Author Topic: detail test  (Read 5789 times)

0 Members and 1 Guest are viewing this topic.

Offline uknerd

  • Contributor
  • Posts: 26
detail test
« on: January 11, 2015, 06:54:51 pm »
Dave

I have to say having watched all three of the digital fps tests, the 60fps video was marginally better then the 50fps video and miles better then the 25fps test

i was watching it on my iphone 5 with ios 8

I have to say the findings would match with my thoughts on the pre-watch predictions.

Nerd.



 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 10415
  • Country: 00
Re: detail test
« Reply #1 on: January 11, 2015, 08:53:08 pm »
What you really need to do is watch it without knowing what the frame rate is.

It's called a "blind test" and that's what makes real science.
 

Offline wraper

  • Supporter
  • ****
  • Posts: 10807
  • Country: lv
Re: detail test
« Reply #2 on: January 11, 2015, 10:19:50 pm »
Don't think that you could spot any real detalization difference on such a small screen. Don't think that you would even see 720P and 1080p to be a bit different. There is no distinguishable detalization difference on my 15" full HD laptop too, however there is on 30" monitor, and not in favor of 60 FPS.
 

Offline wraper

  • Supporter
  • ****
  • Posts: 10807
  • Country: lv
Re: detail test
« Reply #3 on: January 11, 2015, 10:23:31 pm »
And yeah, just checked. Iphone 5 does not even have 720p resolution, not to say 1080p. Only 1136 x 640.
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 10415
  • Country: 00
Re: detail test
« Reply #4 on: January 12, 2015, 09:16:19 am »
I was going to mention screen size too...

How about a new full disclosure rule for those saying "60Hz is better!!!" What are you watching it on? What's the refresh rate of the device? Did you do a blind test (where you don't know what rate/resolution you're actually watching)?

It could all be placebo effect without that.

 

Offline AndreasF

  • Frequent Contributor
  • **
  • Posts: 251
  • Country: gb
    • mind-dump.net
Re: detail test
« Reply #5 on: January 12, 2015, 09:42:39 am »
Ever since Dave started with the high frame-rate videos, I've often had to scale back to the next lower quality setting (720p, high-frame rate) because with 1080p the video would start to stutter/hang frequently for a few seconds (most likely my PC beginning to start its age), which is a real shame, because the perceptual difference between low and high frame rate seems marginal in my opinion (maybe that's me showing my age), especially for the type of static shots in Dave's videos, whereas the difference between 720p and 1080p is HUGE when viewed on a standard 20+ inch screen.
my random ramblings mind-dump.net
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 10415
  • Country: 00
Re: detail test
« Reply #6 on: January 12, 2015, 11:05:20 am »
the perceptual difference between low and high frame rate seems marginal in my opinion (maybe that's me showing my age)

I don't even know how people are claiming to be able to tell the difference when all the evidence is based on watching completely different videos. It could easily be down to different camera focus in each video ... or even just people's own imagination and confirmation biases.

The only way to be sure would be to produce two different versions of the same video, one at 30fps and one at 60fps then get people to vote without knowing which was which. The "30fps" version would have to be made by discarding every other frame from the 60fps version then upscaling it. It would have to be done very carefully (and even so the process might skew the result).

If Dave does that and there's a very clear result in the votes then I say go with it.

Otherwise...nah. There's simply not enough motion in the videos to justify it, not if it's causing problems/delays.

It's a cost/benefit analysis. If the cost is zero then you do it, obviously.

I'm not against 60fps because I'm a curmudgeon (although it helps), I'm just not convinced the self-proclaimed videophiles (videophools?) claiming to see a difference are really seeing one. Not when it turns out they were watching on an iPhone (or whatever).

 

Offline bktemp

  • Super Contributor
  • ***
  • Posts: 1623
  • Country: de
Re: detail test
« Reply #7 on: January 12, 2015, 03:20:50 pm »
If you look at areas with high motion, the 60fps version looks best.
The main problem is the framerate conversion: 60fps to 50fps can be either done by dropping every sixth frame, but this introduces jitter, or by interpolating between frames. The result is motion blur, because almost every frame consists of 2 input frames merged together.
The same applies to the 25fps version: It is even worse, because every output frame is based on 3 input frames from the 60fps source.

Regardless of which framerate you choose, you always get the best possible quality by using the least amount of format conversions (framerate changes, recompressions etc.) during editing.
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 10415
  • Country: 00
Re: detail test
« Reply #8 on: January 12, 2015, 08:11:59 pm »
If you look at areas with high motion, the 60fps version looks best.

Obviously.

But there aren't many areas with high motion.   |O

 

n45048

  • Guest
Re: detail test
« Reply #9 on: January 12, 2015, 09:10:23 pm »
There won't be a difference in image quality. Motion in 60fps will appear much smoother (but the average viewer probably won't see much of a difference in the content Dave produces). I dabble in video production and to me the difference is very pronounced but I'm also happy with the 25fps stuff. The 60fps for this purpose would probably be a bit of a waste of disk space. It will also exclude people with average internet connections from seeing the 1080/60P content (although I think Youtube automatically renders it down to 1080/25P anyway).

If Dave wants to improve the image quality itself, he is better off recording in Avid DNxHD format (if his camera has a HD-SDI output). I use the Atomos gear (http://www.atomos.com/) and the image quality in DNxHD (or Apple ProRes) is phenomenal. Of course you wouldn't upload the DNxHD files (as they are up to 220Mbps codecs) however the resulting rendered files are far better than re-rendering a H.264 or AVCHD files. Adobe Premiere supports it natively in the newer versions but you can also download a free codec from Avid to enable importing into older versions or other NLE tools. Personally I'd avoid Apple ProRes as it tends to have some compatibility issues plus their codec isn't free.

You'll even find that the final rendering time with DNxHD is much quicker than working with already compressed video.

A bit of an overview here: http://www.redsharknews.com/post/item/88-why-you-should-use-dnxhd-and-prores
« Last Edit: January 12, 2015, 09:23:40 pm by Halon »
 

Offline Circuitous

  • Supporter
  • ****
  • Posts: 236
  • Country: us
    • Corgi-Tronics
Re: detail test
« Reply #10 on: January 12, 2015, 09:59:10 pm »
I don't like the high frame rate videos.   I have the same problem of having to drop back to 720 resolution.  It doesn't seem to be bandwidth, probably a problem trying to render all of that detail fast enough.

Offline Phil_L

  • Newbie
  • Posts: 4
Re: detail test
« Reply #11 on: January 13, 2015, 02:25:25 am »
For those of you having problems stuttering at 1080/60FPS, if you're using Chrome try disabling hardware acceleration: http://www.reddit.com/r/youtube/comments/2l283t/1080p_60_fps_videos_lagging/

Fixed it for me with my old HD5770.
 

Offline bktemp

  • Super Contributor
  • ***
  • Posts: 1623
  • Country: de
Re: detail test
« Reply #12 on: January 13, 2015, 06:12:12 am »
Are you saying the blurring evident is because of encoding changes of frame rates? So the 60fps image is camera blur and the 50 and 25 are made more blurry because of the encoding.
It is not the encoding itself. Framerate coversion is an additional step Dave does.
If I remember correctly, he mentioned once one of his cameras can only do 50fps but not 60fps and the other camera can do 60fps but not 50fps.
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 10415
  • Country: 00
Re: detail test
« Reply #13 on: January 13, 2015, 11:38:12 am »
If I remember correctly, he mentioned once one of his cameras can only do 50fps but not 60fps and the other camera can do 60fps but not 50fps.

It's weird that modern cameras are still sold by region with 50fps and 60fps versions and no way to select frame rate. Any TV capable of playing back 1080p will adapt to the incoming signal but most PC monitors are running at a fixed 60Hz so you'll get jitter/blurring when you play "PAL" video on them*.

What are "PAL zone" filmmakers/bloggers supposed to do to be able to record at 60fps for the Internet? Import their cameras from the USA, pay extra for the privilege and have warranty problems (even if they honor the warranty, will they have a 60fps replacement camera if it breaks...)?

I bet the NTSC/PAL cameras are identical inside. Is there a way to load "NTSC" firmware onto a "PAL" camera to get 60fps for blogging?
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf