The ghost of squig

6 minutes ago, Anaconda_ said:

@The ghost of squig - What he's saying, is the macro blocking in blue skies is only present if you have your project set up incorrectly, and the blocking you see in Resolve won't be present in your final export. It sounds to me similar to turning on Proxies, the software plays a lesser quality version of the file for smooth playback while editing... Before you rule out shooting blue skies, why don't you specifically shoot them, adjust your settings to fit your file and then do your own tests.

The screenshots I posted are from the final export.

5 minutes ago, Emanuel said:

Will this be effective?

Nope. Like @deezid said: "I really hope BMD adds a setting allowing spatial noise filtering to be turned off. In this example the grain would become smaller at least and less blocky."

Otherwise I can see a Z6 with ProRes raw landing on my desk sometime soon.

"This should not affect renders"

Looks like I won't be filming any blue skies with my Pocket 4K, which is a shame because we get the occasional blue sky in Australia.

5 hours ago, deezid said:

These blotches exist in both ProRes and Braw footage from the camera and are caused by its internal noise reduction - which cannot be applied to cDNG which is therefore not affected.

It's a good thing I got one of the new not defective in any way cameras that can't be downgraded to CDNG then. I didn't pre-order because this time I wanted to wait until all the bugs were fixed ??

9 minutes ago, thebrothersthre3 said:

So as long as good cards are used there should be no issue? I had an important clip get corrupted on my XT30 using a not so great card, multiple times. Never happened with Sandisk. If thats all it is I am on board. 

No issues to be concerned about that I'm aware of. I've been using the same 4 Lexar and Toshiba 1066x 64GB CF cards for years. The only way to ensure against data loss is to shoot with a camera that either records to 2 cards or records internally and outputs a clean feed simultaneously.

8 hours ago, thebrothersthre3 said:

My only worry with using a 5D mk3 or any Canon with magic lantern is the unreliability. I've heard of people losing all their footage, which of course is not acceptable for pro work.

Right now I have my XT30, Black Magic Pocket OG, and Pocket cinema Camera. Might consider selling the XT3 for a Pocket camera 4k.

I am tempted to get a 5D mk3 but again Magic lantern kind of scares me.

The only instances I'm aware of where all footage has been lost is due to card failure, which is just as likely to happen with any camera. This unreliability thing is a myth perpetuated since day one of Magic Lantern. Sure nightly builds have had some bugs, but anyone would be crazy to shoot a job with an untested nightly build. The nightly I'm using has been around for a year, and it's solid. I've been shooting with ML for over 7 years, and I've never had any worries about its reliability.

8 hours ago, Geoff_L said:

Yes, I do not know what to think about the reliability, as I came across different opinions. Some warn everybody after they lost their footage... others say they never have any troubles ! But I think I read more of the latter.
Interesting that you keep your XT30 and not the XT3 . Both are on my list, and If I go the bmpcc4k route, I'll need a still camera !

I follow the ML forum closely, it's definitely the latter.

I'm not taking any stills with the X-T3, I know I'll want to keep it, but I don't make my living shooting stills so it has to go.

7 hours ago, Geoff_L said:

XT3 + bmpcc4k would be a great combo too... less cumbersome than BM + 5D for my needs. But probably out of budget. With that being said, I really like the tokina 11-16 f2.8 and the sigma 18-35 f1.8, and both can be shared between the XT3 and bmpcc4k. Plus, the budget could be less stretched than buying the same focals for a full frame...

If you're mostly shooting daylight and 5600k, sure.

6 hours ago, webrunner5 said:

I know  a lot of people do it, but I would Never depend on ML for anything I can't repeat easily. Ergo, A Wedding, Baptism, Sporting event, on and on. I am a big fan of the 5d mk III but not if I was trying to make a living using ML on it. It is too unpredictable. Sure at low resolutions it works, but why even bother as a Pro doing that.

It's just code man, not voodoo magic.

8 hours ago, thebrothersthre3 said:

I'd love to see some comparison footage. 

Wasn't expecting the Pocket to be significantly better than the Fuji. Its a very tempting option right now along with a hacked 5D mark 3. How does prores compare to the HEVC on the Pocket?

https://drive.google.com/open?id=1bTrb4QlNxZRQW2dZPMQ5_7U5b0jMxrQ5

I haven't shot any ProRes, been quite content recording Braw to a T5 drive and UHS-1 Sandisk extreme pro cards. The P4K pairs better with the 5D than the X-T3, much easier to match in post. Keeping in mind I've been working on doco with no control of the lighting, the X-T3 would do a lot better with controlled lighting.

8 hours ago, Geoff_L said:

I am exactly thinking the same... For still + video, could be a good combo. Lenses and batteries can be shared, even if it appears to be better going external on the BM. Given the prices of the used 5D, it roughly makes the pack at the same level of the others hybrids (if keeping the BM accessories to a minimal).

I wonder something, maybe @The ghost of squig has an answer : I see everybody rigging the bmpcc4k (with quite big rigs for some), which is, given its purpose as a cinema camera, a logical thing. But, what if you try to use it with the minimal accessories, e.g only a smallrig cage (+ a t5 ssd, or only a fast sd card) and a little mic ? I believe the battery life is only 20min, but if you have 4 or 5 canon batteries with you and a powerbank to charge them in an emergency, could it work to avoid the big external powering solutions ? I know it depends on the shooting style and the type of work, but, still, I wonder ?

I wouldn't bother with LP-E6 batteries, the run-time is too short. I haven't bought a cage, I have it rigged to 15mm rails with the Sony L battery hanging off the back of the rails and the T5 mounted to the rails under the lens. If you're booming you can mount the L battery on the top thread. I have v-mounts and powerbanks, but I'm satisfied with the Sony L setup for now. It's probably the cheapest option, and I only have to charge 2-3 of them at the end of the day.

7 hours ago, webrunner5 said:

I see no Macroblocking on any of them if I expand them, probably not a fair judgement. But I see that the Canon 5D mk III is killing it on DR if they are all exposed the same? But your links are hard, to impossible to open on my computer? Appreciate your work on the comparisons.

I recorded all 3 at different exposure levels, the ones you can see were the optimum exposures for DR. The 5DMk3 CDNG raw format gives it the edge in the grade.

Have you tried dragging and dropping the images?

6 hours ago, mercer said:

Yeah, it’s pretty amazing the fine detail the 5D3 is picking up. If this was just a comparison between the P4K and the X-T3, I’d have no idea the house was stucco. 

Not bad considering the MK3 is 1080p and the other two are 4K.

5 hours ago, thebrothersthre3 said:

The Pocket 4k and XT3 look really similar in that test. The 5D stands out with more contrast.

Zooming in the Pocket shadows look the least pleasing, less color and more noise. The XT3 is second being cleaner, a bit more color but its still muted. The 5D isn't as clean as the XT3 but the color is more vibrant. 

The 5D is definitely tempting me right now. But I am unsure how reliable Magic lantern is and of course you don't get 4k 60p or 120fps. 

One of the things I don't like about the X-T3 is it loses chroma in the shadows, that's something the 5D has always handled better than every other camera I've tested over the years. I recently filmed 3 jobs exclusively with the 5D over 3 weeks with a nightly build; it only crashed once, and I only lost one frame. The 5D upscales to 4K nicely.

4 hours ago, mercer said:

I don’t want to clog up this thread praising the 5D3... most people probably know my feelings about it. But I have never had any glitches with ML Raw on the 5D3. Also, even though 4K(ish) 24p is possible with the 5D3, I don’t really consider it usable... you may be able to get some static wide shots at infinity but for the most part, don’t count on it.

We should be doing this in my comparison thread. My bad.

10 minutes ago, Emanuel said:

In that thread above-posted someone claims cDNG is the worst sample when some other notices that's make little sense from lossless... ProRes is reported as identical.

Anyone with the OG Pocket have any issues with compressed CDNG or ProRes macroblocking?

47 minutes ago, Emanuel said:

I believe this will depend on our grading too.

As far I've seen posted cDNG is even worse with Blackmagic cameras but should it be going lossless?

ProRes 422 HQ should be clean, at least. If not, there's nothing related to codec or RAW implementation.

 

It's surely not good news...

It's definitely compression artifacts. Grading accentuates it, but even ungraded it's noticeable. Any large monochromatic areas could be problematic. I haven't tested ProRes, in my experience ProRes is less prone to macroblocking like that, but I'll do a test anyway. I don't remember having any macroblocking issues with CDNG 3:1 on the BMMCC, but I didn't use it much. Lossless would definitely solve the problem, adjusting the variable compression algorithm might help too.

This is how it looks rendered out to ProRes4444. Braw up first, followed by the X-T3 and 5DMk3 raw: https://drive.google.com/open?id=1bTrb4QlNxZRQW2dZPMQ5_7U5b0jMxrQ5

After a week on the the road with the 3 cameras shooting 5 events over 3000 kilometers I got a good feel for the 2 newbies vs the 5DMk3 under a range of uncontrolled practical low lighting conditions. It didn't take very long for me to decide to ditch the X-T3. I love the design, and it can produce a good image, but under those kind of conditions HEVC 10bit is no match for 14bit lossless raw or Braw. Even custom white balancing with a grey card was hit and miss, and slight underexposure was disastrous. I did love the OIS lens, so much so I'm getting an OIS lens for the Pocket 4K. 

I just posted some dynamic range test pics here:

With the Viltrox speed booster, the Pocket 4K has a 1-2 stop low light advantage over the 5DMk3 raw, which has been very handy. I did see some moire on some grey blinds, and there's the macroblocking in blue skies issue, but apart from that the Pocket 4K has been behaving well. I used a Sony L-series battery to Canon LP-E6 dummy battery, and a generic NP-F970 battery lasted around 1.5-2 hours. Interestingly, the batteries still had a 60% charge when the P4K was ready to conk out.

I really hope Fujifilm get it right with the XH2 or X-T4 with ProRes raw or some other flavour of raw over USB-C or to internal XQD/CFexpress cards. Perhaps Fujifilm could do raw to USB-C on the X-T3 with a firmware update?

5 hours ago, JordanWright said:

Can't say i've ever seen macro blocking when shooting a blue sky but maybe I haven't been looking for it, here's 2 3:1 braw clips on my hard drive and they look good to be with a fairy aggressive grade. They might not be the best examples.

Those examples look OK, but it's only been a problem with clear blue skies. I did a test with the P4K Braw, X-T3 HEVC 10bit, and 5DMk3 magic lantern 14bit lossless, graded in Resolve and exported to ProRes4444, only the P4K shot has the macroblocking in both 3:1 and Q0 at varying exposure levels. It may not look so bad on a 27" screen, but I wouldn't want to risk it on a 40 foot screen. These are screenshots taken from the ProRes4444 file.

Pocket 4K

X-T3

5DMk3

7 hours ago, Emanuel said:

That's new... Really. I've never seen anything like that. It is more about any potential artifacts of aliasing or moiré : ) hardly to be found in motion to be sincere. cDNG is also more detailed and better IQ overall : -)

I find it amazing it's been over a year and nobody mentioned it, I noticed it the first day I got the camera. I'm fine with the detail of Braw, but the macroblocking is a worry.

5 hours ago, MeanRevert said:

There's a whole thread about that Blue Sky thing on the BMD forum:

 

https://forum.blackmagicdesign.com/viewtopic.php?f=2&t=93737

Thanks for the link.

4 hours ago, JordanWright said:

Yeah, my eye isn't good enough to distinguish between them, I prefer the workflow of BRAW but I totally understand why people prefer CDNG. Most of the time Im shooting in ProRes anyway, only using BRAW for bigger projects at the moment!

Try shooting a blue sky, Braw is macroblock hell even set to Q0 and 3:1. Braw has looked good with everything else I've shot, but I'll have to rely on my trusty 5DMk3 with magic lantern for anything with a blue sky in it.

1 hour ago, webrunner5 said:

We are talking Raw over HDMI. Now SDI sure. I am really not sure they are actually getting true 14bit video on ML. I doubt back then Canon converters were really that good, especially on the cheaper models.

"14bit raw and 12bit raw on the 5D Mk3 is imperceptible"  That statement sort of highlights my assumption. There should be a noticeable difference, especially in video..

This debate was settled a long time ago, but just to recap: ML raw is true 14bit raw, you can open a ML CDNG file in ACR and manipulate it just like a Canon CR2 raw photo file, the DR is identical, and there's no noticeable loss of DR in 12bit. More bit depth is good, but there are other factors that determine available dynamic range: https://www.magiclantern.fm/forum/index.php?topic=11097.0

I haven't worked with ProRes raw, how well FCPX recovers highlights and shadows on ProRes raw files will be a bigger determining factor than whether the Z6 is outputting 12bit or 14bit raw video.

6 hours ago, webrunner5 said:

I would guess 12 bit Raw, which is damn good don't get me wrong, is the best they are going to do with the Z6. You are not going to match Photo DR, just like @androidlad said.

Why not? Magic Lantern raw matches photo raw dynamic range. The visual difference between 14bit raw and 12bit raw on the 5D Mk3 is imperceptible.

18 minutes ago, Papiskokuji said:

Interesting read ! Even more so that I have the same set of cameras !

I share most of your thoughts although there’s a slight contradiction in what you say about the low light between the 5D and the XT3. You say at 3200 it’s the same between the two but the 5D has fixed pattern noise. So it’s not the same at all ? ! And I concur ! I love the 5D raw image quality but fixed pattern noise shows quite easily ! And the Fuji has also the edge in termes of dynamic range. That being said all three cameras have a wonderful image quality. They all have their pros and cons. In low light, the bmpcc4k is great ! Even without noise reduction. It’s enough for me. With noise reduction, I’d even say it performs better than the X-T3 thanks to its more robust codecs.

Maybe I’ll write a longer report about all those cameras so we can argue a bit more

Hehe, bring it on.

Actually it's the Pocket 4K and 5DMk3 that are about on par for noise at 3200 ISO, so far the X-T3 noise is very grain-like, and looks good to me up to 10,000 ISO with NR set to -4.

15 minutes ago, Andrew Reid said:

The Fuji 18-55mm is a stunner on the X-T30 for documentary use though. It feels small enough to be balanced with the camera controls, and it's not always trying to escape forward and downwards from your grip. The OIS is amazing. The AF is amazing. It is even parfocal and you do not see it trying to refocus during a zoom. So you can do the odd zoom or even fast crash zoom and focus stays nailed on the subject the whole time.

It's a great little lens, too bad it's not f/2.8 throughout the range. Great bokeh, but a bit lacking in character for my liking,. I'm gonna try the Black Promist with and without the Iscorama. OIS anamorphic, how cool would that be

35 minutes ago, ade towell said:

According to Mattias there is no rolling shutter with the EOS-r ?

?

4 hours ago, Mattias Burling said:

In s35 mode on the EOS-R I don't think you could get five out of five on a blind test next to the C100mkii.

I did some side by side with my BMMCC as well and the R has more detail and almost identical DR to the BMDs Prores HQ.

Personally I use it in full frame HD for mini docs all the time. It's either going online for social media or projected on big movie screens. So no need for anymore detail. I always have my BMD on standby but so far I just keep using the R. 

 

The 20ms rolling shutter on the 5DMk3 is a little too high for my liking for shoulder-rigged doco work, the EOS R has more rolling shutter than the 5DMk2 (25ms), way too much for my liking. Are you using an OIS lens with the R?

One of the things I like about the X-T3 as @andrew_dotdot mentioned, the 1080p resolution is on par with the 5DMk3 raw, and the 1080p rolling shutter is only 12.6ms, slightly lower than the BMMCC.

After a couple of days playing around with the 3, I can get them all to match up pretty well with Filmconvert and a bit of tweaking, sometimes I couldn't tell which was which. What made the P4K and X-T3 shine was my old Leica 35mm Summicron-r, it really helps take the 4K digital edge off, and adds some nice flare, not always welcome on run and gun doco work, but that's why I carry a mattebox. I've got a 1/8 Black Promist filter on order, it will be interesting to see how that works with the 18-55mm Fujinon.

What did surprise me is the X-T3 is cleaner and sharper than the P4K at high ISO, my Viltrox speed booster hasn't arrived yet, that extra stop of light and S35 crop should even things up a bit. At 3200 ISO the P4K looks to be on par with the 5DMk3 noise wise, albeit without the fixed pattern noise.
I'm shooting some stuff for a doco next week where the P4K will be on sticks shooting a 2 hour 12:1 Braw master, whilst I move around a dining table with the X-T3 shooting FHD on a monopod to get close ups.

It's been overcast here, I'm waiting for the sun to come out so I can do some definitive dynamic range tests. The 5DMk3 still has the edge in terms of post manipulation, Braw doesn't have as much highlight control. 10bit HEVC and Braw are quite close in post.

The P4K and the X-T3 both lose saturation and detail in the shadows, the 5DMk3 does better there, but it's noisier in the shadows. I saw a test on the X-T3 vs the Nikon Z6, the Nikon did much better than the X-T3 in the shadows, if only Nikon would bring out an update with a good internal 10bit codec, with an XQD or CFexpress card, it could do internal raw ?.

9 minutes ago, andrew_dotdot said:

Is that old one the "95MB/s" card? -- I have a 64GB one of those, which I was pretty sure worked with all codecs, but I didn't dare put it on my list, since I'm at the office and the card's at home and I wasn't 100% sure it was the 95. If that's so, then I can agree with you, it has worked for me on the 400Mbit/sec codecs.

Yeah it's the 95MB/s write card. Interestingly, Sandisk lowered the specified write speed on the 170MB/s read cards to 90MB/s write. 400Mbits/s write is only 50/MB/s so both cards have plenty of headroom. The faster read speed will be welcome ?

I've been using an old Sandisk Extreme Pro 128GB SDXC UHS-I with no problems recording 400Mbit/s HEVC. I just ordered 2 of the newer 170MB/s read ones, glad to hear they're working OK.

24 minutes ago, mercer said:

To everything... there is a season.

Although I completely agree with you, I think we’re spoiled in a lot of ways and I’m noticing it more and more as a newer generation of shooters show up on this forum.

I’d love 16 stops of DR and DPAF with internal Raw and/or ProRes but those specs won’t make me a better filmmaker. I’d love an IBIS mirrorless that has the latitude of an Alexa Mini... but then what would I strive for and work toward?

It's all about having the right tool for the job and efficiencies. On a 30 day Alexa shoot, a 1st AC would add $8000 to my budget, including food, accommodation, travel, etc.A gaffer, that's another $8000. Steadicam operator for that Alexa? There goes another $10,000. All fine when you have a 500k budget, not so much with an 80k budget. Tech like DPAF, dual ISO, and Braw makes lower budget projects with high production values more viable.

8 minutes ago, mercer said:

Yes, please post some side by sides when you get a chance. The X-T3 has a nice look to it and I love that it shoots cinema 2K at a hefty bitrate.

Will do. The X-T3 does look good, but like every camera I'm discovering its quirks which have me thinking I'll probably have to shoot with it and the Pocket 4K to get everything I want to film shot. Unfortunately we're still not at the point where there's a sub $10,000 camera that can do it all. By all I'm speaking in terms of internal raw and DPAF in particular.