thanks for posting this! im downloading it right now :)

i'm sure your posting is greatly appreciated by everyone with an interest in this camera. this is the sort of material that leads (hopefully) to making informed decisions....

a couple of questions if i may:

1. what lens were you using?

2. how are you 'holding' it, ie. any rig?

3. how do you find the viewfinder? (yes, i know it's the glass panel at the back ;-))

again, thanks

After reading the BMD forum thread on the "orb" and doing some testing I decided that my camera didn't need re-calibrating and I don't think we're seeing them at 2:10 . I would need to go back to check the size of the actual lights, but in my tests I couldn't create anything like I was shown at John Barry Ltd.
However I do see one light "spilling" over the edge of the escalator balustrade, so perhaps the orb isn't a phantom. I would have set the iris prior to the lights coming into frame, so they would have been clipping. There are other shots showing bulbs where I know I had allowed the camera to set the iris for "no clipping" and those graded well. Doesn't look much of practical problem, really.

The lens used was a Panasonic Lumix G 12-35mm f/2.8 which incorporates OIS. I was just hand holding without other support. At that time I wasn't pleased with the LCD on the back of the camera, noting that I'm used to using an EX1 with loupe on its high quality LCD. In daylight the LCD on the BMPCC tends to be washed out, but is better setting brightness to 100% (default 50%); that significantly reduces battery life!

We're accustomed to a high quality viewfinder (optical much preferred) for maintaining focus and framing with a waveform monitor for exposure, and initially lacking trust in the camera I was unhappy with the LCD on the back of the BMPCC. However actually the camera focus and iris functions work well and one can get away with just using the LCD for framing. Yes, you have to be able to see the square denoting focus area and seeing when it goes to denote focus achieved, and you don't necessarily want to to set exposure to "no pixel clipped".

I have built a rig using a 5.6" TVLogic field monitor (with loupe) I use that when I'm not trying to look like a tourist. But the waveform monitor can be (surprisingly) misleading because the level at which white falls (using film log) is influenced by ISO setting. The camera is native 800ASA and at that white is 100IRE, but at 200ASA clipping occurs at 75IRE (in common with the Cinema Camera); proportionally at other ISO settings of 400 and 1600ASA. So recording film log you can generally be content to use zebra set to 100% and be reasonably confident if no zebra showing then you'll get everything. I like to give a bit more to shadows and will let lights clip (see Bob's query on "orbing", above) whereas pressing the iris button will ensure no pixel is clipped. Using RAW gives more freedom in capturing everything, but that isn't available yet on the BMPCC.
I've largely ignored recording in "video" mode so haven't checked whether ISO setting affects the waveform in that mode.

Incidentally, after some trouble my workflow is to take ungraded ProRes HQ 422 (10b) clips into Vegas Pro 12 for editing (to essentially final cut) and then export this to Resolve through the "FinalCut 7/Resolve" path. This maintains the edited format (video and audio tracks) for grading of the individual clips. Then I render the whole thing as a single clip (ProRes 422 10b) and take that back into Vegas as a new video/audio track for final post (titles, audio etc). Not as tidy as going back and forth between FinalCut and Resolve, but it works without hassle. Resolve 10 now handles avi files (cineform codec) so I'll need to sort out another workflow for those. When RAW recording becomes available there will be need for initial 'one light' grading before cutting, but Resolve will batch render individual clips and my workflow will not need to change (apart from that extra step).
One little trick I was caught by is that in Vegas you can insert a cutaway by placing it on the video track above. Perhaps you can't do that in FinalCut because I had to actually cut them into the video track for a correct transfer into Resolve.

Looks great! Are you saying you DID correct and grade this already? Can you post one of your initial clips strait from camera for us to play with in Vegas?

https://www.dropbox.com/s/p0w5glqsvwditut/Blackmagic%20Pocket%20Cinema%20Camera_1_2013-09-02_1313_C0000.mov

is ungraded and direct sunlight was entering through an atrium. Yet those highlight areas are detailed as are the recesses of the store.

getting it and thank you! I enjoy playing with other camera's footage to see how it is. :)

Any moire or aliasing? It looked very clean to me.
Please post pic of your rig.
Don't suppose it has sdi for a monitor?

there is Moire. but otherwise looks great. you can find the moire in the bench across the street behind the harmonica player in one scene I noticed.

Great camera but moire is a deal breaker for me - when my 5 year old HD consumer camcorder has no moire, I expect all future tech to not have it as well. It's in the Pocket, the Cinema and we have yet to see the BM4K so we'll ahve to wait and see.

I have Moire in my T3i now, so when it is time to upgrade, I want it to be an UPgrade and not just a sidegrade.

[I]"Great camera but moire is a deal breaker for me"[/I]

The forget about using anything digital.
Aliasing is an inevitable problem of all digital recording systems. It can be wrangled by oversampling, in theory at least double, in practise somewhat more.
Even if you can avoid the problem in the original image every time the recording is resampled the same challenge arises.
Only yesterday I made a SD DVD from 1080i footage shot with two EX1 cameras. Absolutely no aliasing / moire in the original watched pixel to pixel on my monitor but a quite noticeable amount on certain costumes when watched from the SD DVD.
Clearly the camera did an excellent job and Vegas could have done a better job however when down sampling there's a very real trade off between resolution, aliasing and render time.

Bob.

isn't Aliasing the jagged resolution lines that should be strait, but Moire is the patches of orange and green that occurs because of bayer whatever in DSLR video sensors? Two different things, right?

here's my very unscientific test showing the moire 2011 canon dslr on the left, 2007 Canon HV20 on the right. Both cameras set to FULL AUTO mode for this test.
Moire is not as obvious in the still shot, but any movement of camera sends them dancing about like Austin Powers at a disco!

Yes the DSLR looks a little better with colors and detail, but NEITHER Canons have true 1080 resolution sadly. And neither are graded at all.

Bob, have you tried using TMPG enc? It works much better than Vegas or DVDA for making SD DVD. Native Vegas encoding gives me moire/alias blizzards.

Regarding aliasing and moire, I've seen it in RED footage, so the real question is how much is too much? My GH2's are okay for me, my 7d wasn't.

[I]"isn't Aliasing the jagged resolution lines that should be strait, but Moire is the patches of orange and green that occurs because of bayer whatever in DSLR video sensors? Two different things, right?"[/I]

No, two different symptoms of the same part of digital sampling theory, the Nyquist limit. Going from that simple theory to explain all it's manifestations in digital image recording is way beyond me.

If you want a simple example you can try yourself grab two piece of flyscreen and slide them around on top of one another and watch what happens. One piece is the sensor and one is what's in front of the camera. Camera designers face more challenges than what you'll see with that example.

Bob.

Moire is that rainbowing you get when shooting certain things. Screen doors create it and fine patterned clothing among other things. Aliasing usually occurs in areas of high contrast -- often around fine lines in the image. I get it with piano keys sometimes, the black and white, or high tension power lines.

Okay these are lame explanations, just google for examples, you know it right away when you see it. And all the footage you've ever shot will mysteriously start showing aliasing and moire, because now you'll be hunting for it.

right. I'm showing the moire in the bottom photos - that wood grain looking stuff is really orange and greenish from the Canon T3i.... why is it a big new problem when it was never seen in a much older HD camera? I've seen aliasing artifacts in grids etc, but now with those crazy colors it stands out because of the BAYER INTERPOLATION or whatever they are doing with the sub pixels to make video... just seems funny an older camera is better than the newer stuff at obvious moire.
Hoping the 4k camera with global shutter won't have it with the multi colors.... no way around the aliasing even though it should be better with 4 times the pixel count

Can't open your photos. However the moire you've noted in my shot is a result of having the fine mesh on the bench in focus. This will happen with any bayer sensor, just as superimposing two meshes unavoidably generates moire patterns. I should have removed it, but I didn't.

I should have fixed that shot, so I have.

Ours came back from BMD a couple of days ago. I [I]think[/I] the "white orb" problem is fixed, never entirely certain if it was a problem.
Either way the firmware was upgraded without me having to stress out doing it and we got an extra battery as a gift, nice.

Bob.

The firmware upgrade I did myself. In my camera I've only seen one example of "orbing" (that clip of mine you pointed out), but I guess I may as well send it in for recalibration. I'll try to do a test that will let me compare before and after. My earlier attempts to intentionally record orbing weren't at all successful.

Bob, if I am looking for what you refer to as 'white orbs' at 2:10, I see bright white light fittings, are these what you refer to?

If so, I am wondering just what you expect to see?

Yes, I didn't notice those until Bob pointed them out; looks pretty much as you expect over exposed lights to look. If you go to the BlackMagic Design forum discussions you'll find a thread on orbs, which basically is about over exposed lights (and the Sun) being rendered larger than the actual source; sort of a glowing orb. BMD offer a recalibration of the sensor/camera to fix the problem. In my own tests it doesn't seem to be a big problem, but it is there.

Any electronic camera is going to be subject to 'orbs' as they seem to be called now. In essence it is just clipping of the video signal, and there are a multitude of places where this can occur, and not all of them have to be in the camera.

Even though Blackmagic say the BMPCC offers 13 stops of dynamic range and 'no pixel will be clipped', it can still occur, the clever gimcrackery is to create a set of modifiable transfer curves in the signal processing to compress the high level signals gently without clipping them, and I am guessing this is what is done when the camera is 'calibrated'.

At the same time, and this is very difficult to do well, you need to raise the signal in the lower luminance regions to bring out the detail, while not raising the noise floor, primarily a function of the sensor. To their credit, BM seem to have got this working very well, although close examination of the clip shows some definite low luminance noise in some scenes, or 'noise in blacks'.

However, looking at that clip, I was not offended by the lights, and frankly, I did not even notice them until alerted to their presence. I have seen far worse examples of peak clipping than those, and the only one that may not be able to be explained as 'what the lamps actually looked like', is the one that appears on the edge of the escalator handrail.

Overall, I thought it was a nice little program, well executed. Well done.