helium

On 3/12/2019 at 2:29 PM, deezid said:

With these strong amounts of spatial noise reduction creating all kinds of detail smudging, color bleeding, low frequency texture removal etc as well as edge enhancing creating halos and fringing - nope.

Are you claiming that you can detect these defects under normal viewing conditions or just that they're visible zoomed in?  Would you, for example, bet the house on being able to tell difference in a blind test, without being able to move your chair or hit the zoom key?  If the answer is yes, please link to full rez examples, so the rest of us can test our own eyesight under the usual viewing conditions.

If that sounds like I'm being over-generous with your time, you've repeated these same claims on several discussion sites, so surely you must have examples on hand?

I tested a cheap third party battery, recording continuously for 50+ minutes with the screen at 35% (bright enough for indoors and cloudy days).  Wasn't all that hot and no difficulty getting it out.  Don't ask which one, I'd rather not offer a recommendation without more use.

Anecdotal evidence suggests that the most common battery to jam is the counterfeit LP-E6, not the acknowledged clones.  Somebody else claimed that the authentic Canons have latches more apt to jam on the BMPCC 4K, though it's possible this guy actually had a counterfeit, as reputable dealers don't appear to offer an absolute assurance you're getting the real thing. 

If the BMD battery suddenly jammed after repeated successful use, what other explanation than that it swelled?

At some point, this may get sorted out....

24 minutes ago, Emanuel said:

LOL Of course : ) Nuts now. WTH is that?? : D How have you arrived to such fine conclusion? : ))

There's no one shooting style. Yeah, right. Thanks for the information : P

Sorry for the irony. I've just found funny. Hey, I also want to apologize myself. Me included, to be here as well : -)

 

PS: A final note, though. Old school is done. Media convergence, introduction of large sensors, lightweight tools, from early last decade $100,000 to $1000 price range for S35 acquisition and so on. It is late 2018 today, folks!

If anyone's LOL, it's BMD.  They're selling thousands of "cinema" cameras, with no run and gun conveniences, to people who'll be shooting nothing but crowds, skylines, receding tail-lights and slomo gimbal reviews.  

And all ETTR.  Sounds great, from the marketing point of view.

22 minutes ago, Shirozina said:

Sorry I missed that option. When you have full control of the lighting in the scene you adjust that to best fit the DR of your chosen camera with no need to compromise on anything!

I'm sorry you did too.  What I didn't realize was that the 300+ pages here contemplate using this camera in uncontrolled conditions, and not shooting anything that *must* be exposed correctly and consistently -- you know, like faces.  

Perhaps there might be better camera choices for that sort of photography?

33 minutes ago, Shirozina said:

No the best practice is to relate the brightness range of your scene to DR of the camera and make any necessary compromises to areas that fall outside it with regards to highlight clipping and /or noise floor. You can use ETTR, histogram, false colour etc or all of them. ETTR or 'exposing for the subject' are not correct for every scenario.

Yeah, well, that's a basic presumption of cinema lighting -- you light the scene.  But you expose for the subject.  If there's no lighting, then we're back to the more haphazard approach of available light stills photography, where the best compromise is sought.

9 minutes ago, Emanuel said:

 ETTR, that is, expose to the right is the way to go with this camera. 

Only if what you're shooting is unconnected street or nature scenes and you're using it like a stills camera.  Otherwise, the best practice for this camera is to expose for the subject, as you would with any cinema camera.  

1 hour ago, tonysss said:

OT Pocket has 12,5 ,  it's about 1 stop more ?

Watch the video.  They rate the original BMPCC at 11.2 and the BMPCC 4K at 11.6.  They appear to be using the lower more conservative set of values (SNR=.5) in the chart reproduced above.

In any event, they state plainly that by their measurements, the P4K has more dynamic range.  This is consistent with what most shooters, including this one, have discovered.  

40 minutes ago, Álex Montoya said:

So the Pocket 4K fares 11.6 stops @iso400 by Cinema 5D's strict methodology, almost half a stop more than the old Pocket. That's pretty awesome.

Where did this come from?  They tested the original Pocket tested at around 12.5, which was consistent with other third party results, and the P4K hasn't been tested at all.  Or has it?

22 minutes ago, grex said:

the highlight recovery in these shots seem pretty great but deezid claims that the DR in the highlights are worse on the pocket 4K. Does this only mean that you can blow out the highlights further in the OG pocket and still retain clean data? We're not talking about highlight roll off are we? If i am able to recover a blown out shot like the one you are presenting here with the 4K, I'd be pretty happy. I'm more concerned with the roll-off.

Take it for what it's worth, but I find the performance of the two cameras nearly identical in the highlights, after grading -- meaning the same amount of data is preserved in both cases.   At least, this is the case with limited testing. 

The shots above are not actually "blown out", so there's nothing to recover.  It's just that the highlights are mapped into the super-white regions in the debayered .dng.  But the data is intact, it just has to be brought down.  You'll see the same thing on the original Pocket.

6 hours ago, tonysss said:

Yes, unfortunately ?  ,  my OT BMPCC is visibly +2 stops DR up from my GH5. I see it in all videos of new pockets, dynamic range that is noticeably down ?

It's interesting that you're able to perform DR comparisons by mentally comparing unrelated and differently graded shots seen on youtube, and are confident enough in the results to announce them.

The usual complaint about youtube "reviews" of BMD cameras is that the tuber has no idea how to grade the material, so the tests are meaningless.  But it's gone one better now:  comparisons of unrelated shots (from youtube?) done by memory.

13 minutes ago, JordanWright said:

The door can be easily removed. Pressure near the lock can tigger the door to open. Its entirely possible at this point that im being stupid.

 

No, that's not normal.  Pressure on the grip has no effect on mine.    I'd say exchange the camera, but of course you can't at this point.

50 minutes ago, Snuff said:

I downloaded the video from vimeo and saved that PNGs from the edited video.

In that case, it's as pointless trying to grade the .png as it is the downloadable video.

The few .dngs made available here by members here offer a much better example of BMPCC 4K performance, but I gather examining those frames isn't of interest to the critics here.  It looks to me, subjectively, that the BMPCC 4K has more usable dynamic range than the BMPCC, at least in part because it's so much cleaner in the shadows.   But for some reason there seems to be a strong interest in pushing another story here, and without reference to original footage.

12 hours ago, BenEricson said:

It's a 10mb 4k png still... Yes, I think a lot can be revealed about the image when raising the gamma slightly. The file is basically flat anyway...

 

I don't know where the .png came from, but even if it was exported prior to grading, exports out of NLEs typically clip out of bounds (but recoverable) data, making analysis of actual DR impossible.  The file may be flat, but it's levels are truncated.  It's not original log.

If you have grading skills, there is log/raw footage available for download, which would provide a more accurate view of the cam's DR.  But that might not lead to your preferred conclusion.

1 hour ago, BenEricson said:

Looks really nice, but when you bring the levels back up, you start to see those highlights that I keep seeing in the P4K footage. If you add a bit of a shoulder, you lose detail but it smooths out the highlights. You probably lose a stop of DR as well. 

I'm not a colorist, but the c300ii I use often looks closer to the version with the curve, right out of the camera. I find the OG pocket to similar in that regard. The highlights have a lot of roll off.

 

 

This thread is aggravating enough to revive the dead....  You do realize you're grading a grade, not camera original?  There's no way to know how much original highlight detail was in the shots, because current levels are baked in.  It's like trying to grade a bluray and concluding that 35mm stinks, because you can't push the highlights without clipping.

50 minutes ago, mercer said:

So, I’m not really invested in this conversation since it is doubtful I will buy a P4K for a long while, but here’s my 2 cents... 

There are some flaws with this camera that didn’t exist once the OG Pocket matured via FW updates. Those FW updates fixed the original camera. So, I know everybody loves their new P4K and don’t want to mention any downsides but they are there so ignoring them won’t help in the long run. We already know that people from BM check this post, so if you have an issue, scream it from the rooftops, so BM will address it with a FW update... they’re really good at that.

Ben has some valid points. This camera should theoretically murder the GH5, but from what I’ve seen... it doesn’t. The DR is similar or only modestly better. The GH5 actually handles the highlights as well or better and I am seeing something funky with the color... specifically with greens... I don’t know if it is a slight IR pollution?

All in all, I am impressed with the P4K, albeit a little clean and pedestrian for my goals, but this isn’t our spouse we’re talking about here... it’s okay to mention its shortcomings.

So you're not really interested in the conversation, but simply must offer up your two cents anyway, which is that BMD had better drop everything until the BMPCC 4K murders the GH5 -- murder coming up, next firmware release! -- because you're not satisfied with the difference between the two cameras, based on what you've seen on youtube.  

Between a host with all the self-control of 5 -year old, and a readership which evaluates cameras based on hearsay and computer monitors, this place is really something.   

In other words, another one bites the dust.  Goodbye, and good luck.

9 hours ago, BenEricson said:

I can’t believe that the highlights clip that hard on a clip so under exposed. The handling of highlights are what kills this camera for me. The OG pocket had so many flaws but the image was not one of them. 

I think the GH5 might have an upper hand with the baked in color profiles as well. I always impressed with the files straight out of camera with then GH4. Has anybody been messing with shooting non log footage? Can you bake in LUTs to the BMP4K? 

 

It's raised beyond the point of no return, an acceptable grade would require more attention, but are the highlights any better for you here?  

41 minutes ago, jonpais said:

@helium Actually it is possible to say the poor GH5 results are due to incompetence, because the skin tones don't even fall on the vectorscope skin tone line - they're far too magenta, and far too washed out.

 

Putting skin tones on the line would be a grading choice.  Normalization is a different process -- it takes all the log values and converts them to rec. 709.  If this process wasn't done correctly, skin tones would be off, as would everything else.   But the fact that skin tones aren't on the line doesn't in itself prove that the normalization is off.  The trouble here is that there's no way to evaluate the comparison.  Skin tones may be that bad, pre-grade.  Or they may not.  GH5 skin tones usually incline to green.

53 minutes ago, A_Urquhart said:

The same person would have graded the shots in that video. If they are amateurs , this shows that the Pocket 4K was easier to grade for an amateur. Not a bad thing?

It may not be a bad thing -- if in fact the BMPCC 4K can survive incompetent grading -- but based on their only comment on the matter, it sounds as if they "normalized" the log footage by looking at the vectorscope and making adjustments -- not a promising or reliable workflow, for normalizing log footage.  And that was apparently the extent of the grading they did on both shots.

Without more information, it's impossible to say the poor GH5 results aren't the result of incorrect handling. 

1 hour ago, jonpais said:

And if a person is such an authority, they should be able to convince with facts - not make up stories about Sareesh not disclosing information about the camera he's using. 

I assume you're referring to John Brawley, who *is* a recognized authority -- he gets paid to shoot, rather than blog -- because I said nothing about Sareesh's contractual arrangements, not being privy to them, if any.

That aside, this place is too damn funny.  Here we have three guys defending the wisdom of a blogger, while they by their own implied admission know nothing about the log post-production process.  Normally, you'd expect folks to remain silent, under those circumstances.  But of course this is the internet....

20 minutes ago, webrunner5 said:

I would venture to say Sareesh probably has a better Grasp of what video is all about more than Anyone on here, and probably the vast majority of people in the world. Is he always right, nobody is, but I will take his opinions as real truth.

And who has a camera longer than 2 weeks that reviews anything other than their own personal camera? That is 336 hours. Surely a person can figure out a most things in that amount of time. It's not like the first camera that BMD has ever made. It's not like the sensor is new either. It's pretty much a dumbed down GH5s with ProRes and their UI and color science. Not that that is a bad thing. Especially for the money.

I think I would rather have a GH5s and the new Atomos Ninja V. But that is more money obliviously. Add a Gimbal , a Speedbooster and not a lot you Can't do with it.

Well, that settles it then.  Case closed!   Sareesh is a world-class authority, by decree.

The proof that most users, including Sareesh, never figure it out is as plain as the poor grades and senseless test comparisons (in Sareesh's case) which are routinely uploaded, with apparent pride.  336 hours is nothing.  This has been going on for years.   You still have people offering up uncorrected log footage from their BMPCC, and expecting to be admired for it.

Then again, if you have no intention of either buying or using this or any other camera ("I would think"?), it won't matter much.   Pick your authorities as you prefer.

19 minutes ago, jonpais said:

Funny, because Blackmagic considered Sareesh to be knowledgeable enough to entrust him with a loaner. Now suddenly he doesn’t even know the basics of photography?

John Brawley says over at bmcuser that Sareesh failed to disclose that he had a pre-production unit - but in fact, Sareesh discloses in the very first minute of his review the firmware version he’s using and that he fully expected many of the issues he experienced would most likely be resolved before the camera ever reached retailers’ shelves. That sounds fairly even-handed to me.

755 viewers have found his review helpful and the overwhelming number of the 255 comments on his video praise his work. But a handful of forum members are going to smear his name because they don’t agree with him. 

I suggest you repair to the BMD cinematography forum, and read the discussions initiated by Sareesh, where he betrays astonishing ignorance -- even venturing to lecture BMD staff (Captain Hook), among many others who patiently tried to explain to him the basics, to no effect.

Only BMD and your deity of choice know why Sareesh and other bloggers were chosen for advance units.  With the exception of the Proav UK footage, from a dealer, which didn't try to do much other than set levels, the grading was pretty terrible, even by amateur standards.  Maybe their skill level was deemed to reflect the average Pocket buyer.  Alas.

2 hours ago, seanzzxx said:

He didn't even bother to switch out "GH5" in some of the descriptions but he'll explain to you how this camera works after having owned it for two weeks.

It's worse that that; Wolfcrow is woefully misinformed on basics like log shooting, LUTs, normalization and dual circuit gain.  He has no grasp of what the camera requires in post.  His reviews should be limited to point and shoots.

4 minutes ago, webrunner5 said:

How the hell do you know how I vote! See your doing the same shit your are complaining about. Making someone guilty with out proof. You are the F ing Idiot.

A characteristically civil response.  I don't know or care how you vote. The subject, which you raised, was the current Supreme Court nomination.  Read the post, why don't you? 

But, in any event, thanks for additional confirmation that nobody in her right mind would ever consult on these questions, any more than for camera evaluation.

1 hour ago, webrunner5 said:

He said she said. Kind of reminds me what is going on with this US Senate thing. I bet everyone on here could go back to a girlfriend, a boyfriend in the past, and if they dated a reasonable amount, their X, would all have some person say he, she was a horse's ass. I have no clue if it happened or not. He does not come across as the most laid back person I have ever seen, he is a go getter, a make shit happen guy, and they are not the type to be huggy, kissy, yes dear types. So I am sure sparks flew as they say. Who knows, but to just throw him under the bus with no proof, well pretty week minded if you ask me. Strange ass world we live in now. You people accusing him don't know him from dog shit.

The old "He who is without sin" thingy.

I was going to say, given your views it's easy to see why nobody would ever consult you  on any Supreme Court appointment, but of course the most powerful interests in the country agree with you, at least with respect to this nominee (don't expect them to make the same concessions for a liberal), because they know he'll rule for them.  

It must be great pretending you're a fearless independent thinker, when your opinions just happen to coincide with big money interests.  Then again, the U.S. has the distinction of being the only country in the world where 38% of the population thinks its either in the top 1% or soon will be, so these allegiances are not surprising.

3 hours ago, Dan Wake said:

why it's not smooth? its shaky....

is this camera too heavy?

The operator doesn't know how to move with it. There could also be balance/tuning issues.  The lack of basic operating skill evident in the video doesn't suggest a high degree of competence in the setup.   There are also fundamental limits to smooth gimbal operation without stabilizing arms.  Most gimbal review videos avoid showing how bad it can be, and how bad the operator is.  This explains brief cuts, shots in open areas and lots of slow motion.

In a word, another useless youtube gimbal review....