those of you who are interested, I used Catmull-Rom with an anti-ringing filter for downscaling. I recommend you try it out yourself, just make sure you save the images as PNG and you get the exact same frame in both clips.
 
This just shows how good 1920x1080 can look. It's just that people are used to the awful quality on YouTube (or even worse, groups like YIFY and their awful 700MB releases).

[linusforsell]

Color depth and bitrate > supersampling. With that said though, can't wait to view 4k natively.

[Puppet]

1) You're just experiencing the placebo effect. On a 1920x1080 monitor the 4K and 1080p version looks almost identical.

2) The video is not uncompressed. If it was truly uncompressed then it would be far more than just 1GB. something like 100GB would be more reasonable.

 

Here are some comparisons in case you don't think it's mostly placebo (when you hover over the image with your mouse it becomes the 1080p version, hold the mouse away and it's the 4K version):

Comparison 1

Comparison 2

Comparison 3

 

For those of you who are interested, I used Catmull-Rom with an anti-ringing filter for downscaling. I recommend you try it out yourself, just make sure you save the images as PNG and you get the exact same frame in both clips.

 

This just shows how good 1920x1080 can look. It's just that people are used to the awful quality on YouTube (or even worse, groups like YIFY and their awful 700MB releases).

 

There is a difference in those comparisons, if you look at the highlights in the hair number 2, there's definitely a change. Not sure exactly whats going on, but the 1080p version looks over sharpened or something.

 

I don't think it's too detrimental to the image though and I don't really see the point in watching rendered 4k footage on a 1080p montior unless there is a bias in bit rate.

 

Watching footage shot in 4k and rendered in 1080p is so much better though, obviously. 

[LAwLz]

There is a difference in those comparisons, if you look at the highlights in the hair number 2, there's definitely a change. Not sure exactly whats going on, but the 1080p version looks over sharpened or something.

There are a few changes here and there. The difference is very minor though and I do not think it's enough to warrant a ~3 times as big file and much more work decoding it.

In comparison 1 you can see a tiny bit more noise in the outline of his head (just to the left of his eyebrow).

Comparison 2 is probably the most obvious one. The straws that are in focus are a bit sharper in the 4K version.

In comparison 3 you can see that the "valleys" in the skin (in the tiny strip of skin that is in focus just below her eye) are slightly more defined.

 

Even with a change in bit rate it's not always worth it. The 4K video has a bit rate of 65.7Mbps. The 1080p video has a bit rate of 25.3Mbps. Bitrate:quality does not scale linearly.

 

 

 

Color depth and bitrate > supersampling. With that said though, can't wait to view 4k natively.

Color depth is barely worth mentioning since 99.99% of video is 8bit (at least for the output).

I'd say it goes like this:

Codec > bit rate > resolution

[Puppet]

There are a few changes here and there. The difference is very minor though and I do not think it's enough to warrant a ~3 times as big file and much more work decoding it.

In comparison 1 you can see a tiny bit more noise in the outline of his head (just to the left of his eyebrow).

Comparison 2 is probably the most obvious one. The straws that are in focus are a bit sharper in the 4K version.

In comparison 3 you can see that the "valleys" in the skin (in the tiny strip of skin that is in focus just below her eye) are slightly more defined.

 

Even with a change in bit rate it's not always worth it. The 4K video has a bit rate of 65.7Mbps. The 1080p video has a bit rate of 25.3Mbps. Bitrate:quality does not scale linearly.

 

I do agree, I don't think it's worth the extra effort especially in a moving image when you aren't really going to have the time to pixel peep this much, but just wanted to point out that there was a difference and also invite a discussion as to why?

 

 

The darkest hair in the centre of the red circle has this sort of 'halo effect' around it in the 1080p version (look at original link). I'm no compression expert, at least on the technical side of things, that's the reason I was describing it as looking over sharpened because it's the only time I have seen this defect. Assuming the codec is the same, what is causing this? Color depth, bit rate, or resolution? 

[LAwLz]

I do agree, I don't think it's worth the extra effort especially in a moving image when you aren't really going to have the time to pixel peep this much, but just wanted to point out that there was a difference and also invite a discussion as to why?

 

<image>

 

The darkest hair in the centre of the red circle has this sort of 'halo effect' around it in the 1080p version (look at original link). I'm no compression expert, at least on the technical side of things, that's the reason I was describing it as looking over sharpened because it's the only time I have seen this defect. Assuming the codec is the same, what is causing this? Color depth, bit rate, or resolution? 

It was probably caused by the downscaling. Even the 1080p file was downscaled slightly since the file is 2048x1080 (not 1920x1080 like the images I uploaded). It looks like it's just a ~1 pixel outline that is slightly brighter.

[Puppet]

It was probably caused by the downscaling. Even the 1080p file was downscaled slightly since the file is 2048x1080 (not 1920x1080 like the images I uploaded). It looks like it's just a ~1 pixel outline that is slightly brighter.

 

Okay, yeah after thinking about it you are probably right there, again I don't really know a lot about the technical side of things but when editing photos I'll always sharpen for original and sharpen for web (downscaled) separately or you'll get issues like this, it's the last thing I do. Wasn't thinking about that way because it is video. I suppose on Vimeo you don't really have that option as it's just the one upload.

 

So if anything it's really a user error and they should upload two versions with separate post sharpening in the different renders if they want to be really fussy.