Lintelfilm

On 2017-6-30 at 9:46 PM, Andrew Reid said:

Dod Mantle can be seen with a Canon XC10 outputting 1080p to an Atomos Ninja Star, presumably for ProRes which can be edited straight off the bat. The funny thing is, Oliver Stone obviously didn’t need a very fast turn-around for the project as it was shot over a period of 2 years! Maybe the choice of the Ninja Star was to allow smooth playback of the files for the whole crew to look at during the day or at the end of the day, on a laptop, who knows…?

Read the full article

@AndrewReid The XC10 outputs 10bit 422 via HDMI in 4K mode, and paired with the Ninja Star it's auto-downsampled to 1080p. That's probably why the Ninja Star - 10bit HD rather than 8bit 4K = more bit depth and less data so the cards go further. I do the same sometimes if I know I won't want to crop in post.

By the way, was sad you never got around to posting that XC10 video shot in Italy. I really love this little camera. It has a lot of annoying 1st generation quirks in terms of usability but the image is fantastic. If the C200's supposed 400mb/s codec update is equal to the XC10's that camera will be insanely good.

I've given up on MLRAW on the EOS M. As was said above, the bitrate hack is better. RAW has too many issues with the *oh my god please go away why wont you go away* pink/focus dots - and horrible aliasing/moire with anything other than 3x crop mode. I've tried 10 bit and it's not any better. I've tried a load of different resolutions, aspects, etc but between all the artifacts, the low resolution if you want longer than 4 sec clips, etc, it's pointless. The image just doesn't deliver anything very exciting IMO. The bitrate hack on the other hand gets rid of the horrible macroblocking, ups resolution slightly and seems to improve moire quite a bit. 

As a stills camera I like it but AF is awful. The 22mm f2 is great though. 

12 hours ago, Axel said:

Actually it's just another representation of the color range, like an earth map compared to a globe. Once you get accustomed to it, it's even more intuitive to use. I always complained about the 0-100% acuity of the sliders. In the end, if fast CC is asked for, this can be an advantage. I am defending it only to this extend. Apart from the quick wysiwyg-approach, it feels too much like iMovie. They should make the tool more precise. You probably heard that the Dashwood effects (360°/VR-tools - a hint for 10.4? - , but some other FxFactory products too) recently became free, because Dashwood was hired by Apple. I installed them and tried White Balance. Too simple. Already existed within the color board (it just doesn't tell you how much in Kelvin you had shifted the balance), like the majority of third party CC-tools. 

The color board also has the advantage of being as well an effect as a clip property. If you open the color board window with cmd 6, you can jump from clip to clip in your sequence, and the Color effect is automatically applied to every clip you changed. I see Color Finale as something in between the color board and Resolve. More precise than the first, but not as powerful as the latter.

What most people forget is that the native FCPX colour tools are for colour correction. They're not grading tools and have never claimed to be - the effect is even called "color corrector". I do correction with the native tools first (I especially love the 3 saturation sliders for darks, mids and lights 0 Colour Finale can't even do that). Then I use Color Finale (largely just for the curves and often the awesome vector tool to grade and Film Convert last for final touches.

I do think Apple should add curves to the native tools though. It would save me a lot of time in the long run as sometimes I only use Color Finale for curves

5 hours ago, Oliver Daniel said:

 

I've been using resolve on and office the last 6 months - I definitely think it's better than Premiere (which I find to feel a bit ancient in terms of it's usability and UI). 

I saw in your post you haven't tried FCPX in a few years.... do you know the version available today is much much much better now? Give it a whirl!

I like the way Resolve is evolving and some of the features are unique (grading, audio editing, CDNG), however because I find FCPX so much faster and enjoyable to edit with (like the Magnetic Timeline 2) - when I edit wth Resolve I really miss the modern fluidity you get with FCPX. 

Resolve really needs more plugin support from third parties, plus a few cool tricks on timeline editing to challenge FCPX's "scary at first, but really awesome new way of editing". 

That said, Blackmagic are probably the most interesting and exciting video/film industry product company in the world right now. Top job. 

For pros turning jobs around FCPX is the bees knees. It's the Canon of the NLE world in many ways - easy to get snobby about its tech specs but in real world use its super reliable, straightforward to use and performs fantastically where it matters. For factual content and solo filmmakers it's a dream. The filmmakers you see using it are the ones consistently putting out content, not those dissecting the latest greatest tech. 

Its fast, gets out of your way and lets you focus on the storytelling. Nobody should be writing it off as a top contender for any reason now, especially not solo filmmakers.

As a mac user with pro video needs and only mid level photo needs, I do fine with no Adobe products at all.

FCPX is fantastic, especially with Color Finale and Film Convert plugins for grading. iZotope RX plugins for audio and Coremelt for tracking etc. Motion 5 is fine for graphics and FCPX has good native title options. I own logic but rarely use it.

I use Affinity Photo and Designer in place of PS and illustrator and they're more than good enough for my needs.

The one thing I really feel the lack of is a Lightroom alternative. It still boggles my mind that Apple dropped Aperture and replaced it with ... Photos. I bought Cyberlink Photo Director 8 recently to try as a photo library, which seems adequate if rather slow at loading thumbnails. To me it's weird that apple offer an excellent and well supported pro video app but not a pro photo  manager. You'd think the two would go hand in hand.

If this rumour is true (and I'm sure it isn't) the really big news is that canon are overhauling their stance on hybrid video, ditching decades of conservatism regarding releases and showing they are genuinely concerned about mirrorless domination. It has to happen sometime but Canon's MO is incremental change with hardware releases and that has basically never changed. 

14 hours ago, Eno said:

 

The V-Log is very nice if you have the skills or desire to grade it.

Not if it's the same Vlog that shipped for the GH4. My grading skills are perfectly adequate and never had a problem with the hundreds of hours of Blackmagic film or C-log I've worked on, but the externally recorded 10 bit v-log I got from my GH4 was truly abysmal. Infinite skill and patience can't do anything about the weird banding and colour artefacts. Great if it's different on the GH5 but the Cinema5D analysis didn't fill me with hope.

Panasonic can shuffle around their company as much as they like, but until they offer colour science on par with Canon and Nikon, a log profile that is useable and good video AF, i stand by my switch from the GH line to an 8bit 1080 camera (C100Mkii) and for small body 4K an XC10. 

The problem is with V-Log. It was an outright disaster on the gh4 - even 10bit to an external recorder it showed huge, horrific banding artefacts like those separated colours in the 5d article. It doesn't surprise me at all that the GH5 is the same. They simply haven't fixed v-log. I'm not very hopeful what difference high bit rate will make.

Fair enough the GH5 10 bit might be better with a non-log profile, but it's hard to get excited about 10bit with 10-11 stops DR. 

Andrew you yourself said you couldn't see any difference between 8bit and 10bit external on the GH4. Why do you think the GH5 will be any different?

xx

On 20/11/2016 at 5:14 PM, hyalinejim said:

@Lintelfilm I think with my new workaround I've found a way to shoot in moderately low light, avoid ghosting and mush as much as possible while still keeping nice colours. Check it out (this one old weird trick... Canon hates him!)

 

Have you tried this in comparison to Wide DR? Is there an advantage of shooting Standard over Wide DR?

Certainly in low contrast situations C-Log should not be used as you're stretching the codec for no reason ... but I find Wide DR a happy medium.

On 20/11/2016 at 3:04 PM, hyalinejim said:

OK, this might be a minor breakthrough in image quality. See you if you can swallow this load:

Don't shoot in C-Log, shoot in EOS Standard instead!

Am I insane? Hopefully not! I was messing about with picture styles as we all know that ghosting is dreadful in CLog but not as bad in the other styles. The problem is, many luts are designed to work with C-Log and it's just so nice in terms of delivering a range of tonality. And while the other profiles may be less mushy and have less ghosting they tend to be too contrasty with nasty highlight roll offs and the colours are different to C-Log. But, I think I've found a way to fairly closely match EOS Standard with C-Log by shooting 5 clicks down in ISO and dialling down contrast and saturation and sharpness (remember that?). Then, by messing around with the contrast a bit in post you get an image with similar colours and tonality to C-Log but it has LESS GHOSTING and a SHARPER IMAGE.

First of all here's a wide shot so you can see how they look similar:

C-Log, 1/25s, f4, ISO5000, Canon LUT.....

 

EOS Standard, (-4, -4, -4) 1/25s, f4, ISO1600, contrast and saturation adjustment, Canon LUT.....

 

Now, let's look at 100% views, C-Log on left EOS Standard on right:

Amazing right? Much more detail in the EOS Standard at 5 ISO clicks down (1.66 stops), as you'd expect. And the noise looks roughly similar (except for the subtle emergence of those fuckers, the little black dots) But the real advantage is in the effect on ghosting. We all know that the contrasty profiles have substantially better ghosting than C-Log even at the same ISO. Imagine how much better the ghosting is at a much lower ISO.

Example: According to my theory if I'm shooting an interview in a dimly lit interior and the meter tells me that C-Log wants to expose at ISO 5000 I would throw my hands up in despair as I know that the person's face will turn to mush and when they move I'll get lots of ghosting. However, I should be able to switch over to EOS Standard with sharpness, saturation and contrast turned down, knock the ISO back down to 1600 and, using my contrast adjustment, still get a comparable image to the C-Log but much sharper and much, much less ghosting.

I'm heading out now to get some more comparisons of the two picture styles. If you want to try this for yourself in the meantime, here's a work in progress LUT to get EOS Standard close-ish to C-Log. Just shoot EOS Standard 5 ISO clicks down, and bring down your superwhites before this lut:

https://drive.google.com/open?id=0B1exEpCRAfgFNGF0dUUtb2NjSTA

 

Look at how the blacks/shadow areas are rendered in the two images (the girl's hair in the photo, the squares between the crossed lines on the blanket). Then look at how the highlights are rendered (e.g. on the spiky ball). This is about contrast, not resolution. I don't know if in-camera sharpening is in the equation too (only Canon know that).

I made a post related to this on the other XC10 thread about C-Log:

C-Log is for grading - you can put that contrast and sharpness back in there (pretty much anyway - not perfectly with an 8-bit codec). The happy medium is Wide DR.

Resolution is the same though.

On 28/02/2016 at 4:22 PM, nikos said:

Ok, I did some more testing....

Xc10 c-log has less resolution than the standard (look1) profile.

Is this normal in c100 c300 c-log?

ar you ok with that?

 

On 21/11/2016 at 11:34 AM, hyalinejim said:

@Lintelfilm has a CXXX series camera I think. He might be able to take a look.

Resolution in C-Log on all Canon cameras (C100, XC10, whatever) is EXACTLY the same as it is in the other profiles.

Sharpening is turned off in C-Log by default, so that can be one explanation for the perceived difference in detail.

However - and much more crucially here and on the other XC10 thread I think - higher contrast is read by the human eye as "more detail". This is just a fact. It's an optical illusion if you like.

That's what this is about - perceived detail - not resolution. The resolution is the same.

If you want a "punchy" image with well defined lines/detail, you may be better capturing it in camera than trying to get it back in post from footage shot with a log profile on an 8-bit camera. 

Log footage is low contrast by nature, and therefore perceived resolution will always be lower unless you put all the contrast back in (and probably in the process sacrifice some dynamic range). You can't have it both ways  - particularly not with an 8-bit codec.

I always used to marvel at how much "punchier" graded HD images from my BMPCC were compared to my 4K GH4. This was because I would be able to put loads of contrast back into the 13-stop prores image without dropping down to an 8-stop result. Whereas I'd always try and coax as much dynamic range as I could from the GH4 and always end up with mushy detail. Of course the GH4 had far more resolution than the BMPCC, but stretching the lower dynamic range resulted in images that were less contrasty. Add the BMPCC's higher bit rate and colour depth to its superior dynamic range and you have what is arguably a more detailed image - certainly a richer one.

While I really do love 4K on my XC10, I think there is a real danger on forums like this of pixel peeping in such an analytical way that you forget to look at the image as a whole. Nobody watches films at 300% inches from their screen. 4K is largely a marketing thing pushed by TV manufacturers (yes I'm talking about you Panasonic, Sony, Samsung). It's an easily understandable number that can be stuck on a camera or a TV to sell them as "better". For high-end cinema 4K is certainly becoming an almost-necessity, but we independent filmmakers would be wise to remember most of our favourite films from the past several years were shot on the 14-stop dynamic range, RAW/prores, high colour-depth shooting Arri Alexa in HD/2K. None of us left the cinema saying "I'm going to watch a David Fincher film next time because he uses RED cameras so I can see more details in the actor's shirt."

Sorry, I'm having a ranty day. x

On 18/11/2016 at 5:54 PM, hyalinejim said:

Regarding ghosting, Canon say it's due to noise reduction and (I guess) that we must all bite the bullet:

I mean, does this footage look usable? That's not a rhetorical question. I'm genuinely wondering if people think you could use shots like this without the general viewer thinking wtf is that?

 

Personally I think this looks pretty awful. I haven't seen much from my XC10 that looks that bad, but if I did I'd be loathe to use it.

1 hour ago, Axel said:

Never use the built-in luts. You can't color correct the image underneath (the way you could with LUTutility on an adjustment-layer) and bring back the highlights. You can only grade on top of that lut, after it has limited the signal! 

Indeed! I never wanted to use them - in fact I didn't even know they were there! But after I updated to 10.3 the metadata from my cameras saying "shot in Canon Log" started getting read by FCPX and automagically processed. Not what I wanted at all. Particularly as I didn't know it was doing it or where to find the place to turn it off. 

The function of the Log Processing however is not to act as a grading LUT. It does have its uses. It's for editors working with log footage who want to work with a more "realistic" image prior to the grading stage. The processing would always be turned off again prior to being handed off to a colourist.

However I'd ETTR'd all my shots so the built in processor was just blowing everything.

Suffice to say my post was just a warning that Log Processing got turned on automatically for me, so watch out!

Just to let you other FCPX users know, since updating to 10.3 my Log footage ( CLog on XC10 and C100 too i think) has been displaying as totally over exposed. I thought I was doing something wrong in camera but it turns out FCP has been automatically applying Log Processing! I was annoyed but relieved to discover I could use the two days of footage I thought I'd ruined. To turn off log processing go to the info pane of the inspector and choose "settings" view. Then select "none" for log processing. 

16 hours ago, mercer said:

This is why I stopped using 100% zebras and started using 70%. At 100% I could not save any of the highlights, they were gone. But I actually have had two cameras. BH accidentally sent me a return, but I tried it out and had those results. After speaking with a few people on this site, I was told to use 70% because of the whole 18% grey card exposure rule for C-Log. I was told, if you don't use a grey card to expose (I never have) than the simplest, safest way was to use 70% zebras. When I received my new camera. I went right into using 70% and it seemed to work fine. Since then I was told to use 100% but I haven't had the camera out in a while, to give it another test. 

Mercer I think you're an FCPX user? Having looked at the documentation from apple the log processing feature was actually present before 10.3 so if I were you I'd find those old clips that were getting blown and see if that was the problem. In the inspector choose the info tab and at the bottom change the drop down menu to "settings". Set log processing to none and voila, no clipping! I bet that's what your issue was too. For some reason it only started happening to me after I updated to 10.3 but it's worth a try...

4 hours ago, hyalinejim said:

On mine, all of the info is there sitting under 100 IRE on the waveform monitor as would be expected. Is your software somehow interpreting the footage as 16-235 levels? Can you upload a clip for us to check out?

FFS!!! I was right!!! Thanks for making me look at this further. It's not the levels though.

The new FCP 10.3 now has built in Log Processing that appears to be getting applied automatically! It was detecting CLog (somehow?) and applying it's own F-ing LUT to the footage in my browser! Unfortunately this LUT is just blowing the crap out of my levels. 

For anyone else who runs into this issue, you need to go into the "i" tab of the inspector, select "settings" view from the small dropdown menu at the bottom and then change log processing to "none". Yes, it really is that hidden away and complex. FFS!

So the good news is that the two days of shooting I did last week for a job that is getting far too close to deadline were not entirely wasted. The footage is actually absolutely fine now I've taken this new auto-LUT off it! It also means I'm not going insane and can go back to my original method of exposing using the 100% zebras as a safety net. I also don't feel like a complete f-ing amateur getting home with almost entirely unusable dalies!

Jees Apple. It's a nice feature to have but maybe don't set it to on as default and then bury it in the back of beyond.

Phew. Thanks guys. I might start a thread just to warn others.

3 hours ago, hyalinejim said:

On mine, all of the info is there sitting under 100 IRE on the waveform monitor as would be expected. Is your software somehow interpreting the footage as 16-235 levels? Can you upload a clip for us to check out?

I've just been wondering the same thing. I've only noticed the blown areas since updating to FCP 10.3. Perhaps I need to see if I can tweak the interpretation levels somehow. 

I'll also look at the footage in resolve to see if that changes anything. 

Thanks!

Quick tip: I think there are a few XC10 users here using FCPX? Alex4D's free Unsharp Mask plugin works great. It's default sharpness is far too strong but if yopu apply it and then bring the radius down to about 1.9 and crank the "amount" slider up it really brings the detail out in a nice organic way. If you're using grain remember to stack it under (i.e. apply it before/top of the inspector) so it doesnt sharpen the grain too. This technique with a fine 4K grain (e.g. Film Convert's 35mm) on top of a low-radius unsharp mask gets fantastic results, great with 4K but even with HD footage too.

For best HD results: upscale HD to 4K > add subtle unsharp mask > add a contrasty curve > add 4K grain = HD that looks very much like 4K!

1 minute ago, UHDjohn said:

I could be wrong and it could be the LUMA channel but I'm assuming it's like their DSLR's which all show the clipping in the green channel only but I'm pretty sure when I had the RGB parade up and was pointing at a 24 patch colour checker I saw how it reacted. In the situation you describe, which is going to be a tough test for any camera, your best approach is to use an external monitor with scopes and make sure you get a custom WB. 

Yeah I really miss not having a waveform on the XC10. I really hope Canon add it with a firmware update.

57 minutes ago, mercer said:

This is why I stopped using 100% zebras and started using 70%. At 100% I could not save any of the highlights, they were gone. But I actually have had two cameras. BH accidentally sent me a return, but I tried it out and had those results. After speaking with a few people on this site, I was told to use 70% because of the whole 18% grey card exposure rule for C-Log. I was told, if you don't use a grey card to expose (I never have) than the simplest, safest way was to use 70% zebras. When I received my new camera. I went right into using 70% and it seemed to work fine. Since then I was told to use 100% but I haven't had the camera out in a while, to give it another test. 

Thanks yes this is what I've done. However the 70% zebras are really for exposing isolated subjects (usually people/skin) properly and not for protecting highlights (hence they disappear again after about 85% which is annoying if you're thinking about highlights.)

2 minutes ago, UHDjohn said:

The Zebras work on the green channel so with blue sky you can be not showing any clipping but be very overexposed. ETTR is not a good idea with any in-camera profile as it's an exposure strategy for shooting RAW. With RAW you get as much info onto the sensor before clipping to maximise DR and keep out of the noise and then apply a tone curve in post production. With non RAW capture the tone curve is 'baked in' to the data so if it's a low contrast scene you will overexpose if you ETTR and you have to pull down the values in post but you are moving values to a different part of the tone curve where with 8bit they can get stretched out and cause banding. Like I said this camera needs scopes ( as they decided to do with the XC15) but more importantly an RGB parade so you can see the individual channels ( not sure if they have done this with the XC15) - or use an external monitor / recorder but this kind of defeats the ergonomics. 

Thanks good to know. But in my tests it was actually the red channel that got exposed correctly and the green and blue that got blown. The zebra said the green (if it is the green it's reading) was fine. I wasn't really exposing to the right I was just trying to get as much latitude as possible as it was a very high contrast scene with dark shadows under trees and bright skies so I needed to use as much of the right and left as possible.

ARRRGH! More XC10 issues!

I've been using 100% zebras to help meter and of course guard against over exposure for a while now and it's been working great. However all of a sudden it seems to have stopped. Now, if I expose the sky in a landscape shot so that the zebras are a notch under 100% (i.e. no zebras showing but a slight tweak to up the exposure and they start showing), the sky gets completely blown - and I mean completely. No information at all! I've now tested this several times and the 100% zebras are clearly not accurate. Add on top of this there are supposed to be superwhites and I'm very concerned my unit has become defective. It's possible it was always like this but I'm pretty confident it wasn't as I've been exposing like this for a while now.

Can anyone contribute? How do your cameras respond to using 100% zeberas to expose to the right (just before clipping)?

Arghhh!

3 is the highest I'd go with the UltraCon. Personally I think 2 is the best balance but what's right depends on focal length quite a bit and the XC10 has quite a range to choose from! 

Note that the HDTVFX filter is simply an UltraCon and Digital Diffusion added together. The UltraCon doesn't diffuse at all so if you want to take a little of the digital look off the XC10 the HDTVFX is a good choice as you get the contrast reduction too. 

For documenting stuff the way youre talking about the DAF really comes into its own. Stick the very decent Canon 18-135mm STM on a DAF enabled Cinema camera and you've got a low light XC10 with pro audio, shallower DOF and controls where you need them. I have a C100 MkII and personally I think it's the best option now as you get 60p, better low light, better colour than C300 yet it's smaller. Stick a Ninja Star on if you need the C300s intraframe 422 image but for what you're describing you don't. 

5 hours ago, Mattias Burling said:

No this is not about Depth of field. It never was about depth of field. I suggest rereading my posts or just letting it go. Its not about exposure either.

A 50mm f1.8 on a s16 has the exact same dof as it does on a FF from the same distance.

The exposure is also the same.

So is the compression/separation.Switch to an equivalent lens and the compression/separation changes accordingly to the focal length.

Its simple physics.

Mattias compression is only about distance from the subject. It has nothing whatsoever to do with lens choice OR sensor size. You can test this without a camera using your eyes and your feet. Wider lenses have the ability to get more FOV in the frame but they can't magically see around or behind objects. Just as longer lenses can't magically make background objects larger.

This discussion/disagreement is about equivalent focal lengths and whether or not the same image characteristics can be recreated regardless of sensor size if the correct lens spec is used. Depth of field doesn't have to be part of that discussion but it can be and has been (I think you talked about bokeh for example).

So yes a 50mm lens at 2m from a subject creates the compression as it does on a small or big sensor. But so does a 25mm lens kept at that same distance. And therefore the only variables that influence image character are FOV, sensor size and aperture size. These variables however can't magically make background and foreground objects move in 3D space! 

Or to put it another way ...