Home · Register · Join Upload & Sell

Moderated by: Fred Miranda
Username  

  New fredmiranda.com Mobile Site
  New Feature: SMS Notification alert
  New Feature: Buy & Sell Watchlist
  

FM Forums | Canon Forum | Join Upload & Sell

  

Archive 2013 · 5D3 int vs external video recording

  
 
skibum5
Offline
• • • • • •
Upload & Sell: Off
p.1 #1 · p.1 #1 · 5D3 int vs external video recording


MOTION/FRAME CHANGE COMPRESSION:

Looking into it a bit more, focused on a blu-ray disc case, hand-held moving camera about a bit, moving in and out of focus, so plenty of changes frame to frame.

Comparing ProRes HQ with NinjA to internal IPB looking frame by frame in PP there is a difference on some fast moving, defocused stuff the internal gets a weird painterly effect that is partly smooth and partly ragged while the Ninja 2 recording stays smooth. So frame by frame over stuff like that the external record looks a lot better.... at least peaked at frame by frame. Very clearly better. Having checked out how noticeable it is to the eye in real-time yet though.

Frame by frame, when things are wholesale changing the external recorder clearly gives better compression than the internal IPB (didn't compare ALL-I yet).

Again I need to see how noticeable it is in real life. Also having lots of entire frame changes during big motions or focus chnages probably aren't the most import things since everything will look blurry etc anyway. But perhaps not have a nasty half watercolor half sharp jaggy motion blur would make such things looks worse than maintain a smooth look. Frame by frame, for wholescale entire frame movements and changes there is no doubt the external compression holds up vastly better than the IPB internal. The ALL-I internal may hold up much better though than IPB to such things. I need to see if the ALL-I doesn't lose fine static details or have other issues though.

-----------------------------------------------------------
RED/WHITE TRANSITION:

External compression is less jaggy. Vertically white to red happens a bit more naturally and better with the external compression.

--------
BUT WEIRD:

On left to right the internal compressor seems to try to balance the change on both sides, maybe doing it a bit more on the right sides of transition types while the external seems to try to load up transitions on the left sides and in a less balanced fashion.

So with the external red and white and so on are less jaggy together and smoother on the external recording overall but OTOH the external gets somewhat more noticeable 'dark sharpening halo' looking effects each time going L to R that red has to hit white or yellow or that black has to hit red. With the internal those are actually less pronounced, less false fake artifact halos even if the transition lines are jaggier and when it does have these fake halos they tend to be more on the other side white going to red or red going to black.

Weird.

vertically the external handles color transitions better, but horizontally something about the internal does it almost seems to leave less noticeably artifacts in some sense.

Oddly when there is red before white going left to right, the ninja 2 ProRes HQ recordings seem to have a wide less saturated, darker area right before white that the internal recordings don't. Almost like it has LESS color resolution horizontally using the HDMI out recorder or that it is applies some sort of left edge sharpening for red into white Or maybe without the white jaggy in it makes for a darker edge and the 4:2:0 actually ironically makes a less smooth and yet more natural transition on those sides? Or maybe it is the internal compressor is placing the transition more in the middle and not totally left biased

Not sure if it is the Ninja 2, the ProRes HQ, Premiere Pro or the 5D3 but the external recoder is giving a dark, de-sat halo line on red to bright white transitions (not so much on white to red) going horizontally. The internal recorder is free of that transition artifact! A bit of a shame since otherwise you can see the external is holding up better on any motion and it has less jaggy red/white transitions at many angles.

Very, very odd. Why would the external ones seem to a giant seeming sub-sampling transition going from any sort of red to white left to right and the internal not?

I also see that black to pink gives a smear of pink black inbetween using the external recorder and yet using the internal it is mostly free of the smear!?!

----------------------------------------------------------------------------------------------------
COLOR/BRIGHTNESS SHIFTS:

I could swear the external recordings seem to clip on the brights a bit quicker. They seem a bit brighter overall Maybe applying the same colorista effect to each was no good, maybe the details in highlights were still there and I shouldn't have brightened it up as much. Need to look into it. Does it clip away more DR externally or just place thing differently.

The external also seems to be a bit green-ish cast compared to the internal. Although I see people complaining, I'm not sure the external is the wrong one though. The blue on the blu-ray case looks to have some green in the blue to my eye and that is what the external recording shows and yet the internal recording show the case to have more of a hint of purple to it than green tinge which seems less accurate. Unfortunately by the time I checked the lighting had change though so I need to re-check when I can view the case under the same color temp lighting as when the footage was taken.



May 01, 2013 at 06:57 PM
skibum5
Offline
• • • • • •
Upload & Sell: Off
p.1 #2 · p.1 #2 · 5D3 int vs external video recording


Watched in real time all the compression artifacts of the internal are not nearly as noticeable. Carefully looking you can spot the worst sets though. Although the eye isn't as trained to care when lots of stuff is moving and changing. So what looks RADICALLY worse frame by frame looks almost the same in many cases, although in the few worst bits you could see the internal one looked a little wonky and the external looked perfect.


May 01, 2013 at 07:26 PM
saneproduction
Offline
• • •
Upload & Sell: Off
p.1 #3 · p.1 #3 · 5D3 int vs external video recording


Weird you compared the external recorder to IPB instead of All-I, because someone who wants the best quality would either record external or use All-I. Also doubt Pro Res HQ makes much difference over regular Pro Res 4:2:2 for most applications. The main reason we use external recording on the C-300 is to keep the footage edit ready without ingest and go from the internal compression at 50Mbps to the Pro Res 422 125Mbps. Never use Pro Res HQ except for effects shots (and then hardly ever do that).

Would love to see a comparo of 5DIII All-I transcoded to Pro Res 4:2:3 vs Pro Res 4:2:2 recorded via HDMI feed, but Internet samples are less good to me as I would like to see it on a native 1080p broadcast monitor. I may do this test. If you put up a split screen sample that would be cool.

Mike



May 01, 2013 at 09:29 PM
skibum5
Offline
• • • • • •
Upload & Sell: Off
p.1 #4 · p.1 #4 · 5D3 int vs external video recording


saneproduction wrote:
Weird you compared the external recorder to IPB instead of All-I, because someone who wants the best quality would either record external or use All-I. Also doubt Pro Res HQ makes much difference over regular Pro Res 4:2:2 for most applications. The main reason we use external recording on the C-300 is to keep the footage edit ready without ingest and go from the internal compression at 50Mbps to the Pro Res 422 125Mbps. Never use Pro Res HQ except for effects shots (and then hardly ever do that).

Would love to see a comparo of 5DIII All-I transcoded to Pro
...Show more

There had been early complaints about ALL-I and mosquito noise and such. I was used to using IPB. ALL-I is variable bitrate and much more inefficient per same bitrate so I wasn't so sure it was necessarily better. I will compare it to ALL-I. I'm sure the motion difference will be less in that case.

After that I will try to do a very careful comparison between all three.

Do you have any clue why the external recordings are getting more bleed on black to red and red to white and red to yellow (if a bit less of red to black and white to red and yellow to red) horizontally? I'd think the external 4:2:2 would be largely the same horizontally and better vertically in all cases.

I'm settling in on one pass of USM 21 strength and 0.5 radius and then one pass 21 strentch sharpen tool. Seems to avoid bring up too many artifacts while making it pretty sharp.

Seems a bit crisper now with 1.2.1 which is very nice.



May 02, 2013 at 12:19 AM
saneproduction
Offline
• • •
Upload & Sell: Off
p.1 #5 · p.1 #5 · 5D3 int vs external video recording


Interesting stuff for sure, sounds like you have tested more than I. I have not seen anything straight out of camera with the new firmware or really compared IBP to ALL-I, just don't like formats that have interframe compression such as HDV as acquisition formats. As delivery formats they are fine, but inter-frame compression is too CPU intensive for editorial purposes and I don't think they do a good job with motion, hand-held footage, whip pans etc. Statics scenes are no test.


May 02, 2013 at 12:46 AM
skibum5
Offline
• • • • • •
Upload & Sell: Off
p.1 #6 · p.1 #6 · 5D3 int vs external video recording


saneproduction wrote:
Interesting stuff for sure, sounds like you have tested more than I. I have not seen anything straight out of camera with the new firmware or really compared IBP to ALL-I, just don't like formats that have interframe compression such as HDV as acquisition formats. As delivery formats they are fine, but inter-frame compression is too CPU intensive for editorial purposes and I don't think they do a good job with motion, hand-held footage, whip pans etc. Statics scenes are no test.


Yeah, I do a lot of tripod mounted nature stuff and mostly super slow pans (quick pans are tricky at 24p too due to judder and with DSLR, jello, unless you are skilled at how to handle things I think) and scenes where not everything is all changing at once and have an nvidia card that PP now hooks into to use GPU decoding for IPB h.264 to help speed things along.

I definitely want to compare everything to all-i too though.



May 02, 2013 at 05:19 PM
skibum5
Offline
• • • • • •
Upload & Sell: Off
p.1 #7 · p.1 #7 · 5D3 int vs external video recording


And of course external lets your record everything so that includes the 5x zoomed mode so you get zoomed 3:2 video at something like DVD resolution (nice for distance limited wildlife stuff, it's much better than trying to scale into the regular mode video when distance limited, obviously it's hardly full glory HD but useful. There is no way to record that using the internal system since they don't allow it.

So you get back DVD zoom mode video that I think a few rebels or something had.



May 03, 2013 at 02:45 PM
skibum5
Offline
• • • • • •
Upload & Sell: Off
p.1 #8 · p.1 #8 · 5D3 int vs external video recording


shooting some highly detailed woods at 24mm, f/8 moving the camera all over the internal IPB compression goes to junk (although motion does make it tougher for the eye to notice, frame by frame some frames are absolute junk though), using all-i fixes up those junky frames a ton (didn't compare yet all-i to external for that)

I didn't compare tripod based, mostly static scene but branches moving and water flowing in a stream type stuff to see if all-i vs ipb makes much difference in compression then or whether the better efficiency of IPB even pulls ahead (from what I hear it's similar and all-i just eats HD space for no reason).


Edited on May 03, 2013 at 03:54 PM · View previous versions



May 03, 2013 at 02:47 PM
saneproduction
Offline
• • •
Upload & Sell: Off
p.1 #9 · p.1 #9 · 5D3 int vs external video recording


skibum5 wrote:
Yeah, I do a lot of tripod mounted nature stuff and mostly super slow pans (quick pans are tricky at 24p too due to judder and with DSLR, jello, unless you are skilled at how to handle things I think) and scenes where not everything is all changing at once and have an nvidia card that PP now hooks into to use GPU decoding for IPB h.264 to help speed things along.

I definitely want to compare everything to all-i too though.


The ASC cinematographers manual has information about rate of panning at various focal lengths to avoid stutter. Note that you would need to convert the focal lengths for your 24x36 sensor and probably go slower to minimize rolling shutter (jello).



May 03, 2013 at 03:04 PM





FM Forums | Canon Forum | Join Upload & Sell

    
 

You are not logged in. Login or Register

Username       Or Reset password



This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.