IRND Filters

I recently shot a movie of the week in Vancouver and was lucky enough to get some early samples of the new Panavision IR ND filters.  I tested them here on both the Alexa and the Ursa Mini 4.6K.

You can see the Ursa Mini 4.6K is much more susceptible to IR pollution in this torture test under tungsten lighting.

I also compared the Schneider IR filters and you can see in the above link they really don’t measure up, whereas the Panavision IR’s look really great and were very consistent across their range.  I went onto use the Panavision IR ND filters for the movie and I thought they were great and really consistent.

Up until using them on this film, I’ve pretty much only used the wonderful Mitomo TRUE ND’s, but as some will know these are obscenely expensive and hard to come by.

Well, I recently discovered, much to my pleasant surprise, that you can now BUY the Panavision IRND filters, because they’re made by Lee, and Panavision own Lee.  And they’ve just started to sell them to us the general public !

Lee call these “ProGlass Cine IRND filters”.

Here’s the official page from LEE.

And here’s the official Panavision page.

And just to make sure they’re the same filters I checked by camera testing them!

So here’s a quick exterior IR filter test.  I shot using the Alexa Mini, shooting ProRes 4444 Log C 1920.

I shot the Panavision IR filters, my new LEE IR filters (which should be the same) and some Tiffen IRND’s as a comparison, along with the Alexa Mini’s own internal ND.

Below is a little edit of the 2.1 ND in all situations, so you can compare.  I was surprised how well the Tiffen did actually, as I’ve seen them be a lot worse.  The 2.1 ND is a real test because at that density you generally see lot’s of variation in colour and consistency.  As an extra check I’ve also shot using no filters at all.

Here’s the Digital Pigeon preview file (faster load)

Or watch here on Vimeo.

All I’ve done here is apply the Resolve / Arri default REC 709 LUT and I just used a curve to open the blacks up a little bit more.  I then applied this grade to all the clips.

I used shutter angle / shutter speed to regulate light levels and used false colour on the grey chip to maintain exposure.  It was all shot over about a 40 min period with very little cloud, so it should be consistent.  All the exposures are noted on the slate.

As expected, there are little variations in exposure and most surprising to me was that I found the Arri Alexa Mini internal ND’s to be a little green and differing in density on this default grade.

But if you want to check for yourself, go ahead and download the original camera files here courtesy of my friends at Digital Pigeon.

I’d be really interested to hear if anyone sees anything different, feel free to leave a comment if you do.

There are some tiny differences in the way the edges are finished on the LEE filters compared to the PV filters but the glass is the same, as are the pouches. I really like the little notches on the PV filters to let you know by touch how many stops they are.  I would pay extra to have my filters engraved with my name the way the Panavision ones are, I wonder if that’s hard to do?

 

 

Posted in Equipment, General | Tagged , , , , , , | 3 Comments

Intrigue – How far will you go ?

 

I’ve been flat out in pre production on a new Telemovie I’m shooting in Vancouver, but I wanted to share this little short film project from earlier in the year.

Following on from the launch video I did for the Olympus EM-5 Mark II, I thought it was time to put some of the newly updated features to the test on an actual short narrative style shoot.

I’ve been doing a lot of testing of the new FLAT video profile for the Olympus E-M5 Mark II and I was really keen to find out how far I could take the image if it was recorded to a more robust codec using an external recorder.

After the initial launch of the E-M5 Mark II Olympus have been taking feedback from video users and have included some of that feedback into firmware updates to further improve the video capabilities. Most significantly they’ve created a FLAT profile for shooting video and the thinking is to try and squeeze a little more DR into the image by starting with a lower contrast and flatter picture profile. This then allows you to put the colour and contrast back in during the final grade.

About the Shoot

I wanted to see how the camera and new profile would produce in a very minimal environment.  For me, this is when using a mirrorless camera really excells.  I can be very discrete and shoot without any requirement for permits.   With just two crew and the two actors, I was keen to try the camera in a few scenarios where I knew it might shine.

So I had three days.  Luckily I had some friends in LA on a recent trip to California that I was able to rope in for this shoot.

Now who hasn’t thought about driving the open highway in a classic car like a nice ’65 Mustang ?

So I had in mind to shoot some in-car work and also to visit some of the more iconic locations of California.

Starting at a wonderful standing “movie version” of an American gas station, we set the scene between our couple who meet for the first time.

The Story

After some initial frisson over the bowsers he leaves before they can talk, but she eventually finds a little surprise he’s left her.  A ring !

Later she realises that he’s left her a message inside the ring, it’s an Instagram handle.  Looking it up she realises he’s left her a picture clue for where he might be next, so she takes off to track him down.  After arriving at the beautiful Joshua Tree, she again finds another ring.  Checking in on Instagram, he’s left yet another clue, and she’s off again, this time to the snow capped mountains of inspiration point, about 90 mins outside of LA.

Once in the beautiful mountains she finds a third ring and it points her towards Santa Monica Pier where she finally meets up with the mysterious stranger.  They have a dance by the sunset before he finally offers her a final ring.

The Rules

I had no tripod.  This is all shot handheld.  No gimbal, no monopod.  There are a couple of shots where I had the camera mounted using a suction mount when it was on the car, but that’s IT !  So everything you see here is done using the amazing 5 axis image stabiliser.

The Outcome

With a grand total of four people, three days and 600 miles of driving this is what we got done !

The Gear

Everything was recorded on an external recorder over HDMI using the “clean feed function” of the camera.  I used a 5″ Blackmagic Video Assist recorder to record the HDMI signal to 1920 ProRes HQ on it’s internal SD card.

Intrigue was edited and graded using Blackmagic’s DaVinci Resolve (which is available as a free download here)

If you’re interested in what the visual differences are with the flat profile, you can get a sense of where the grade started to where it ended up by looking here. The first image is the final grade and the second is the ungraded shot.

 

I used a combination of lenses, mostly the outstanding Olympus 12mm F2, and the Olympus 9-18mm F4-5.6.  I love these lenses because they are so lightweight !  For ND I used the SLR Magic Vari ND and in addition to these lenses I had some PL mount Zeiss MK 3 Superspeeds from my friends at Blacklist Digital and an MFT–> PL lens adaptor from Hot Rod Camera.

Thanks to

Special thanks to the awesome Jane Harber,  Armen Taylor on acting duties and especially to Jessica Clarke-Nash, who helped shoot and cut this for me. My good friend Hook did the grade.

I also really want to thank Kristy Galea and Olympus Imaging Australia for her awesome support and making this shoot possible, and Quett Lai for his masterful technical support.

Download your own files

For direct viewing in vimeo go here.

If you want to download some juicy large ProRes files to try editing and grading your own clips then go here.

Very special thanks to my friends at Digital Pidgeon for making these files available for download, it’s not easy doing such fast hosting of large video files.

Posted in Equipment | 35 Comments

In Memory of Morgan

I just learned of the sad news of the passing of a colleague Morgan Evans.

You have probably seen him on your screens and not realised it.  I was lucky enough to know this amazing stunt performer and director and got to work on a few shows with him.  I’ll never forget his boundless enthusiasm and energy and love of life.

I still remember seeing one of his VCA films as a young filmmaker and being so astounded at the audacity of his stunts, humour, production value and the political commentary he managed to combine in one simple film.

Although we hadn’t crossed paths lately, I miss you already and the world is poorer for not having you around.

Morgan Evans (pointing behind camera), Tony D’Aquino and I working on a TVC in Vietnam

 

 

 

 

 

Posted in Uncategorized | 3 Comments

Ursa Mini 4.6 And Alexa side by side

So here’s an interesting one.

I had a chance while shooting some tests recently of the new Olympus firmware upgrade to the Olympus E-M5 Mark II to shoot some tests in a studio under controlled conditions.

I set up a 3200K balanced lighting setup and then also able to shoot the Ursa Mini 4.6K in RAW and then shot the same again with Arri RAW on an Alexa XT.

This was shot using Cooke S4’s and both cameras were exposed for ISO 800.

I designed this scenario to show some contrast and I’ve also added a brightness ramp on the back wall.  You can work out how many stops over and under from the Fc readings here.

 

Tungsten Set

 

 

 

I think the UM4.6K stacks up very well against an Alexa shooting RAW, but I’d love to hear if you think it does.

The only thing I’ve noticed now that I’ve had some time with these files is that there appears to be a small amount of IR contamination in the blacks around Maddy’s shirt and the fabric on the table on the UM4.6K.  This can be easily corrected but it’s unexpected.

I used a mitomo True ND6 here and these filters have been very reliable for me on other cameras.  I have to admit a lot of the other shooting I’ve been doing for Blackmagic I’ve been using the firecrest ND’s and haven’t seen this problem before.  It may well be that the TRUE NDs are letting in some IR that the UM4.6K is seeing, but I’d have to test further for that and haven’t had a chance to.

In any case please do enjoy these camera original files hosted by my friends at Digital Pigeon.  I’d love any feedback too if you’re finding the files fast to download or not.

Here are the Blackmagic shots.

http://bit.ly/1U7P4OB

And then the Arri Alexa RAW

 

A002C001ALX_1.1.1

Arri Alexa RAW open gate Cooke S4 32mm

http://bit.ly/1Z2VONe

A002C007ALX_1.5.1

 

Arri Alexa RAW open gate Cooke S4 75mm

http://bit.ly/1o24uWv

 

 

 

 

 

 

 

 

 

Posted in Equipment, Uncategorized | 18 Comments

Ursa Mini 4.6K Carnival

With the recent shipping news on the much anticipated Ursa Mini 4.6K camera, I wanted to give people a chance to download and grade some more files.  I have a few other more “test” style shots coming, but for now I’m keeping it fun.

I spent a couple of hours wandering around the Santa Monica Pier at sunset and this is a selection of shots I managed to get.

I shot Ursa Mini 4.6K PL and no ND.  These are RAW 4:1.  I had the Angenieux 16-42 and the 80-200 TLS which is a rehoused Nikon lens.  Both are T2.8 lenses.

Because they’re compressed RAW, only the CURRENT version of Resolve will open these files. It’s a free download.

Also, I want to acknowledge my good friends Digital Pigeon for hosting these very large files.   They are hosting these for me for nix so please pay them a visit in return and consider their specialised high volume super fast file sharing options.

They put all the others to shame and I know from personal experience these guys are the best.

http://www.digitalpigeon.com/

And so the files.

I was going to cut a little piece together, but I figure it’s just best if you get to try them for yourself.  I’ve got some little preview shots but best you download to see what they’re really capable of.

I’ve tried to include the darker evening and mixed lighting shots.  Normally a challenge for any camera.

If you grade and share these files online, please remember to acknowledge the source. I’d love to see you post your grades back here in the comments too.

 

1741 C0016_1.2.1

http://bit.ly/1ZhoO3S

1741 C0018_1.3.1

http://bit.ly/1R8mgTO

 

1747 C0027_1.3.1

http://bit.ly/1R8mlaf

 

1758 C001_1.5.1

http://bit.ly/1R8mQAZ

 

1811 C017_1.6.1

http://bit.ly/1R8n0Ix

1812 C018_1.7.1

http://bit.ly/1R8nezw

1818 C0024_1.8.1

http://bit.ly/1R8nlLg

1819 C026_1.9.1

http://bit.ly/1R8ntKK

1823 C031_1.11.1

http://bit.ly/1R8nChe

1826 C35_1.1.1

http://bit.ly/1R8nOgG

1748 C0028_1.4.1

http://bit.ly/1R8nZbE

 

 

Posted in Equipment | Tagged , , , , , , , , , , | 17 Comments

Olympus E-M5 Mark II FLAT profile tested

***** UPDATED WITH NEW INFORMATION. SEE BELOW

 

Further to some recent camera testing, here are the results for the Olympus E-M5 Mark II using the new flat profile.  Late last year Olympus released a significant firmware update to the E-M5 Mark II that among other improvements, added an extra video shooting profile.

Using the flat profile allows you to try and squeeze a little more dynamic range into the eight bit file being recorded by the camera.  You should use the flat profile when you know you’re going to be grading later and in doing so you’re going to preserve as much detail for the grade as you can.

Another way we can preserve information is to use a better acquisition codec.  The E-M5 Mark II allows us to record the clean feed from the HDMI (unlike the E-M1) so that means I can use an external recorder.

For these tests I dual recorded, both internally to the none-to-shabby All-I codec but at the same time I also recorded 1920 ProRes HQ files with a Blackmagic Video Assist.  This is a small low cost monitor / recorder that’s very well suited for this job.

When it came time to compare these files I also learned something.  Now, with the FLAT profile loaded, I expected that the internally recorded ALL-I files would look almost the same as the externally recorded ProRes HQ file, but I’d find the ProRes files would be more be even more robust to work with because they have less compression and are 10 bit files instead of 8 bit.

What I didn’t expect was there was a slight difference in colour hues.  This was confusing because bit depth and compression shouldn’t affect how the colours are reproduced.  So I went back to Olympus and it turns out that they have used two different colour gamut standards.  So internally the camera uses the older REC 601 colour gamut standard, but on the HDMI output they use the current HD spec of REC 709.

Now I don’t think that many people will notice or even worry about this difference and the external recording is at least using the more up to date gamut standard.

So to reiterate here again, this is the same lighting setup I used in the recent E-M1 test I posted.

Here’s the lighting plot.

Tungsten Set

And here’s what you’re looking at on screen.

I also set up a daylight colour temperature version.

Daylight Set

 

And here’s what you’re looking for here…

And then the same candlelight version.

So unlike the E-M1, because we’re now shooting big boy formats like ProRes, these files are much bigger than the internal camera files are normally, so be prepared !

You can find these files, hosted and shared by my good friends at Digital Pidgeon, who offer a very fast and secure file sharing service.

Here’s the edited ProRes and All-I internal comparison original file.

1gb E-M5 Mark II Colour Hue intercut

 

As a little coda to this, I’ve since learned the following….

I’ve confirmed the HDMI output is 4:2:2 @ 8 bit.

I’ve also learned that the frame rate is DIFFERENT on output compared to what’s recorded internally.

This means, you choose 24 FPS in the menu, internally it’s recording 23.976 (the us / japanese version of 24 fs) BUT on the output over HDMI it’s straight 24 FPS.

Same for 30 fps.  It’s recording 29.970 internally but outputting 30 FPS over HDMI.

 

 

 

 

 

 

 

Posted in Equipment, Uncategorized | Tagged , , , | 33 Comments

Olympus EM1 for Video

 

This is the very first of a series of camera tests I shot recently in Sydney.  These are studio tests and I’m also currently shooting some more real life scenarios which I’ll be aiming to share with you in the coming weeks.

I’m planning to show you some examples of the Olympus EM1, the video oriented EM5 Mark II as well as the new Olympus PEN F.  I also used an external video recorder on the EM5 and had a look at the new FLAT log-ish profile Olympus introduced recently with a firmware update after some user feedback.  In theory this should give us a bit more room to move when we’re grading.

I also tested the Blackmagic Ursa Mini 4.6K and the Alexa XT, but we’ll get to those later.

P2090024.jpg

So first off we have the Olympus EM1.  I have to say right off, this is my absolute favourite Olympus camera to shoot stills with.  I just like the way it handles and the sensor is really nice.  The pictures sit very nicely when graded and I don’t have to work them very hard at all.  Skin tones especially are very natural looking.  Their recent firmware update also made the camera significantly better.  I can now use silent shooting mode which is AWESOME on a film set ! Not only that but, for the first time you can now shoot video at the more cinematic friendly frame rate of 24 frames per second !  *technically 23.976 but we’ll discuss that later

While it is the flagship Olympus camera, shooting video is not really it’s forte.  It’s mostly limited by the data rate limited IPB codec they use on this camera.  Basically the bucket the camera has for storing video data is very small.  It’s more like a small bowl than a bucket.

I like the challenge though of making it look as good as we can get it.  I actually think these pictures hold up pretty well.  I first of all went into the custom picture profile menu and created my own profile so I could try and maximise the amount of dynamic range the camera can capture.  I do this by turning down the contrast as low as possible to -2.  I also turn the sharpness down to -2 so as to reduce the harshness and electronic looking image post sharpening can introduce.  We can always sharpen the image later in the edit if we need to, but if it’s baked in from the start it can’t be removed.  Lastly I also turn the saturation down a little to -1.

Once I have this custom profile loaded I can then try to make sure we get as much DR as posisble into the tiny little bucked the codec allows us to capture.   The data rate is very low at about 26 mbps.  The more modern ALL-I codec on the EM-5 MarkII by comparison is 77 mbps.

As I mentioned, I think these picture do hold up well even at such a low data rate.  You do notice the image though is quite soft.  This is a combination of the data rate and the way the sensor is read in the camera when in video mode.

P2090007.jpg

To show these cameras off, I came up with three scenarios.  One is a scene lit by tungsten lighting, one is daylight and one is candlelight.

Here are the lighting plots.  I used the same lighting setups for all the cameras I’m testing.

Tungsten Set

 

So here’s the main tungsten lighting setup. I very deliberately say tungsten because in motion imaging, the sensors tend to be most optimised for daylight, usually around 5000K.  But we often use tungsten lighting to light with ! Tungsten is also a beautiful light source to work with.  Unlike LEDs and Fluros it’s got a beautiful and even spectral emission plot.  In simple terms, it emits all the colors we want to see in very even amounts.  A lot of other light sources like LEDs and Fluros are “peaky” so there can be big gaps in their spectrum of light that they emit.  So it made sense to me to start with and focus on a tungsten lighting setup.

You’ll notice in the lighting plots I will put the CRI, which is a score out of 100, 100 being a perfect and flat light in terms of colour.  I also note the exact WB point too.

I then tried to create a range of exposure zones and then had our wonderful model Madison walk through them.

Check out the lighting plot above.  You can see the key is a 12k Dinette Wedge or Booklight. It’s 12 x 1 K par64 cans, bouncing into a 12 x 12 Ultrabounce that then goes through a 12x 12 HiLite diffusion frame.

I created some negative fill on the camera left side, the 12×12 solid black.  Then on the back wall you’ll notice I used a 2K Fres to do what i call an exposure ramp.  By skidding the light along the wall, I can create a hot spot that falls off over distance.  This is really useful to us to compare when we try to recover over and under exposure. Incidentally there’s also a bare bulb prac light and of course a colour chart and cube.

So firstly I metered the end position of where Madison gets to.  Here when she’s looking at camera the meter says 198 foot candles.  At the base sensitivity of the EM1, where in theory we get the most dynamic range, that means an exposure of T5.6 at 1/50th of a second exposure.  I like measuring in footcandles because it’s an absolute measurement and independent of ISO.  It’s especially useful in scenarios like this because we have different cameras.  It’s also very easy to work out exposure ratios too.  400 Foot candles is double 200 foot candles and that means +1 stop…!

Now take a look at this hilarious image.

 

Now, where Madison starts she’s about 1 and 1/3 of a stop hotter to where she ends up.  At it’s very brightest, the back wall measures 2 and 2/3 of a stop hotter and the dark side of her face is as much as 2 stops under. She has a range to walk though of exposure that will remain constant.  This shows you also what I have set up and look for when comparing images.

For me I’m most interested in the extremes.  What is it like at what I call near clipping and then clipping.  How much detail is there holding in those highlights ?  Then I look at the same in shadows.  What detail is in near blacks and then true black.  I’ve tried to show those for you.

So.  Once I’ve built the above setup, I then do bracketed exposures.  I expose at the “On” exposure, indicated by the “0” on the slate.  Then I do a +1 stop version a +2 stop version and a +3 stop version. So you can then imagine the values indicated above, modified by the over exposure numbers.  At +3 stops, the back wall in effect is 5 and 2/3 stops over exposed.  Then I do an underexposed version at -1 and -2 stops.

I then did a daylight setup, though this time much simpler.

 

Daylight Set

And finally a candle light version.  In this one the camera was set to ISO 800.

 

Lenses were the Olympus 25mm and the 75mm primes.

I took the rushes into Resolve to edit and grade. Resolve is great because they have one of the most powerful colour correcting tools built in as well as great editing functionality.  It also reads almost any codec from any camera and works with it natively in 32 bit float precision.  What’s even better, is that it’s free with mostly the exact same functionality as Resolve Studio.  All you miss out on is the improved speed of using multiple GPUs (if you happen to have them) and a few things like noise reduction.

So because I’m going to give you the camera original files to work with yourself, I did a very simple edit and grade on these rushes, trying get them looking as nice as I could, showing them in their best light.

Here’s my best go at grading this camera for these setups. Click on the EM1 graded .mp4 file to watch or download.  In the link you’ll also find a folder containing the EM1 bracketed rushes as well as the actual 1920 444 ProRes if you’re super keen to pixel peep or try grading something better.  I’d love to hear how you find the over and under exposed images.

I’d love to add my files are being hosted by a terrific new file sharing service I’ve just discovered called Digital Pigeon.  Unlike a lot of other cloud services, these guys are specifically setup to handle very large files, very quickly AND they offer some cool services like an annoted video reviewing system.  You should try them out if you need to get big files moved around for client review.  They’re much much faster than the other clour based services.

Now are you ready for the end result ?

Here’s the link !!!

http://bit.ly/1PxEDPZ

So my takeaway for the EM1 for shooting video is that the codec really limits what you can do with the camera because the data rate is so low and this means the image is a bit soft and get’s mushy very quickly in low light and when shooting at higher ISOs.  The actual pictures can look really lovely when the exposure is centered and meaty.  Normally the solution for a low data rate codec would be to use an external recorder like the Blackmagic Video Assist, but unfortununitly, the EM1 doens’t offer live view output over the HDMI port.  Let’s hope they address this in the future.  But, if like me, you already have an EM1, this is a great confirmation that the recent 4.1 firmware update at least get’s you 24 fps and in the right circumstances you can make it work. I would also like to see Olympus make sure that we have the option of true 24 FPS as well as 23.976 which is what it currently shoots when shooting at 24 FPS.

You should note that I’m an un-paid Olympus brand ambassador and Olympus supplied me with all the camera equipment used in this test.  This post is un-moderated and hasn’t been edited by Olympus in any way. I have been using Olympus equipment since I bought an E1 in 2004.

 

Posted in Equipment, General, Photography | Tagged , , , | 8 Comments