Jump to content


Photo
- - - - -

5D3 vs D800 video ISO comparison


  • Please log in to reply
9 replies to this topic

#1 Drew

Drew

    The Controller

  • Video Expert
  • 10638 posts
  • Gender:Male
  • Location:GPS is not reliable in South East Asian seas

Posted 23 April 2012 - 11:45 AM

Nofilmshool released a video comparing the 5D3 against the predecessor 5D2 and king of MP, D800. The FOV of the D800 is cropped from the FF sensor, ostensibly because of the high pixel count and downsampling ratio. Basically it's 91% of the full width, so FOV on the widest lens will be 9% narrower.

Anyhow, note that the comparison is a bit skewed. The D800 allowed 422 uncompressed capture via HDMI and that's what was used in the comparison. The 5D3 used ALL-I processed in 5DtoRGB at REC709. Strangely the 5D2 was processed at REC601. It's obvious that the resolution from 422 uncompressed is nice but the noise is pretty bad. I'd say over 2.5 stops because there's no compression and noise reduction incamera for the D800.

[vimeohd]40113110[/vimeohd]

What's amazing is that at ISO25600, the candle looks like a 100W bulb!!! That's seeing in the dark!!!

Drew
Moderator
"Journalism is what someone else does not want printed, everything else is public relations."

"I was born not knowing, and have only had a little time to change that here and there.


#2 Drew

Drew

    The Controller

  • Video Expert
  • 10638 posts
  • Gender:Male
  • Location:GPS is not reliable in South East Asian seas

Posted 23 April 2012 - 02:11 PM

Here's a comparison of 5D3 clips sharpened in CS5.5

[vimeohd]40184470[/vimeohd]

The picture does sing a lot more once it's worked on. Not much sharpening artifacts at all if you don't go nuts!

Drew
Moderator
"Journalism is what someone else does not want printed, everything else is public relations."

"I was born not knowing, and have only had a little time to change that here and there.


#3 A.Y.

A.Y.

    Wolf Eel

  • Member
  • PipPip
  • 183 posts
  • Gender:Male

Posted 24 April 2012 - 07:52 AM

[youtubehd]lIKcfKhE-ew[/youtubehd]

These side-by-side high-ISO comparisons show the video performance from both cameras are "closer" up to ISO 3200 not "lopsided". At ISO 6400, D800 noise starts to become quite visible - this is the price one pays for 36mp I guess.

The daytime dynamic-range test seems to show D800 has a slight edge; however, IMO, the difference is most likely only noticeable during a side-by-side comparison.

My question is why the vimeo piece from "Joe Marine" produced such lopsided results while other similar comparisons show far closer performance?

Edited by A.Y., 24 April 2012 - 08:37 AM.


#4 Drew

Drew

    The Controller

  • Video Expert
  • 10638 posts
  • Gender:Male
  • Location:GPS is not reliable in South East Asian seas

Posted 24 April 2012 - 10:58 AM

I suspect the difference is the HDMI output, which bypasses NR and also the fact that the Adorama tests adjusted aperture to maintain exposure. Whereas the first test just left exposure as is to see sensor sensitivity. This is of course assuming he didn't screw up the test or the Ki Pro didn't capture badly.

Regardless, while the noise looks crappy, the fact is that the D800 produced a useable, sharper image at a lower ISO via HDMI capture. For some scenes, it's easier to add light and have a sharp pic. Even though the 5D3 ALL-I takes sharpening well, it can't beat 422 uncompressed.

Drew
Moderator
"Journalism is what someone else does not want printed, everything else is public relations."

"I was born not knowing, and have only had a little time to change that here and there.


#5 Drew

Drew

    The Controller

  • Video Expert
  • 10638 posts
  • Gender:Male
  • Location:GPS is not reliable in South East Asian seas

Posted 01 May 2012 - 04:19 PM

Yet another 5D3vs D800 test:

http://www.dslrnewss...canon-5d-mkiii/

Here's a video sample:

http://tinyurl.com/7klttj9

Drew
Moderator
"Journalism is what someone else does not want printed, everything else is public relations."

"I was born not knowing, and have only had a little time to change that here and there.


#6 foxwinter

foxwinter

    Sea Wasp

  • Member
  • Pip
  • 2 posts

Posted 09 May 2012 - 12:51 AM

This comparison is more about video information between 5D3 & D800



#7 loftus

loftus

    Blue Whale

  • Member
  • PipPipPipPipPipPip
  • 4571 posts
  • Gender:Male
  • Location:Winter Park, Fl

Posted 09 May 2012 - 03:55 AM

This comparison is more about video information between 5D3 & D800

The similarities and differences are pretty similar in most of these tests; basically for most folks - stick with the system you're invested in.
If you are not vested, shoot mostly stills and you want the camera with the best DR with clearly superior shadow detail and resolution even as high as 6400 (dp preview) go Nikon.
If shooting higher than 1600-3200 and 6fps vs 4ps are important parts of your shooting, and as a video buff moire is now an issue, go Canon. Unless of course that HDMI output is important.
Nikon D800, Nikon D7000, Nauticam, Inons, Subtronic Novas. Lens collection - 10-17, 15, 16, 16-35, 14-24, 24-70, 85, 18-200, 28-300, 70-200, 60 and 105, TC's. Macs with Aperture and Photoshop.

#8 HDVdiver

HDVdiver

    Eagle Ray

  • Member
  • PipPipPip
  • 327 posts
  • Gender:Male
  • Location:Adelaide, Australia...Great White Shark country.
  • Interests:Dive,dive, dive...

Posted 14 May 2012 - 01:54 AM

The similarities and differences are pretty similar in most of these tests; basically for most folks - stick with the system you're invested in.
If you are not vested, shoot mostly stills and you want the camera with the best DR with clearly superior shadow detail and resolution even as high as 6400 (dp preview) go Nikon.
If shooting higher than 1600-3200 and 6fps vs 4ps are important parts of your shooting, and as a video buff moire is now an issue, go Canon. Unless of course that HDMI output is important.



Also have a look at:

"Full Frame Shootout: Nikon D800 / D4 / Canon 5DmkIII"



The D800 tended to have noticeably better video resolution.

#9 Drew

Drew

    The Controller

  • Video Expert
  • 10638 posts
  • Gender:Male
  • Location:GPS is not reliable in South East Asian seas

Posted 14 May 2012 - 05:34 PM

One thing Phil forgot to mention is that the D800 FX video mode is actually a 1.2 crop, which can mess up the lens choice. Instead of the 16-35, one has to use the 14-24 to get the a similar (but slightly smaller) FOV @ 16mm.

Drew
Moderator
"Journalism is what someone else does not want printed, everything else is public relations."

"I was born not knowing, and have only had a little time to change that here and there.


#10 HDVdiver

HDVdiver

    Eagle Ray

  • Member
  • PipPipPip
  • 327 posts
  • Gender:Male
  • Location:Adelaide, Australia...Great White Shark country.
  • Interests:Dive,dive, dive...

Posted 23 May 2012 - 11:29 PM

I'm actually hoping the rumors are correct and Nikon will release a D600:

http://nikonrumors.c...now-at-80.aspx/

At 24 megapixels full frame it could actually be a better all-round camera for cinematography in terms of low light performance and video artifacts. Now that Nikon are serious about competing in the video DSLR market (and given that they don't have a pro-video range to protect) they could potentially unleash the demons with this camera. :P