HD200 to EX3 "wildlife" teething pains (with video)
Yes I know already - I'm retarded, I don't know what I'm doing, I'm an insult to the EX3 community and my mother wears army boots. Now that we've got that out of the way I've got a few observations I'd like to share, and a question I'd like to ask.
I've been shooting with an HD200/100 for two years. I love the camera. It's like a third hand and I shoot great images with it effortlessly. But I sold my 200 seven months ago to buy a RED. Then the Scarlet was "released" a week before my camera was supposed to ship, so I bought an EX3 instead of either of them... I shoot a lot of wildlife so I ordered Mike Tapa's MTF Nikon adapter for the EX3, just like the one I've always used on the 200. This weekend I shot with the EX3, for the first time ever, while it wore a Nikon 300mm 2.8 ED lens on it's nose - just like I always shot wildlife with the JVC. Here are some of the differences I noticed between the two cameras. Observations: 1 - it's so great not having that lapse in time while waiting for the tape to roll on the JVC. New world.. 2 - The JVC has a focus assist option in the LCD which turns the highest contrast areas of the image blue. This is great for use with the Nikon lenses mounted because that option is in the camera. The EX3's focus assist option is part of the stock lens, which is in the bag at your feet while you're shooting super telephoto. Both cameras have adjustable peaking in the lcd, and both lcds are of a similar size. So really, even though the EX3 has a much higher resolution monitor, out in the field they're really not all that different. But the focus assist option on the JVC, which works with the Nikons, makes focusing easier than on the EX3. At least that was my first impression. 3 - When I shoot with big glass mounted to either camera, my right hand holds the tripod handle and my left hand runs the focus/aperture/shutter/record button. This really sucks because while you're following a bird, for instance, you have to look away from the bird if your finger doesn't find the record button right away. And when you're shooting such telephoto it's real easy to lose your subject, so the last thing you want to do is look away. On the JVC, the way it comes stock at least, there was nothing I could do but deal with it. On the EX3 we've got a remote. This remote will live in my left hand. Maybe I'll even velcro it to the lens hood. That's gonna be great.. 4 - Side note for both cameras. Someone recently asked me if I use the eye piece or do I look at the lcd out in the open. On the JVC I always used the LCD out in the open. On the EX3, after only an hour of use, I found myself adopting an entirely new technique which is much better. I now use the eye piece with the boot turned backwards - left eye. This frees up my right eye to sight down the body of the camera and lens, keeping a much wider view of what I'm chasing in the sky, or of what's about to run across the camera's view. This is a huge improvement. No idea why this never occurred to me while shooting for a million hours on the JVC exactly this way - yet it hit me in less than an hour while shooting with the EX3...? 5 - The EX3 is lighter to carry. Maybe you think this isn't a big deal, but wait till you have to carry the thing for nine miles with 26lbs of lenses and a huge tripod to carry. Believe me, it's a big deal. 6 - Here's a gripe. While I am completely new to this camera, and for sure I don't have it set up even close to optimally, for the life of me I can't get it to stop clipping the highlights. Yes I know there's a histogram, and yes I know all about zebras. I'm actually thinking maybe this camera is defective or something. I'm talking NO ZEBRAS, and the histogram right SMACK in the middle of it's range with nothing anywhere near either end. Image looks unbelievable in the monitor, and in the thumbnails menu. Beautiful. But when I get them into FCP, it's like they were taken with a different camera on a different day. What wasn't even close to over exposed suddenly has no information at all. I honestly think something might be wrong. With both me AND with this camera. 7 - C-A. I used to get this under extreme situations with the HD200 also. Once Paolo offered me some time setting my HD200 up with the Nikon lenses, I never had it again. Maybe this is what I need again. I only saw it once, at wide open aperture - again, with the histogram way in the safe zone. I could actually see the CA in the monitor, which is great because I never could see this on the JVC monitor. But it's still there, which kind of sucks for me right now. I do have faith though. If we fixed the JVC, for sure we can help this camera as well. 8 - Shiny blacks, soaking wet, in hard bright sunlight ALWAYS freaked the JVC out. It was almost funny what it did to the image. Purple haze, blue fringe, green teeth - you name it. Especially if you were zoomed way in or had big glass mounted to it. The best of the best said it wasn't actually the JVC in particular, but more it was the 1/3" chip dilemma. Well they must have been right because this EX3 handles the exact same situations with no problems at all. Question: Can someone please help me solve the clipping problem I'm having? Again, please understand I do know how to use a histogram, and I do understand how zebras work. Neither of these indicated the conditions that the camera actually, apparently, recorded. Do you think something might be up with the camera I have? It IS from one of the first batches that arrived here in the US.. Finally, here is the test footage I took today, COMPARED to similar footage that I took with the JVC - with the same 300mm lens mounted. Obviously I need to do a lot more reading than simply the manual or the seven billion pages on EX picture profiles.. |
Excellent summary Eric!
What value is the zebra set to, and what gamma are you using? Can't remember the details off the top of my head, but one of the gamma settings extends the dynamic range quite a bit (at the expense of flat-looking pics straight from camera, ready for grading). If your zebra is set for peak white at 95% or whatever and you get them out of the shot then there should be no clipping, and as you say the camera might be faulty. Why not mount the stop-start remote on the pan handle? Steve |
Setting the exposure properly with these cameras is quite diffrerent from any other camera I've ever used! It depends on your Gamma settings mostly. Are you using Cine Gammas? This camera (and my EX1) tend to compress mid-tones around 65% IRE, you need to set one of your zebras for that and make sure certain things like faces, skin tones etc don't go higher than that. I use STD4 mostly due to the lack of a knee setting in the Cine Gammas. Some like those looks better though so you should definatley experiment.
Here's a thread that discusses it at length: http://www.dvinfo.net/conf/sony-xdca...ure-cine1.html There are many others on this forum! Just search under "Exposure EX1" Adam Wilt had an excellent description of how the Gammas all affect the image on ProVideoCoalition.com see here: http://provideocoalition.com/index.p..._camcorder/P3/ Good luck! |
Surely the only effective way to know exactly what you're getting is to set the zebra for white clip level, then you know that if you see zebras there's nothing there.
If you set it for mid tones you're just guessing what the whites are doing. Steve |
You can assign the focus assist function to any of the assignable buttons. It's in the "others" menu.
I suggest you give either cinegamma 2 or cinegamma 4 a try, that should help with highlights. If you don't like the cine gammas then you will need to set the knee to suit your shooting style and help with clipping. I like to set my zebra to 95%. Then I expose so that the brightest part of the image is just showing the zebras. I don't use the histogram unless I'm doing chroma key. |
Quote:
This is mentioned in MANY, MANY posts here, most explain it with different terminology, you lose detail in whites (such as Erics white feathered bird clip he references) and midtones become pasty long before you reach 100IRE. |
Is this not something you deal with in grading, when you change gammas etc., to strecth the dynamic range you're bound to end up with skewed results, but you put them right in grading, so boosting contrast, selectively bringing highlights down and deepening blacks, colour correcting mids etc.? Sorry post production is not my area!
Steve |
Quote:
If you protect the highlights all the time at any cost then you will have a lot of trouble in many situations where you really need the shadow areas exposed correctly and you don't have much choice of background. If I can expose for a skin tone I can then make a better judgement of how much to back off to bring highlights back in or shadows. Using mid tone (70%) zebras also helps to eliminate any exposure issues due to a badly set up or old viewfinder, because 70% is a good exposure for certain surfaces. In other words you have a direct reference. Exposing for the highlights can lead to trouble with a bad viewfinder. |
Quote:
|
Quote:
In quick use I find it's also easier to look for highlights with 95% zebras and eliminate them than searching for "mid-tones" which are more of a variable quantity. Steve |
Quote:
Quote:
With 8-bit recording on lower end cameras, and that includes the 3xx series, I would not want to have to bring up or bring down any extreme end of the picture. If you are doing higher end work then you will be using higher end equipment and have a crew and the ability to control the light somewhat. At the lower end you have to be more considerate about where you shoot, when you shoot, and what you can do to manage things. Can't protect the highlights and shadows simultaneously? Simple, change the shot angle, shoot at a time of day when the light is less problematic. Now with wildlife a magical moment only happens once, and yes ideally you would want all extremes protected, but the fact is that this isn't always possible. You can strive for it, but when it comes down to it that is why all serious wildlife videographers have the best camera they can possibly get their hands on, and more often than not that still means film. That's one reason why the RSPB for example are reluctant to move to HD across the board for their film unit. The cameras simply aren't up to snuff in that capability nor the framerate ability. Not even the Varicam. Using 70% zebras isn't an iron clad rule. You have to judge it and make a call. It is a guideline. Quite often on my full size cam I will have both zebras enabled, but on the EX3 it is too confusing to do this. But generally I can see what is plain white by looking at the screen. The 70% zebras offer me not only a guide, but also as I mentioned an insurance policy against knackered or badly setup viewfinders. Using peaking can also sometimes give a false impression of the overall brightness of a scene if it is set high. All said and done I'd much prefer a realtime waveform display like some of the Panasonics are starting to do. |
OK so thanks very much for your responses. Very informative. Enlightening. Thanks for the focusing button tip Alister, and everyone for the theories.
It was really late last night when I posted. I was drooling. I can add to this now. I actually took a photo of what I saw in the monitor, just in case the exposure that the EX3 recorded turned out to be very different than what the histogram, zebras, and monitor were indicating. Turns out what got recorded actually was very different. By the way I have been up and down and back again on the monitor adjustments. Nothing's helping. Here's the scene Here's what the monitor showed - note, NO zebras, and a nice, safe spread on the histogram Here's what the camera recorded - note, whites clipped, CA around highlights, sensor having a heart attack What I would like to know is simple - if I can't rely on the zebras, if I can't rely on the histogram, if I can't judge by the monitor (someone please tell me that image in the monitor is showing signs of clipping) - what do I do now? Simon, you state here: "generally I can see what is plain white by looking at the screen." Well this is exactly what I 'used' to be able to do so well with the HD200, but cannot yet with the EX3. There was no histogram on the JVC, but there were zebras. And when they first showed up, this was a warning - the way it should be. But so far this EX3 is treating me like a pregnancy test treats a teenager - no warning, just the answer. On a related note, today I remembered watching a guy shoot at Mavericks last month with an EX3. Guy was real proud of his cam. Said he'd shot the previous day from a photo boat out there by the break (on the day I saw him we were both on a cliff about a mile away from the break). By the next week his footage was up on the Mavericks website. It looked horrible. Yes that's right, totally blown out. And while I feel his pain, white water in broad daylight is a tough subject to nail, this guy wasn't even close. It was really poor footage of a really great subject. So now I'm thinking back today, to that guy, and I'm wondering if maybe it's not just me here struggling all on my own? Remember, my buddy Brian totally fried his subject (model about six feet away) on an overcast day while shooting fully auto ....and she has dark skin and was wearing a cream colored sweater. |
I have an EX1 and shoot outside all the time and I'm not running into any clipping problems. In fact I'm surprised it doesn't clip, especially on very white sand under bright sunlight. I am using a PP from the Vortex EX1 Training video, I think the zebras are set at 95% and it's a Cine4 gamma. The default settings that came with my EX1 clipped bad... I couldn't even shoot anyone wearing a white shirt outside, even on a cloudy day, it just blew it out. Once I changed settings I've never had clipping problems since.
|
Thanks Buck, I haven't tried 4.
Tomorrow... just for kicks, when your EX1 clipped, did you get proper indications on the monitor/histogram/zebras? That your whites were too hot? Or did the footage look exceptional in the monitor/histogram/no zebras and only when you captured it did it look horrible? |
Quote:
|
[QUOTE=just for kicks, when your EX1 clipped, did you get proper indications on the monitor/histogram/zebras? That your whites were too hot? Or did the footage look exceptional in the monitor/histogram/no zebras and only when you captured it did it look horrible?[/QUOTE]
Eric, I'm not sure because it was early on but I don't recall seeing zebras or off the charts histogram levels. I'm sure I was using "auto" exposure, I was still learning the camera. Even with "auto" exposure now I don't get clipping although I try and shoot mostly manual. I don't even know what the default settings were, I changed my Profile while watching the Vortex DVD, pausing the video to make the changes as I watched. Even skin tones used to blow out easily in sunlight, like the forehead and cheeks on my boy, but now they're completely fine. I tried one or two other Profiles too that are in the Profile thread but other than that I haven't experimented on my own with Profiles. I really like the one I have so I don't see any need to keep tweaking, at least not yet anyway. I know I have a lot more to learn. |
What gain are you shooting at? And is any auto gain or auto iris being applied. Try with -3 db and repoyt back.
|
I'm not surprised to see so much CA and a generally soft picture as you have an extender on the back of the lens and extenders are notorious for producing soft pictures with lots of CA.
CA is primarily a function of the lens, while you may have been able to reduce the effect to some degree with careful camera setup on the HD200 it will still be there, just hidden. Stills lenses are designed to have nothing but air between the rear of the lens and the sensor or film. A video camera has a big glass prism in the way. The larger the sensors the bigger the prism. This added glass will bend different wavelengths of light by different amounts adding to the CA. Video lenses are built to allow for the extra glass but stills lenses are not. Cinegamma 2 won't allow you to go above 100IRE so if your zebras are set to 100% you may well never see them. Cine 4 goes up to 109IRE (I think). Going by the pictures you have posted I would say that the recorded image looks pretty much like what's shown in the VF. The clipped areas are very very small and only just going in to either clipping or a hard knee, the histogram simply isn't detailed enough to show up such small specular highlights. As the birds head is facing the opposite way in the final grab it's difficult to judge the overall exposure of the white feathers, the VF was indicating 44% which looks to be about right to me. Overall the scene would be a challenging one for any camera. Brilliant white bird against a dark shadow filled background with bright sunlight overhead. |
|
|
Quote:
|
I concur with Alister. Also depending on which gamma setting you had you may be able to bring back some of that highlight detail with care. Especially if you could actually see the detail in question in the LCD.
Having said that though the EX's histogram is not very good. As Alister mentioned it simply isn't detailed enough to show small areas of highlight blowout. As well as this I find that what you see in the EX3 LCD is extremely sensitive to the way that you have set up to the bars. On my 510 I find that although I regularly check the viewfinder setup I rarely have to adjust it. With the EX3 I find that I have to constantly adjust it to bars throughout the day. This is why I mentioned that my zebra method was a guide too, because knowing the characteristics of the EX I find that it is always best to bring things down a tad from what on most cameras I know I could leave it set. I wouldn't be at all surprised if Alan Roberts settings, when they are made public, recommend setting the black gamma to a certain level even if the cine gammas are used. This way you could bring down the exposure and still have shadow detail (better than bringing up shadows in post by far because it is done pre-compression in camera). I wish there was a straightforward answer to this, but hopefully Alans settings will shed some light on things when they appear. |
Quote:
|
My two cents:
Try Cinegamma 1. I used to use CG4 to avoid highlight clipping, I thought CG1 was dull. But after extensive testing I found that CG1 gives the best result in high contrast situations. CG4 compresses the bright region of the picture too much which sometimes results in banding effects in the highlights. Also - and this is another topic - I noticed from your footage and also from the still of your LCD that you are using auto/high speed shutter. This is something that I have found many still photographers that move to video do, probably out of habit. Maybe you have a good reason for doing this that I don't understand but my advice is to always shoot at 60 fps (or 30 if you need the light) to avoid the home movie look that fast shutters give. To me that is an amateur-give away on par with blown out whites. Just a thought... Otherwise - great footage! Good luch with your work. |
Quote:
Quote:
Kill the auto white balance too because that'll give you colour shifts like nobodies business. |
Quote:
Thanks. Paul. |
The gamma curve shouldn't produce any banding. That's more likely due to the monitor.
|
It should also be remembered that any zebras operate over a certain window. Say you set the Zebra to 70%. It will start to show at 70% but once you get to around 85% it will stop showing. So if the zebra's were set to 70 it is still possible that they would not have been showing as the clipped areas are above the 70% window.
|
Quote:
I tend to switch to Cine1 when I need to be careful about highlights and it works ok for me so far. |
Something strange is going on there?
The 44% center brightess level indication on the swan certainly appears brighter than in the actual image. Like you mentioned, the histogram does not seem to reflect what you actually shot. Although, as mentioned, it may not work good enough to show small incremental hot areas. Also agree with everyone here 100%, do not use auto! This may have caused this issue. I have not seen this with my EX1. odd. |
For whatever it's worth, CA was a lot worse on the JVC HD100 (110,200,250) series.
The tele end at wide open you always had to be careful, especially under high contrast. |
Quote:
My theory on this is that by using this sloping gamma curve you are compressing a lot of luminance information into a very small range of the 8-bit sample. The shifts in luminance are therefore being quantified a bit too much. There is simply not enough zeros and ones to describe the shifts in luminance. I am sorry if I did not manage to make my point clearly. English is not my native tounge. |
Quote:
|
Thanks guys for offering up such great help.
Simon, the zebras were ON - like I already said, and not one line showed up anywhere in the image at any point that I was recording. They were set at 90 though, not 70. Zero gain. Again, not one zebra line showed, histogram was safe, monitor looked 'underexposed' even, and it clipped, which is surprising to me to say the least, and which is why this post. Thank you for the shutter speed tips. They are my common practice, and actually how I shoot/shot every scene (30fps - 1/60shutter) minus the turkey vulture (where I added text in the video explaining that I had used auto shutter specifically there). Call me an amateur, but I find it useful when panning across the sky chasing a bird as he flies past the sun and back again. The Egret scene immediately followed the vulture, my bad for shooting the still of the monitor before I changed the shutter mode back, rather than after. If you will notice, in the still shot of the monitor the camera is not recording. I re-set shutter just after I took the photo/before I recorded the image. The Egret footage in the video is taken at 1/60th shutter. Your well trained eye can probably tell by the motion blur in the frame grab, right? Trust me, or not, my clipping issue - so far as I can tell - is not related to any shutter speed/techniques that I use. Alister, good on you for noticing the extender. I purchased it about a month ago and it was sitting in my pack fixed to the end of my still cam with the 400. Temptation was irresistible. I HAD to try it. However I only put it on at the end of the shoot, for the pair of Mallard ducks (in the video), and for this Egret. In the video, where the much smaller (farther) Egret walks across the dark background, I actually only had the 300 mounted. No extender. But to be honest it was pretty wide open for that shot. Two ND filters, 1/60th shutter, no gain and almost wide open. I bet if I had stopped the lens down and used 1 ND filter instead I could have reduced CA. I appreciate the critiquing of my shooting techniques everyone. Thank you. I hope to always keep learning. What my real point is here, is that on this particular camera there actually exists a HUGE difference in what my monitor is showing, compared to what is being recorded. Let me say that again - there is a huge difference in what my monitor is showing, compared to what the camera is actually recording.. So thank you Alister for pointing out that the histogram is not very accurate. That solves one mystery. Simon, could my NOT setting the monitor up by the color bars be causing/adding to this discrepancy between the viewed image and the recorded one? |
Quote:
Quote:
Steve |
[QUOTE=Eric Gulbransen;986089]Your well trained eye can probably tell by the motion blur in the frame grab, right? Trust me, or not, my clipping issue - so far as I can tell - is not related to any shutter speed/techniques that I use.
QUOTE] Do I detect a hint of sarcasm here? ;-) I am sorry if I offended you. It was definately not intended! Actually, I did not even think as far as Allister did concerning the auto shutter possibly being the source of your exposure problem. I just wanted to make you aware of the fact that while using the shutter to control exposure is an excellent choice when shooting stills it gives some very obvious and nasty effects when shooting video. I noticed this on some of your shots on the video clip you linked to (with my well trained eye) and it was confirmed by the still of your view finder. Earlier this year I worked with some footage from Kongo shot by a stills fotographer who was new to video. He was great at composing his images but unfortunately he did not realise the difference between using the aperture and ND to controle exposure as opposed to shutter speed. The images looked like holiday movies. I hope this can be of some sort of help to you (or somebody else). :-) |
Quote:
They did seem particularly interested in the 700 though, but their set of circumstances is particularly demanding so it would need to prove itself. |
RSPB have a highspeed Arri that'll do 150fps so until there's a video camera that'll do that they'll keep using it (I'm doing a Phantom HD shoot for them but that's even more pricey than shooting film!). But, just like everyone else they know the benefits of shooting tape - cost only being one of them.
Last time I spoke to Mark they were very keen to get something ASAP,just had demos of 2700 and pdw700 and as I said they were waiting on my 720 results from the PDW700 to help the decision - just taking longer to get the 720 update than I' hoped! Steve |
Interesting. Sounds like they really want to bite the bullet now.
Any idea why the update is taking so long? A lot of guys in the US have already got their updates. |
Apparently there's a queue and you wait until your number's called, but I can't believe there are that many that need doing, it's been available for about 2 months now.
Steve |
All times are GMT -6. The time now is 03:48 PM. |
DV Info Net -- Real Names, Real People, Real Info!
1998-2024 The Digital Video Information Network