C100: Banding issues in footage please help

Splendid

Member
Hi Guys,

Today i was testing my C100 on the beach and i noticed (for the second time) there are wierd issues in the sky of the footage. Actually the clips are unusable i hope someone can help... (Lucky i was just testing)

On most of the footage its not visible, but it was when the sun was in my back. I shot on C-Log with NR of 5 and a 55-250 STM lens, however i also noticed this issue a couple of months ago on the same location with the 24-105 L.
Iso was not high, and it was not really underexposed... well i made a short video to be more clear:

https://www.youtube.com/watch?v=3gpZ-Ig5ynM&feature=youtu.be

Thanks for any help, much appreciated :-)

Regards,

Rob

EDIT: Above video was in AVCHD internal because i was on a glidecam, however i did some more testing on the Ninja blade, and it seems in C-Log i need to overexpose quite a bit especially when using ND's...
 
Last edited:
Ahh, the downsides of an 8bit codec. If it's becoming more severe after transcode, take another look at your settings and workflow. Although not the source of it, it could be amplifying it a little more than it should.
 
Ahh, the downsides of an 8bit codec. If it's becoming more severe after transcode, take another look at your settings and workflow. Although not the source of it, it could be amplifying it a little more than it should.

Thanks for the reply, however the whole reason i bought such a pro camera is to not get these kind of issues... a bit disappointing really. :crybaby:

Guess i need to watch carefully at my exposure in certain cases. BTW i never experienced any of these kind of issues with my DSLR's, but then again those images were mediocre at best compairing to the C100 :laugh:

EDIT: I read i another thread that WDR might be better, and less noisy because of the 8-bit limitation, will test that tommorow

Cheers
 
Last edited:
I shoot exclusively with c-log(nr off/view assist enabled) and never run into any of these problems. Did you use the waveform to dial in your exposure?

I suggest doing another round of tests. Try capturing 5 shots without the waveform(eyeballing it) but before changing locations enable the waveform and take a photo of it reference. Do another 5 shots trying to dial in it exclusively using the waveform, again taking still photos of the final waveform. Bringing all of them into an nle to compare should quickly show you how far off the eyeball/lcd is. For more thorough testing, bring along the dslr for side by sides :)
 
Were you shooting under ISO 850 in C-Log? I've heard that can introduce issues like the ones you're experiencing.
 
I shoot with the rec407 file (based on WDR) that's posted in the stickys. I've had very few banding issues compared to other cameras - I transcode to ProRes in FCP7, then import the files without reprocessing them into FCPX. Banding is much more of a curse than color space with 8bit though, especially downstream for broadcast. I've seen material that looked clean, shot with the AF100, that started looking pretty crappy when aired, especially in skys and dark areas. So far, no problems with the C100.

However, I've basically chickened out and only use the C100 for web or local now. Anything for national broadcast I shoot with my Pocket cameras. Those little things have their issues too, like moire, but that's easier to deal with in post than banding - and honestly, moire doesn't crop up as much as some believe.
 
Nucoda FilmMaster, Pablo, Baselight, Mystika, DS, Smoke, Nuke, DF all have anti-banding filters, either native or through plugin's, i'm not familiar with MC/FCP/FCX/PP/lightworks, but i would google anti-banding filter + the name of your choice of NLE and see what turns up

d
 
Thanks all for responding!

No i did not shoot under 850 ISO and i always use the waveform in the way that is suggested by canon (youtube tut). Also it has nothing to do with workflow and transcoding, as the issue is also visible on my ninja blade in prores hq...

However i shot a music video today for the first time in WDR, thus far i like it, and i did not have any issues thusfar, but then again i did not do any shooting on one object against the bright sky (Whats seems to cause these problems)

Regards,

Rob
 
Thanks all for responding!

No i did not shoot under 850 ISO and i always use the waveform in the way that is suggested by canon (youtube tut). Also it has nothing to do with workflow and transcoding, as the issue is also visible on my ninja blade in prores hq...

Rob

With the c100 I can underexpose by 1.5-2 stops and still save most footage if its shot externally to a 10 bit container. So your trouble is inherent either to the way in which it was captured, or post workflow. Internally things are a bit more sensitive and need to be done with more precision. Having said that my c100 never leaves its C log profile and I have never had an issue with banding or noise.
 
With the c100 I can underexpose by 1.5-2 stops and still save most footage if its shot externally to a 10 bit container. So your trouble is inherent either to the way in which it was captured, or post workflow. Internally things are a bit more sensitive and need to be done with more precision. Having said that my c100 never leaves its C log profile and I have never had an issue with banding or noise.

You are right, 99% of the time its fine, and i also almost always record in prores externally. However since i had the camera it happened to me twice now.


What NLE?

Would you mind sharing the original clip for a test?

I use premiere pro CC

Here is the AVCHD file straight from the card:

https://drive.google.com/file/d/0B4Dvjb4nEVFWaTBaV1UzckpDVXM/view?usp=sharing

Here is the one i recorded couple of months ago, it seems ok, untill you start to grade it, recorded on the ninja.

https://drive.google.com/file/d/0B4Dvjb4nEVFWMXdoSkQzU3dkM28/view

Both clips have banding issues in the sky.

Cheers
 
Last edited:
While external recorders have some merit since you can tap into an uncompressed signal, there is no benefit to the 8bit vs 10bit problem. 422 color space will help in many scenarios, but hardly this one.

The C100 outputs 8bit and wrapping that into a 10bit container adds nothing to help with banding. Even if the file is captured at 8bit, as soon as we bring it into our NLEs every computation will be made at a significantly higher bit depth. If you work with DaVinci you get 32bit floating point.

Different NLEs DO have different optical engines however, and final output might very well vary.

What was the post workflow for this file?
 
95% of the time i use the ninja to record, because its a great monitor too, i use a custom prores sequence in pp, when i do record with the C100 in avchd i almost always transcode it to prores with clipwrap.
I sometimes use gopro footage, and i transcode everything to prores because its so much better to work with. (rendertimes and whatnot)

I mostly use filmconvert for the grade, but it depends of the project. For the normal "look" i use the tools within PP. I rarely use a LUT because i feel i then loose control.
 
Here is the AVCHD file straight from the card…

Thanks.

Well loading it up in FCPX (in my case) it has got about the same about of banding even before we touch it. Even if it might have been better if this was due to some "post issue" (which could be fixed) I'd say it simply 8bit vs an extremely flat sky. I'm shooting 8bit myself all the time and I never see banding. I don't peep for it, so I'm sure I could find some, but it's not an issue for me.

In your case it's a combination of a prominent sky and very little gradient. A few whisks of clouds or just a little more natural contrast and you would probably have been fine.

If you have a clip like this that is essential, you can try breaking the banding up with some added noise in post. Doesn't have to be much. Just enough to redistribute the light values. Might even bring a little of that "film panache".
 
Thanks.

Well loading it up in FCPX (in my case) it has got about the same about of banding even before we touch it. .

Yes that is what i meant ;-) Its not a major issue because most of the time the footage looks great. I guess i have to be careful in such circumstances.

We never stop learning :laugh:

Thanks for your input

Cheers
 
The C100 outputs 8bit and wrapping that into a 10bit container adds nothing to help with banding.

Myself, as well as many others are well aware of the fact that the c100/300 output more than 8 bits of information via the hdmi(regardless of what oem specs state). There is plenty of evidence, tests and information out there to substantiate the claim. So in fact, it would make quite a large difference shooting to a 10 bit container vs 8 :)

Looking at the test clip, looks like some nasty stuff. That is one flat sky!
 
Myself, as well as many others are well aware of the fact that the c100/300 output more than 8 bits of information via the hdmi(regardless of what oem specs state). There is plenty of evidence, tests and information out there to substantiate the claim. So in fact, it would make quite a large difference shooting to a 10 bit container vs 8 :)

Looking at the test clip, looks like some nasty stuff. That is one flat sky!

May be so, but the problem also exists when recorded on the ninja in prores hq, i tested this the same day on the same sky. You can even see it on the screen of the blade itself ;-)

If i have time i will post that footage as well.

Cheers
 
Myself, as well as many others are well aware of the fact that the c100/300 output more than 8 bits of information via the hdmi(regardless of what oem specs state). There is plenty of evidence, tests and information out there to substantiate the claim.

I have tried to keep myself current on these cameras since they launched. If what you say were true, that would be some sensational news. If you know of plenty of evidence, it would sure be nice if you could link to some of it.

Well encoded 8bit footage is normally not a problem to grade in a quality application. Shooting too flat "internet color profiles" or trying to correct badly exposed footage will probably present problems, though.

No doubt, moving from 420 to 422 will give you benefits if you do some creative color grading, and going uncompressed to ProRes should keep everything nice and fat (but look at the other ProRes thread regarding muddy ProRes 422 on a Ninja Star).

Anyway, if you have got some "evidence" to your statement, I'll gladly accept it.
 
I had three 8bits cameras and there was no problem with banding, with their cine1 gamma. C100 has sometimes problem with banding (with both wdr and log gamma) but cine1 gamma solves this. It is my way to solve this limit in case of flat sky or sunset.
 
Back
Top