Jump to content

C100 Mkii - Resolution and Image Quality - How to Get the Best Out of It?


User
 Share

Recommended Posts

Hello Folks,

So for example I've seen footage on the net where the image from the C100 Mkii is nice and detailed. And, I have seen the opposite. I recently came across this and it seems to me to be a very detailed and clean image: https://vimeo.com/140112461. Yes it was recorded via a Ninja 2, so I'm wondering if this unit helped and to what degree? Moreover, I understand that there can be many variables that come into play when working towards getting the best out of a camera - lenses and apertures, lighting, etc. I guess I'm wondering what one wants to keep in mind in order to get the best out of a camera like the C100 Mkii?

Another question? I understand it, the image has been 'down sampled' to 1080 from a 4k sensor and I even read where someone was saying that because this happens 'in camera' it is the actually equivalent of having a 4k image. Can someone comment as to the 'truthiness' of this?

This camera has quite a bit going for it in terms of built in tools that help you work despite the fact that it is having its ass roasted by the competition.

Link to comment
Share on other sites

Shoot in log and expose to the right. Don't believe the 35% middle grey Canon says in log. Aim for 50-60% with caucasian skin tones. ProRes is nice but it's really not much more detailed than the mp4 or AVCHD if you expose correctly. There are tests around youtube and vimeo outlining this. 

Link to comment
Share on other sites

Another question? I understand it, the image has been 'down sampled' to 1080 from a 4k sensor and I even read where someone was saying that because this happens 'in camera' it is the actually equivalent of having a 4k image. Can someone comment as to the 'truthiness' of this?

That is indeed how it works. It downsamples a 4K sensor readout to 1080p very early in the image processing pipeline to improve SNR and color. Here's a long boring white paper from Canon explaining the finer points. http://www.cs.ucsb.edu/~mturk/imaging/Misc/EOS_C300_New_35mm_CMOS_Sensor_WP.pd

I don't know much about internal vs external recording as I haven't tried them both, but count me as a +1 for exposing to the right on Canon cameras. As long as you avoid clipping any color channels, it seems to respond well to bringing the image down in post. 

Link to comment
Share on other sites

Shoot in log and expose to the right. Don't believe the 35% middle grey Canon says in log. Aim for 50-60% with caucasian skin tones. ProRes is nice but it's really not much more detailed than the mp4 or AVCHD if you expose correctly. There are tests around youtube and vimeo outlining this. 

Hi Kubrickian, thank you for the straight and concise dope here. You have mentioned C-log as the way to go and I'm curios as to why you would say that when others are suggesting the wide dynamic setting. Would appreciate hear your thoughts here.

Also, can you (or anyone) point the way towards learning more on exposing for middle grey with Canon C-log? And exposing for 50-60% for caucasian skin tones? I could really use a course here.

And one more thing, the subjects of the film I'm shooting are in are South Asians. Any thoughts of exposing for their skin tones?

Big thanks to Renaissance for confirming the 4k to 1080 with the Canon White paper :)

Link to comment
Share on other sites

I have a Mk II and I did some extensive tests with external ProRes recorder. There is tiny bit of improvement but for me it was not nearly worth the upgrade in file sizes. If I were shooting a 30 second spot that was going to be on television I'd probably use it.

Check out what this guy has been doing with the Mark II. Great spots!

Link to comment
Share on other sites

That was good! It's a strange camera. At first the image seems average but the more that I use it I fall in love with it. It's easy to operate and very user friendly. I was comparing some stuff I shot the same day with a Blackmagic Pocket Camera and the C100 and was more pleased with how the C100 footage turned out.

Great to hear such a comment. Thanks for that!

For me the lenses are important esp. with the C100. With EF-S lenses the output is very soft, only with L Lenses the Camera really shines. 

I've got a small arsenal of L glass left over from my 5DM2 but I'm going to need something wider in moving towards the C100's 35mm sensor. Three lenses are on the table: Canon 16-35mm f4 IS, Sigma 18-35mm f1.8, Canon 17-55m f2.8. From everything I've come across, the Sigma takes the cake in sharpness. But then again it's missing the IS. Does anyone recommend one over another and why? My feeling is that I could get away with no IS for run and gun on the Sigma as it is still a fairly wide lens and that can help level the bumps out.

In my experience shooting Wide DR gets noisier and has more artifacts than c-log. I like to grade. When we say expose to the right we're talking about checking your waveform. Look for skin tones halfway up the meter. 

Got it Kubrickian. Thank you for the clarification :)

Much obliged to everyone here and thanks again for your comments!

I have a Mk II and I did some extensive tests with external ProRes recorder. There is tiny bit of improvement but for me it was not nearly worth the upgrade in file sizes. If I were shooting a 30 second spot that was going to be on television I'd probably use it.

Check out what this guy has been doing with the Mark II. Great spots!

Too funny Aaron, this kind of material always makes me think of starving africans and refugees. But yeah the images theselves hold up rather well ;)

Link to comment
Share on other sites

I've got a small arsenal of L glass left over from my 5DM2 but I'm going to need something wider in moving towards the C100's 35mm sensor. Three lenses are on the table: Canon 16-35mm f4 IS, Sigma 18-35mm f1.8, Canon 17-55m f2.8. From everything I've come across, the Sigma takes the cake in sharpness. But then again it's missing the IS. Does anyone recommend one over another and why? My feeling is that I could get away with no IS for run and gun on the Sigma as it is still a fairly wide lens and that can help level the bumps out.

hi User - (i usually shoot run and gun) the problem with the sigma 18-35mm is that it's a pretty heavy lens. it gets heavy after shooting for a bit, and i consider myself a really steady shooter. i really like the 24 to 70, 2.8 despite not having an IS. 

 

Link to comment
Share on other sites

hi User - (i usually shoot run and gun) the problem with the sigma 18-35mm is that it's a pretty heavy lens. it gets heavy after shooting for a bit, and i consider myself a really steady shooter. i really like the 24 to 70, 2.8 despite not having an IS. 

 

Hi Tokhee, Thank you for your thoughts. Do you realize that there is only a 6 gram difference between the 2 lenses you have commented on? Also, as mentioned, I'm in search of a wider angle than 24-70mm.

Link to comment
Share on other sites

Guest Ebrahim Saadawi

Answering your first questions, the C100 image is exactly 1080p, it resolves 1080p lines, as high resolution as 1080p can ever get, meaning it's exactly similar resolution to all the 4K cameras after downsampling (also give perfect 1080p) 

-The difference is, the C100 has a 4K sensor, but instead of giving you the 4K file, it downsamples it elegantly to 1080p, this means a few things, first, you skip the step of having to downsample in post, and the internal downsample is actually extremely sophisticated and occurring on raw RGB data instead of in your NLE on compressed 4K data, while the downsides are, you don't get 4K resolution, you can't deliver in 4K, and you don't get the reframing/cropping advantages of 4K. So saying it's basically as good as 4K is half true. 

-The internal AVCHD is not softer than externally recorded ProRes from that camera, it just has less blocking with moving subjects and finer colour edges only visible keying fine details. But as of resolving power, overall image, identical. It's a perfect codec size to quality ratio, somehow at 24mbps 4:2:0 Canon did it, it grades extremely well in C-Log with heavy pushing, again somehow. 

-As for why you see some soft and some sharp samples around, let's list the elements that affect this. 

The sensor (C100 in this case) is NOT the only element affecting resolution, 

it goes like this 

1- The scene/light (detailed wide fine images give higher perception of resolution and close-up shallow shots give much higher perception of softness. 

2- The lens, this is the intermediate step between the scene and the sensor, and it plays a significantly high role in resolution. Note that the 4K sensor on the C100 doesn't need very expensive glass to resolve sufficient resolution, having an EF mount most Canon lenses (designed for 20-50mp) are more than enough, some being better than others. For example a 99$ Canon 50mm F/1.8 II, stopped down to f/2.8 is eye soreingly sharp, giving an exact same perceptual resolutiom/sharpness as the Canon 50mm T 1.2 CN-E at F/2.8, while for example using the same 50mm 1.8 at 1.8, the image C100 is going to look soft, or a 35mm f/2, 18-55mm kit lens, 28-135mm, so just choose wisely by reading the available canon lenses reviews regarding sharpness/contrast, The-digital-picture.com is my one and only source. Choose a sharp lens, and it's doesn't exactly mean an expensive one. 

3- Then after the light exists the lens, it reaches the sensor of the C100, which the most vital part along side with the lens to determine image resolution, for example if  you use a tack-sharp Canon 135mm F2 L on a Canon 550D, the image will end up soft, because the sensor is soft. This is not an issue with the C100 thankfully as described earlier, 

4- After the image leaves the sensor, it goes to processing. This includes all tricks to alter the image, to simplify it's the things you can change with the camera settings, as in shooting at flat C-Log, WDR, a custom PP, altering contrast, exposure, saturation, paint settings, changing noise reduction, shooting at interlaced, at lower resolution etc, so regarding getting a high resolution, you must first of all realize the difference between resolution and sharpness. Resolution is how much information/detail that is in the image, while sharpness is a digital enhancement technique that emphasizes fine edges of the image to make it appear of higher resolution. So, shooting in Canon Cinema-Lock mode, you might be shooting a wide detailed shot using a sharp lens and shooting at 1080p, and find the image to look soft, as you chose an in-camera setting that doesn't apply in kind of sharpening whatsoever. In fact, the 6K Red dragon images raw viewed directly, 6K, look SOFT, for the same reason, it's organic resolution, the information is all there just not edge-enhanced to pop. That's a look some film-emulating DPs like, but applying digital sharpening either in-camera or in-post hugely affects the final viewer perception to clarity, thus for the normal guy watching they'll prefer the sharpened image. As a side not, also contrast amount contributes to a lesser degree in percieved image sharpness, higher contrast gives a sharper appearance and low flat contrast gives a softer appearance, try put to identical images side by side with contrast change and you'll be shocked by the preceived sharpness difference. 

-After that in camera processing, the images are sent to the memory cards, using a compression codec. Having all the previous steps done perfectly but compressing to a destructive codec, it all goes to waste. The C100 however does not have this issue, as I mentioned earlier the codec doesn't affect resolution and maintains full 1080p resolving power with minimal compression artefacts. It's a great codec with tiny files. And to remove some tiny motion artefacts and get slightly better keying, ProRes 4:2:2 can be used with a higher bit rate. 

-Then after that, the camera role is now done and it's in the closet having a few drinks to prepare for the next shoot, and now it's up to you: POST-Production, this affects sharpness enormously, again a wrong step here destroys all the previous steps completely. For example importing/exporting in interlaced, in 720p, (these are specifically hugely common when shooting for broadcast, they destroy your image due to data rates expensiveness)  exporting in a 5mbps H.264 codec, reducing contrast, applying a blur filter, anything. So the footage must be treated properly, importing in your NLE at the exactly correct settings of the file, 1080p, at correct frame rate, correct progressive scan, colour grading with care and watching for artefacts, applying sharpness when needed, noise reduction when needed, contrast when needed, and exporting the final render in NO LESSER codec than the original file to keep all the IQ intact, as the C100 is 24mbps H.264, I would export at a high end codec such as a high-end H.264, around 50-100mbps, or to be safe if one has the storage space, export in ProRes LT, or DNxHD, this way your keeping all the data in the bigger container. 

-Then, another step, which is the dreaded upload to the internet. The video sites have different compression amounts, some of them will lower your image quality and destroy sharpness/resolution. So the key is, since YT/Vimeo are going to be performing their own compression, feed it a higher than normal file quality (ProRes, DNxHD, High bitrate h.264). And if you're using youtube, youtube uses a much stronger compression with uploading 1080p files while uses a 4x better compression when uploading 4K files, so to get the absolutely best web quality, you should upscale the footage at the final stage to a high end codec like a ProRes 4K master and then upload that to youtube, now the viewers choosing 4K can see the image that the camera delivered with almost zero degradation. Be aware that vimeo offers only 720p for regular users, 1080p for paid users, and no 4K at all. So uploading at 720p will destroy all the previous work on resolution. 

-Then the last step (sorry it's getting boring, who knew resolution was reliant on so many factors) is the VIEWER. This is, unfortunately, something yo can't directly control easily. For example a user can be watching on an SD/720p panel, a user could be watching on a samsung panel with hideous sharpening, a user could be watching on a perfect 1080p monitor but choosing 720p, a user could be choosing 480p for internet speed, so your only chance is to educate people, either clients or internet audience, write down to change settings to 1080p/4K if internet speed allows to get the full quality otherwise you're not watching the real footage, giving clients high quality files and educating them on which monitor to view on, SD vs HD ones, etc. 

So resolution/sharpness is dependent on

-Scene -> lens -> sensor -> in-camera processing -> codec -> post-production -> upload/delivery quality -> Audience viewing settings/Monitor 

-That why you simply can never judge a camera's resolution on random web footage, you have no idea how the 8 steps were handeled. Only evaluate native downloaded files and see in your NLE. 

EOSHD-era are known for being very generous in providing native files if anyone asks for them.

Sorry for the long post. 

 

 

For me the lenses are important esp. with the C100. With EF-S lenses the output is very soft, only with L Lenses the Camera really shines. 

With all due respect, This is a broad false generalization actually and is incorrect. MANY EF-S lenses are higher resolution/contrast than EF/L lenses, and vice versa. 

Again evaluate each lenses independently. The C100 EF-S compatability opens the door to some killer Canon lenses released lately, for example the unmatched, tack-sharp, ultra wide angle with IS and STM 10-18mm for 250$, 50mm f/1.8 STM, the very sharp and extremely useful general 18-135mm STM IS that covers almost all needs, not to mention the higher end 17-55mm F/2.8 which is a great lens with very ''stable'' IS and shallow depth of field/lowlight. There are plenty EF-S glass that's brilliant. Take advantage of them if you own a C100, you get smooth AF with DPAF and image stabilization and extremely low cost plus full compatibility in distortion/CA digital corrections and all that stuff. I consider the 10-18mm a must have for the C100, together with the 50mm f/1.8 STM, almost free and even if they'll act as back-ups (although they're capable of shooting a high quality feature if needed). 

Link to comment
Share on other sites

Wow Ebrahim, that deserves an applause!

You raise many interesting and necessary points, most of which a known to me but still good to have your sensibilities as a refresher and for others with a taste for the extraordinary.

You know, the most endearing part for me was when you described the Canon 550D sensor as soft. It actually made me wonder if I could use one as a pillow? ;)

Alrighty... back to the front.

 

Link to comment
Share on other sites

That is indeed how it works. It downsamples a 4K sensor readout to 1080p very early in the image processing pipeline to improve SNR and color. Here's a long boring white paper from Canon explaining the finer points. http://www.cs.ucsb.edu/~mturk/imaging/Misc/EOS_C300_New_35mm_CMOS_Sensor_WP.pd

I don't know much about internal vs external recording as I haven't tried them both, but count me as a +1 for exposing to the right on Canon cameras. As long as you avoid clipping any color channels, it seems to respond well to bringing the image down in post. 

I doubt it. That applies to the C300M2, which has dual DV5 processors, so it can handle the load. The C100M2 has a single Digic DV4 processor, which is last generation. Those processors can't handle full sensor readout IIRC. They are similar to the processors on newer Canon DSLRs, except they have been optimized for video. They use the hardware encoder to produce files, which is why it's output options are the same as the XA20/25/G30 line (which use the same processor). In spite of it's "mark II" designation, the C100 has last generation technology inside and has nowhere near the capability of the C300M2.

Link to comment
Share on other sites

Guest Ebrahim Saadawi

I doubt it. That applies to the C300M2, which has dual DV5 processors, so it can handle the load. The C100M2 has a single Digic DV4 processor, which is last generation. Those processors can't handle full sensor readout IIRC. They are similar to the processors on newer Canon DSLRs, except they have been optimized for video. They use the hardware encoder to produce files, which is why it's output options are the same as the XA20/25/G30 line (which use the same processor). In spite of it's "mark II" designation, the C100 has last generation technology inside and has nowhere near the capability of the C300M2.

Even the original C100 does perform a full 4K sensor read-out and downsamples it elegantly to Oversampled Full HD in realtime, without debayering. It's a very sophisticated downsample algorithm to the raw data, though the downsample algorithm did get tweaked with the MKII models (they actually up-scale each 1080p RGB channel to 4K before downsampling to HD) 

The Canon Cameras doing a 4K to HD downsample internally are XC10, C100, C100 MKII, C300, C300II, C500, and the 1DC HDMI output in 4K mode.

Link to comment
Share on other sites

Even the original C100 does perform a full 4K sensor read-out and downsamples it elegantly to Oversampled Full HD in realtime, without debayering. It's a very sophisticated downsample algorithm to the raw data, though the downsample algorithm did get tweaked with the MKII models (they actually up-scale each 1080p RGB channel to 4K before downsampling to HD) 
The Canon Cameras doing a 4K to HD downsample internally are XC10, C100, C100 MKII, C300, C300II, C500, and the 1DC HDMI output in 4K mode.

They are not downsampling, all it is doing is debeyering to a HD frame space, probably with some pixel binning as well. There is no native 4K being converted to HD. Since you have more pixels to begin with the line resolution of output is relatively high (unlike the XA20/25/G30, which are limited by the pixel resolution of their sensor)

The C300M2 debeyers to a 4K frame space, and since it has a 4K native sensor, the effective color resolution of the output with good glass is around 1400 lines. The C300M2 uses 4K as it's native output, HD is downsampled from that. It is not the same as the C100. The XC10 on paper should be similar to the C300M2 (excluding the effects of pixel level light scatter) but is limited by the resolution of the fixed 10x zoom (which apparently provides ~80% of nominal resolution). The single processor in the XC10 probably hurts it in signal processing as well, and there are likely shortcuts being taken that don't happen with the C300.

Link to comment
Share on other sites

I hesitate but here it is .... no cinema just working on exposure and post process with these...

 

Two lenses very different character on the C100 Mk II  

Zeiss Otus 85

https://vimeo.com/139408101

Richard Gale DSO 88 FF

https://vimeo.com/139083257

 

Similar subject matter ... inside soft diffused natural light to outside Texas summer ... 

 

I find the rendering of the Canon to be wholly organic the antithesis of harsh clinical ... colors very natural and accurate. When you nail the exposure

they are wonderful ... overexpose a bit and they still refrain from evoking any sense of brittle on the edge of a break character. 

 

Such a wonderful camera to shoot ... main drawback is all the material you can capture in a short period of time ... lots to filter through in post.

These both were with the Odyssey 7Q+ in Prores HQ ... some say overkill but what a great monitor to gauge exposure and focus.

 

Sorry for the length of the files ... this is about 20% of what I captured with the camera on those outings. And I tend towards a robust file ...

not too many milky blacks .... As I normally do not find those in my subject material.  

 

Bob

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

 Share

×
×
  • Create New...