pocket camera

just got a pocket cinema camera coming from a panasonic lumix, what would cause bricks like a house to have movement to them? im a bit lost on this, i have tested raw and prores hq. thanks

Some of it will be moire/aliasing BUT I found that it's mostly from the reduced size from the resolve player, try export a small portion and see if its gone.

Also you can go into the settings and change the playback options from sharp to smooth to fast and there are also other options like bicubic etc, try different ones untill you find one that looks best to you.

thank you for your reply, i dont see the fast, smooth options under playback, but i will have a look again, thanks

after exporting it is much better! thanks alot, the only other issue i seem to have is highlights. if i have a scene with buildings in it and you can still see the sky when i first load the "raw" footage you can see the clouds but once i convert to rec709, bmd or grade from scratch the clouds are gone, i have tried alot of things, and watch some youtube videos but no luck, this is really only a issue if im shooting towards the sun,you can still see clouds in real life. im confused because i thought this was a big plus for shooting raw, i use a ND filter and expose so there is no zebras. any tips? i figured i would add this to my post as i dont want to clog up the forum with a bunch of questions, thanks

On the bottom left in resolve in the tray where you can select the colour wheels etc the left most tab is camera options. You can select the raw settings of the file, change it from project to clip and then you can tick the highlight recovery box. Otherwise the only think I know to do to drag the highlights down, if nothing was clipping then the information is there.

The information is there.


Sent from my iPad using Tapatalk

yes i use the camera raw and use clip. i have tried clicking highligh recovery and not clicking it, im new to raw i was just making sure i was going about it the correct way. i will try a few more shots this weekend, going off the waveform nothing is near cliping but was still not satisfied with the highlights "the sky" i use panasonic lumix 12-60 and used a ND filter and adjusted apature till all zebras were gone. is it possible that i was underexposed? thanks alot for your replies.

Depends how you set your zebras. If set to 100 and you just step back a fraction from the point where the clouds are clipping, you should be fine.

Brian, try setting zebras at 90% for a little extra protection of your highlights. With 'highlight recovery' turned ON and use 10bit wherever it's an option in the Project Settings in Resolve and young should be fine. You can set your output to Rec.709 in Resolve, but when you capture video in camera, use either the best ProRes Film or raw recording to increase your dynamic range. Don't capture in ProRes Video.


Sent from my iPhone using Tapatalk

thanks, i might need to add i shoot at iso 200 i read some where that iso 800 is best for dynamic range? im also only using a variable ND is a IR cut a must? if so which one should i go with? im going to go out today and set zebra to 90 and see what i get, thanks

Brian, yes, the 'native' ISO is 800 and that can get you better graduation. If your scene is dominated by highlights, it can be a good idea to shoot at ISO 1600. That will actually get you one more stop above middle grey at the expense of one stop below. You might like that if you want to capture clouds or highlights on your subject with more apparent fidelity.

Either an IR Cut or IRND filters or SLR Magic's Image Enhancer Pro is necessary depending upon the scene and subject. If you are only using a variable ND, you may find you don't need any infrared filter for one or two stops.

There are no hard rules. It depends on the strength of infrared radiation in the scene and the type of materials and clothing present. If there's black synthetic attire, use infrared protection as it likely will not be black.


Sent from my iPhone using Tapatalk

yes for now im only using a variable ND i have tried a ND with a graduated ND as well, i dont have any knowledge on IR cut. I know it stands for Infarred but thats it, i will look into picking one up and trying it, thanks alot for the help! i really appreciate it

You might want to search here for threads discussing Infrared Cut filters and IRND filters that attenuate (reduce infrared wavelengths to the same degree they reduce visible wavelengths).

IRND filters cost more than ND filters as they are more difficult to make and they may do the job adequately without cutting infrared. Some IR Cut filters may come with visible tints that you need to correct in post production.

I like the SLR Magic Image Enhancer Pro because it reduces infrared wavelengths beginning about 700 nanometers whereas most IR Cuts such as the Hoya UVIR Cut reduce around 680 nm and that might affect deep red and skin tones.

Filters that rely on a dichroic coating on the glass may also Cut more than you think if you are using very wide lenses. So if a filter is rated at 680 nm for modest wide and longer lenses, when you mount a very wide lens, it will begin cutting a little sooner. Unfortunately I don't have a very wide lens and the spectrometer to demonstrate this.

Most times though you'll be fine with an IR Cut like the Hoya.

looking at them, is a cpl the same as a ir cut?

CPL? Circular polarizer? That's a method of constructing the polarizer which can be circular or linear. For camera optics, you want to use a circular polarizer.

Look up 'Circular Polarization' and 'Circular Dichroism' in Wikipedia just before bedtime. The explanation will help put you to sleep.

Does not have any utility as an IR Cut.


Sent from my iPhone using Tapatalk

I got the Firecrest UV/ IR Cu filters for use in my Micro acsmera (same sensor as the Pocket camera), works great with the least amount of additional color cast. Some IR cut filters add a slight green cast depending on where in the color spectrum they start cutting IR, which can also remove some red, creating a lean townards green.

Cheers

We have four filters usually recommended for use with BMD sensors. If your filter isn't listed here, it may be because it tints the image you see on your monitor more than you would want:

Hoya, Firecrest, Heliopan, Schneider.

There are other specialty vendors such as Breakthrough that may offer good products but they're not yet mainstream. If anyone feels they have a great filter, let us know and we can add it to the list.


Sent from my iPhone using Tapatalk

If your scene is dominated by highlights, it can be a good idea to shoot at ISO 1600. That will actually get you one more stop above middle grey at the expense of one stop below.

Did you actually try that with a BMPCC? My experience tells me you got it backwards. You'll lose about a stop in the highlights.

Also note that even with zebras at 85%, highlights showing no apparent clipping may still prove to be all but useless, for detail or much in the way of adjustment.

No, John, I don't have a BMPCC. You may be right. My advice was based on using the URSA Mini 4.6K so I apologize if that behaviour doesn't apply to the Pocket camera. Something Brian might try as he learns how to shoot highlights with the camera. Thanks for the feedback.


Sent from my iPad using Tapatalk

The other funny thing -- and this is heresy -- is that in low light situations with the BMPCC you get much better results in video mode, than either raw or Prores Film. But you didn't hear it from me.

No John, Rick,is correct, the exposure curve is very similar between the UM and the Pocket, both being "Cinema" sensors, with a similar exposure setup , so increasing the ISO does shift the exposure latitude towards highlights, reducing the ISO shifts it towards shadows. And yes, this seems backwards, but when shooting Rawmor FilmLog, (not Video move), this close to Adam's Zone system, of taking an existing fixed exposure latitude, and shifting it up and down towards highlights or shadows to get the part of the subject you want "properly" expsoed to the desired IRE by adjusting the resulting exposure with NDs or Iris control. BTW I did this with my Pocket camera and use the technique with my Micro Cinema camera.
Cheers

I've had the opposite results. I did test it at the time -- material shot in log @1600 gained a little in the shadows at the expense of the highlights.

As I recall, somebody -- John Brawley? BMD staff? -- explained that 1600 allocates more bits to the shadows.

You may be right John, it has been a while since I tried this, and I might have it the wrong way round, as I mostly shoot at ISO 800 and adjust in post as required.
Cheers

I'll check again at some point, but I also leave it at 800 and deal with any further corrections in Resolve. The tests were a few years ago, but as I recall the shadows were so noisy at 1600, it wasn't worth the little added detail gained from it.

That was my take on it also
Cheers

The overly noisy shadows with ISO 1600 may have been partially addressed in a recent firmware upgrade. I recall seeing improvements in the wedding shot this July. I didn't do a formal controlled test though on the Mini 4.6K camera.


Sent from my iPhone using Tapatalk

Now I'm questioning myself about ISO 1600 improving highlights. Here's the thinking though:

When shooting film, a film rated at ISO 400 or 800 or 1600 is used when you need to see into the shadows as it is more sensitive and used when your light is low. A film rated at ISO 50 or 100 is less sensitive and used when you have plenty of light and want to see into the highlights since it's not going to blow the highlights the way a more sensitive film would.

Since we are dealing digital sensors, the various ISO levels alter the gain on the signal from the sensor. So ISO 1600 boosts the signal one stop. If all other camera processing was unchanged, then 1600 would blow highlights sooner than 800. But the surprising result I'm remembering is that it allocated an additional stop above middle grey so that there was a possibility of seeing those highlights with more graduation. It was surprising because a beach scene could look better at 1600 than 800. But there's less graduation in the shadows.

Maybe John Brawley will remind us of the difference in the log curve shooting ISO 1600. Anyone else remember what was discussed quite awhile ago?

Just a reminder that's only applicable when selecting ISO 1600. The log curve is the same for ISO 200, 400, and 800 in terms of the number of stops above and below middle grey.

Sent from my iPhone using Tapatalk

I vaguely remember that discussion as well and thinking at the time, maybe it works on the Mini but it definitely doesn't on the BMPCC.

Easy enough to test, no? Just barely hit the zebras at 800, and see what happens at 1600 if you stop down one stop. Are they still there?

That ought still to be true and that's simple to test. I think what's needed is seeing what happens above middle grey.

Does there appear to be better graduation going from middle grey to the brightest highlights at 1600 compared to 800. I don't have the kind of chart that might indicate the additional stop above middle grey. But someone here probably does.


Sent from my iPhone using Tapatalk

Almost always you get the most dynamic range from any BMD cameras shooting at 800, aside from the global shutter cameras like URSA and the 4K Production camera / Ursa Mini, which are 400.

To the OP, grading is hard and a difficult skill to master. Your highlights issues speak more to your abilities to colour correct. Perhaps look at getting yourself a copy of the colour correction handbook or one of the "learning DaVinci Resolve" books.

Sometimes lot's of people change the ISO in the camera to "trick" the user into favouring highlights or shadows more.

But if you're exposing properly for 800, then you will almost always get the best result. Doing anything else is almost always a "con" for either monitoring more easily of giving comfort in a difficult situation. Sometimes known as ETTR or ETTL. Changing the ISO is a way of cooking the books to achieve that result, something you could also do at ISO 800.

Separate to this, it is true that there's a slightly different allocation on the BMD curve when shooting at ISO1600 in RAW, the theory being that there's more discretion in the shadows for being able to grade later.

In reality it makes almost no difference in my view to the viability of low light footage.

I almost always still shoot ISO 800 when I'm in really low light because it looks better and grades better. And that's in ProRes, not RAW.

Here's some low light shots form the second season of queen of the south. All shot ProRes and good old ISO 800. First as shot....

ExDStill_QOTS_170131_010438.jpg (262.2 KiB) Viewed 12260 times

And then corrected in Dallies. This is a simple grade by the way, no secondaries because it has to go through CDL. We did some further noise reduction in the final grade, but not on this shot. This is also an example of a skilled colourist, and why it takes a long time to actually be good at it.

ExDStill_QOTS_170131_010447.jpg (762.92 KiB) Viewed 12260 times

JB

Thanks, John. I was wrong then to think ISO added a stop above middle grey. It's improving the graduation in the shadows. John Paines had it right. I think it was a good choice for me to shoot the interiors for my recent wedding video in raw 1600 as the shadow detail was much more important than the highlights in that situation. Thanks again for sharing your experiences as they're much more valuable than charts.


Sent from my iPhone using Tapatalk

Thanks, John. I was wrong then to think ISO added a stop above middle grey. It's improving the graduation in the shadows. John Paines had it right. I think it was a good choice for me to shoot the interiors for my recent wedding video in raw 1600 as the shadow detail was much more important than the highlights in that situation. Thanks again for sharing your experiences as they're much more valuable than charts.


Sent from my iPhone using Tapatalk

You're kind of both right though.

Exposing as if 1600 while actually recording at 800 does mean that you're putting an extra stop of highlight headroom because your midpoint is STILL 800, but you're now shifting everything down one stop or ETTL AND THEN, BMD allocate more bit's towards the shadow part of the curve, not to be confused with DR, just the spread of where the bits are across the DR.

JB

Finally understand. Peace and love!


Sent from my iPhone using Tapatalk

[CaptainHook]

The log curve is the same for ISO 200, 400, and 800 in terms of the number of stops above and below middle grey.

This is not correct. The log curve is different for each ISO.

ISO1600 has the most dynamic range above middle grey, and it shifts by a stop each ISO you step down. Resolve will apply these different curves for the 4.6K when selecting the different ISOs from the RAW tab, but the other/older cameras it will just apply an exposure shift (ideally they would have been different log curves also). In camera with ProRes and BMDFilm it's always a different log curve for each ISO that gets applied (and when monitoring BMDFilm with RAW).

For the BMCC/Pocket you're looking at around 6 stops over middle grey at ISO1600, and then lose a stop to the shadows each ISO down. For the 4.6K its around 7 stops at 1600.

Here's an old marketing image from the website for the BMCC/Pocket that shows over/under at ISO800:

Which is what I was trying to point out, though not quite as elegantly put. Thanks.
Cherts

For the BMCC/Pocket you're looking at around 6 stops over middle grey at ISO1600, and then lose a stop to the shadows each ISO down. For the 4.6K its around 7 stops at 1600.

Over the years, discussions of this issue seem to leave everyone more confused, which is maybe why we're still arguing about it 4 years later.

I can't speak to the Ursa Mini Pro, but what, exactly, in layman's terms, are the practical consequences for highlights in the BMPCC when the camera is set to 1600, for log or raw? In my experience that camera setting has never been a good choice, if the idea is to preserve highlights.

[CaptainHook]

in layman's terms, are the practical consequences for highlights in the BMPCC when the camera is set to 1600, for log

Speaking of the log curves, if you were to exposure a grey card for middle grey at ISO1600 you would have about 6 stops over middle grey before clipping the sensor at the expense of more noise in the shadows (since they're raised and you've given the sensor less light than a lower ISO maintaining middle grey). At ISO800 about 5 stops over but less apparent noise, ISO400 4 stops over and even less apparent noise, etc etc.

To maintain middle grey at those different ISOs you either need to give the sensor more light or less.

Speaking of the log curves, if you were to exposure a grey card for middle grey at ISO1600 you would have about 6 stops over middle grey before clipping the sensor at the expense of more noise in the shadows (since they're raised and you've given the sensor less light than a lower ISO maintaining middle grey). At ISO800 about 5 stops over but less apparent noise, ISO400 4 stops over and even less apparent noise, etc etc.

To maintain middle grey at those different ISOs you either need to give the sensor more light or less.

Forgive me, but it's this answer which has perpetuated the confusion for so long. All idiots like me want to know is: under what conditions, if any, does it make sense to set the Pocket camera to 1600, for log? For raw?

I've never seen any advantage to a camera setting of 1600 -- I see a loss, not a gain, in the highlights -- but I'm willing to learn...

Unlike most video cameras, changing the ISO on BM cameras, does not add analog gain. I think this is where the confusion comes in. On BM cameras, as Capt. hook points out, it changes the log curve of highlights and shadows compared to middle grey.

John, for 90-percent of most shooting, keeping the camera at ISO 800 (or 400 in some situations) is desired. Changing it to ISO 400 is like exposing to the right (giving the sensor more light ETTR). So ISO 800 is "normal" exposure/light to the sensor, 400 more light, 800 less light.

If you think of the Zone exposure system, you determine what in the scene is most important, where is the majority of light (or lack of) at. If mostly in shadows, very little I'd sny hoghlight, then shoot at 400 to get more light into the shadows (you can do the sam by staying at 800 and opening the Iris up one stop). For a beach or snow scene where it is all bring highlights, no to very little shadows, then ISO 1600 makes more sense to reduce clipping possibilities, and give more detail the bright part of the scene (snow, water, beach etc). Again you could just leave the ISO at 800 and stop down 1/2 to 1-stop more and get a similar results.

Changing the ISO on the BM camera is like doing an exposure compensation on a film camera (the +/- 1,2,3 dial, which is just changing the ISO setting on the camera to reduce or increase light to the film at a given exposure)
Cheers

Denny, with all due respect, this argument has persisted for too long, with folks routinely changing positions, or expressing certainty one moment and retracting it the next, as this thread indicates.

I understand the basics of how ISOs work on BMD cameras. The issue here is whether the 1600 ISO setting offers particular advantages, which aren't apparent in simple testing, despite persistent claims otherwise.

The answer needs to come from on high, in my view.

I see your point John, as was trying to answer that. Yes, there is still some confusion about how ISO works on BM cameras/sensors. The only time I see changing to ISO 1600 has any advantage (mostly it is a disadvantage, as noise in shadows is increased) is a when shooting a bright beach scene or in the snow, where you do not have deep shadows, only bright highlights, for the reasons I stated above. As Capt Hook mentioned, You can compensate the exposure at ISO 400-1600 by opening up the Iris or reducing ND to get the exposure you want, which would also reduce noise in the shadows, to get the balance you need for the scene. It all depends on what you are trying to shoot and what you want to see in the final shot.

Their are no hard and fast rules, the camera gives you the flexibility in exposure to get the results you are looking for.

Otherwise, I would avoid shooting at ISO 1600, especially in low light (which seems logical to increase ISO to get more exposure in low light), as,this ismdoing the opposite of what you want, more light on the sensor in low light, less light on the sensor in bright light. ISO 400 equalsmore light on the sensor, 1600, less light on the sensor. This is the backwards bit, that I think causes the confusion.

Hope this helps

Thanks, CaptainHook, for the correction regarding the cameras with my 4.6K sensor in particular. My log curve misunderstanding goes back to 2012/2013! You know there were posts in that timeframe that discussed this topic regarding ISO 1600. Good to have your current input on this.


Sent from my iPhone using Tapatalk

Denny -- the impulse to instruct can be very strong (I also suffer from it, if you'll allow me to say so), but there's a point where I think we have to cede the podium to somebody who can answer these questions with authority. Based on the changing views in this thread, I think we've reached that point.

John, apologies but after reading CaptainHook's post here, and CaptainHook speaks for BMD, I'm not sure where there a need for more information regarding the BMCC/BMPCC use of different ISO values. Beyond the theory, what you see in your experience on your camera is what will determine how you shoot.

Hook didn't say how the bits are placed on the log curve, but in terms of dynamic range, the graduation of highlights (the number of stops above middle grey) is greater using ISO 1600. How the 12 bits are placed on the log curve is still part of BMD's secret sauce I suppose but where they are placed alters how we perceive the graduation within the overall frame.

Sent from my iPhone using Tapatalk

John, apologies but after reading CaptainHook's post here, and CaptainHook speaks for BMD, I'm not sure where there a need for more information regarding the BMCC/BMPCC use of different ISO values.

I'm happy you think so, but I don't agree, which is why I asked for a clarification. Any further objections?

[CaptainHook]

I've never seen any advantage to a camera setting of 1600 -- I see a loss, not a gain, in the highlights -- but I'm willing to learn...

Do you have examples to illustrate this?

What you would practically see is a "flatter" curve that would have less contrast at ISO1600 vs say ISO200.

Here's a visual example using the 4.6K. This a DNG (posted to the forums, can't remember by who sorry) to which i've applied the ISO200 log curve and ISO1600 and maintained exposure (lifted from original to demonstrate highlights detail/loss). This also has an s-curve applied as a quick 'photographic rendering'.

You can recreate this yourself in Resolve with a 4.6K DNG by doing the same, change ISO in the RAW tab to 200 and increase exposure until some highlights clip. Create a new grading version, drop exposure slider by 3 stops and change ISO to 1600, compare the two versions.

The downside to this example is the same amount of light hit the sensor for both ISOs where in reality its likely to have been adjusted (but not necessarily). But for showing highlight detail alone it works as an example. Just not in regard to the typical tradeoff of noise. You will also notice with ISO1600 the shadows are lifted and overall the contrast lower, as mentioned.

for some reason i dont get a email notification when i get a reply to a post, just now catching up, i have read all the replies but i think im going to have to keep reading them as im kinda new and a bit confused, but when i first used my BMPCC i was using is200 and then seen some youtube videos saying to use 800 so i tried that out over the weekend and looks good, i will still go back and read all the replies agian in hopes i can learn more. thanks alot

A nice example, thanks Capt. Hook.
Cheers

CaptainHook, much appreciate that example with the interior shot showing the improvement in the exterior through the window using ISO 1600.

I had a very similar result shooting in the early summer evening (still very bright in Canada) and the exterior rendered good detail while I exposed for the interior subjects in front of a large window. I had already decided to shoot ISO 1600 but didn't do any comparison shots as this was a wedding. I remember being pleased with the way the exterior retained detail.

Since I normally shoot ISO 800 under most conditions, I was seeing better than expected results at 1600. Not a scientific test, but the choice of ISO and the observation of the results was well before this thread began.


Sent from my iPhone using Tapatalk

That has been my past experience also Rick, just checked some old shots, from testing the camera a while back. I has to shoot a Parade yesterday, noon sun, but with a slight overcast, shot at ISO 800, pulled down the contrast in film log, came out OK.
Cheers