How do you sharpen the soft video of 5D3 (premiere CS6)

awful
is that a screengrab from premiere, or does it appear also when you output your final video? I've seen similar artifacts disappear just by restarting Premiere... and then I found they were not showing up in the final render anyway
 
that was visible in premiere, i then exported an mpeg and re-imported that mpeg. That screen grab was taken from the re-imported mpeg. It's also visible (obviously harder to spot without being able to nudge forward frame at a time) when the exported mpeg is played back in windows media player
 
it's probably in the original footage then
(mpc-hc has very nice scrubbing and screen-grabbing abilities)
 
I downloaded the clip to check it in Adobe CC. I see the same artifacts in Premiere Pro CC and also After Effects CC. The really bad blockiness lasts for exactly one frame, then it goes away. I have a problem though in trying to understand what is going on because:

This is not the original IPB footage from the 5D3. This has been transcoded to MPEG 2 at a very low bitrate, 10.9 Mbps. How did you transcode this clip? What program?

The only way to get to the bottom of this mystery artifact frame is to look at the original 5D3 footage in a known good program like After Effects CC. I would have to treat this like a scientific test, without introducing unknown variables like transcoding to a different format.

Maybe it is better to use ALL-I all the time, that is what I have been doing anyway.
 
Last edited:
Thanks Celluloid and Samuel, I appreciate your continued help. I outputted that file from premiere pro to see if the artifacts disappeared.

Here is the original file straight from the camera, untouched. The artifacts appear when focus shifts at around 12 seconds.

www.rideoutfilms.com/original5d3.zip

Thats what I should have uploaded in the first place thinking about it..
 
Last edited:
I got the original file (thanks for uploading) and put it into After Effects CC. There is definitely a problem here. The macro blocking lasts for two frames and then it is gone. I have never seen these type of artifacts happen before. Here is the output from After Effects of the (first and worst frame) as uncompressed png. http://we.tl/cWeaMdAETo

If you look carefully around the frame, you can find blocks all over the place, not just on the hand. I really think this is some sort of anomaly, like maybe some type of electronic interference or a bad block on the memory card or something? I just do not think a normally functioning 5D3 would do this in IPB mode.
 
Celluloid Dream said:
I got the original file (thanks for uploading) and put it into After Effects CC. There is definitely a problem here. The macro blocking lasts for two frames and then it is gone. I have never seen these type of artifacts happen before. Here is the output from After Effects of the (first and worst frame) as uncompressed png. http://we.tl/cWeaMdAETo

If you look carefully around the frame, you can find blocks all over the place, not just on the hand. I really think this is some sort of anomaly, like maybe some type of electronic interference or a bad block on the memory card or something? I just do not think a normally functioning 5D3 would do this in IPB mode.
Click to expand...

Thank you for doing that test. The problem is it's quite hard to replicate the problem so I'm really not sure how to test this scientifically to see if it's camera or memory card.. Any suggestions?!
 
thinking about it.. i can rule out the card as I have a 32gb that I use exclusively for speeches and I just saw similar artifacts in the next wedding I'm working on, but then in the past one of the shots I frame grabbed to show the blocks was during the photographs so it definitely wont have been the same card. They're all the same type of card, Transcend 400x CF cards but it wouldn't have been a single bad card.

Not sure if I've mentioned I use and rely on Magic Lantern too.
 
I'm now considering selling the mkiii and going C100 but not sure if I'll miss the full frame look. What do you guys think?
 
If you don't need the stills (I do on current assignment), then it's an easy switch. The C100 look is just delightful, so much so the full frame vs super 35 debate really doesn't cross you mind much.
 
It's not exactly the same, but pick up a Tokina 11-16 F2.8 and you can at least get something full-frame-ish looking if you swap to the C100.
 
yeah I always ruled it out because of the lack of wide angle until someone here suggested the 11-16 (I've even read about a 10-22) so now I've started looking into it I'm pretty much sold.

I may hire one to do some tests for myself as I can't seem to find any visual representation of the difference between the dof of the c100 and 5d's at the same aperture/focal length. I dont want everything filmed at f1.4 to look like it's f2.8... Please slay me down and tell me that's not going to happen (as thats what I want to hear!)
 
The crop factor of the C100 is about 1.5. The depth of field difference between the 5D3 and the C100 is about 1 and 1/8 stop. 30mm @ f/1.4 on a C100 has a similar field of view and depth of field as 45mm @f/2.5 on the 5D3.

I don't wanna spend $5000 of your money, but the C100 does have a great look. Here is one solid example I found with good production technique. https://vimeo.com/74684667
 
Last edited:
Celluloid Dream said:
The crop factor of the C100 is about 1.5. The depth of field difference between the 5D3 and the C100 is about 1 and 1/8 stop. 30mm @ f/1.4 on a C100 has a similar field of view and depth of field as 45mm @f/2.2 on the 5D3.
Click to expand...

wow.. as big a difference as that? That's a pretty major difference.. I have read in some places the crop is 1.3x and other places 1.5x, I assume at 1.3x the difference wont be so great?

There's no denying the clarity in that vimeo clip is amazing, but it's nowhere near as narrow DOF as I'd usually shoot. (heres an example of my work on 5d mkii's: https://vimeo.com/85602293)

Can anybody confirm if the c100 is a 1.3x or 1.5x and if 1.3x, what would that 30mm f1.4 shot have a similar dof to on the 5D3?
 
To make a meaningful comparison, you want to consider the depth of field at the same angle of view and subject distance. A Sigma 30mm Art at f/1.4 on the C100 would have a very similar angle of view and depth of field (at the same subject distance) as a 5D3 with a 44mm lens set at f/2.5. This is not an exact calculation, but very close. For a practical real world comparison with two real lenses, the C100 plus Sigma 30mm 1.4 would have a DOF look very similar to the 5D3 with the Canon 40mm f/2.8. (The actual focal length of this lens tested by DXO is 43mm)


The C100 sensor is 24.6 x 13.8 mm (28.2 mm diagonally). The 5D3 sensor is 36 x 24 mm (43.4 mm diagonally) for stills, and 36 x 20.3 mm (41.3 mm diagonally) for video. Most of the time the C100, C300, and C500 are referred to as having a 1.5x crop factor compared to the 5D3, but its not exactly 1.5x, more like 1.45x or so. It is definitely not 1.3x.

Have a go here at the Abelcine Field of View Comparator. http://www.abelcine.com/fov/

And you can get an idea of the depth of field here. http://www.dofmaster.com/dofjs.html (You can use Cinematography 35mm film as a reasonable substitute for the C100 on this page.)
 
Last edited:
thank you very much celluloid thats a great help.

Do you know if it's true that you shouldn't switch lens on the c100 without powering down the camera? Using only primes I change every couple of minutes during a wedding and this just isn't practical for me.. is it any more harmful to the camera with the c100 than the 5d or have I always been doing it wrong?!
 
Yes, it is true and you haven't been doing anything wrong with the 5D3. They are different in that respect.

According to Canon in the C-100 manual, page 29, you should turn the camera power off to attach or remove a lens. Also, you are supposed to turn the camera off when inserting or removing memory cards.

Interestingly, and also according to Canon manuals, EOS still cameras like the 5D3 do not need to be powered off to attach or remove a lens. They are however, supposed to be powered off when inserting or removing memory cards.
 
well thats a new bit of information for me. I'll try to get info from the canon rep on whether this is a just in case or a necessary,,
 
Just a little anecdote... I shot an interview with the 5d3, just regular h264. Had it on my plasma monitor while editing and then a seasoned VFX- artist walks into the room (they were doing vfx for a red epic shot 5k commercial), looks at the screen and blurts out: "Man that looks really good, very sharp".

Okay, I had added a slight bit of sharpening in Premiere but I still found it quite funny considering how soft people think the 5d3 is. He had just been watching local tv in SD so I guess that also made a difference.
 
Just picked up a 5d3 last week and I'm surprised how sharp many shots are turning out to be with 10-20 sharpening in FCP. Latitude still wanting. Going to try Sam's Flatt 10 this evening.
 
You must log in or register to reply here.
Back
Top