I have only looked at this shot in the stills on the web, but it looks like to me like something is wrong with a camera setting. It should not be neccessary for the operator to have to work around problems where the camera is treating the sky behind the trees differently than in the rest of the picture.
I don't know what it is, but telling the operator to underexpose slightly may be a fix in this particular shot, but its important to figure out what is going wrong because a professional camera should not do that in my opinion. I hope this is not a by product of the CAC lens function or anything else peculiar to the camera because it looks unacceptable to me . The knee should be creating a gentle roll off not a hard clip taht is different in different parts of the picture.

i neIed to learn a bit more about knee settings , but I noticed that the slope was "0" on your settings - what does that mean? Does that effectively get rid of a gentle roll off? I guess I'll look that up right now.

By the way in my mind the histogram is useless for this kind of exposure issue and useless for most exposure questions except the most general sense of whether you are weighted toward under or overexposure and you ought to know that anyway without looking at the histogram.

Has anyone every seen this on another professional camera.

- Lenny Levy

Leonard,
he was not using the cine gamma curves which roll into the highlights. He was using the std3 setting.

I've got plenty of sky images holding up decent latitude.

Here's an EX1 frame grab. It was shot about 11:00 AM.
I believe it was Cine3

Leonard,

While I fully agree this behaviour is unacceptable, I'm slowly coming to terms with it. You see, we're trying here to squeeze maximum quality from a standard (i.e. non-cine) gamma curve, which has been used in conjuction with considerable colour saturation increase (with both Hisat matrix adoption and several primary pairs adjustmenst), but WITHOUT even trying to optimize the curve's KNEE settings...

As I said before - if you want to get saturated and rich colour, either use cinegamma (e.g. CINE1) which takes care of the proper KNEE setting itself, or adjust the KNEE with a standard gamma (like the STD3 used in this case).

Otherwise, don't expect good results with backlit scenery, cause you will either end up with this ugly patches I have shown, or you will severely overblow the backgroud, or you will get nice and rich blue sky but severely underexposed and over-saturated foreground.

Well maybe I've forgotten what to expect in this situation on an ordinary video camera, but i've been shooting professionally for over 20 years the vast majority with cameras in standard gammas - not cine gammas, and that looks wrong to me.
What looks wrong isn't that the camera can't handle the whites in the sky, but that its treating the sky behind the trees differently. You shouldn't have nasty abrupt clipping with areas in the sky abruptly having no color and others having color, that's just unnacceptable.

Maybe I need to look at those pictures more carefully again.

I am confused about one technical thing though.
I looked up point and slope and those settings don't look weird as "0" is a mid point on the slope, but when the auto knee is on does that mean your point and slope numbers are overridden?

Yep. Leaving Knee on auto overrides all your fine-tuning to the point, slope and saturation.

Leonard, you say you've never saw such behaviour with your 20 years with cameras that didn't even have cine gammas; perhaps this is the clue? Now that the cine gammas are available, the standard curves are more "relaxed" and left to the (knowledgeable) users to shape. My experience is much less than 20 years, but I must admit on the V1 I never pushed the standard gamma so hard!

Anyway, this really isn't nice - that's for sure. This grab has not been exaggerated by the 32bit Vegas stretching; it's exactly how it looks at playback time on my monitor (and it's really difficult to rate it as "severely overexposed", either - as Bill is suggesting):

Maybe I'm over-reacting here - and my head did tend to gloss over some of the technical analysis - but i still find myself disturbed taht an ordinary setting on the EX-1 would do this.

I'm not comparingm it to V1's or any other 1/2" "semi-professional" cameras , but rather to run of the mill 2/3" professional Sony & Ikegami cameras taht I've been using for many many years. You should be able to shoot decent looking pictures in standard gamma without being terrified that if you overexpose your sky the picture will start artifacting. To me its completely unacceptable.

Now I don't ever recall seeing this with any semi-pro video camera either like an DVX100 or an HVX200 or even a PD100 for that matter.

It sounds like the only thing you did unusual was add some saturation to the matrix - is that right?

Am I over-reacting here? I mean this is a totally ordinary shooting situation.

Maybe you are right that you need to avoid the standard gammas. - that would be a drag. Did you try anything other than STD 3?

Leonard, you certainly are not over-reacting; my opinion is the same as yours and this is why I started this thread, and even added a suggestion on improving the algorithm behind it in the firmware update wish thread.

And no - apart from adding some saturation and slightly lowering blacks, I didn't do anything else to the most standard factory settings (it's exactly the Bill Raven's PP which he says he arrived at as the most neutral and natural colour setting, using WFM and things :))

Thanks for posting this Poitr and i will do some test myself.
I haven't really tested this camera very carefully just assumed that ordinary circuits like this would work properly.

As I mentioned I have seen sony send out misadjusted auto knee circuits on top end $50,000 cameras in the past so I'm hoping that's all it is.

Well, there's no doubt that it does look odd with your std gamma examples. I'll have to try to recreate this problem. I guess I have not seen it, since I mainly use cine curves.

Piotr...

Your image41 is not the same one that I saw that was very overexposed. Image41 definitely looks very strange...actually, it looks like the white balance settings are off.

Bill,

The image41 comes from the very same few seconds of shooting; it's actually a couple of frames away from those previously posted. I made the reservation about WB not being right in my original post.

The only difference between image41 and the previous four is that this one has been grabbed from an 8bit Vegas project, while the previous ones - from a one with 32bit video setting.

Do you agree now that it's not the question of overexposing, but of how being on the verge of overexposing is being handled by the EX1? I guess it'd be very important for all of us users if all the interested parties came to a conclusion, as it could be passed on to Sony as an important issue for the future firmware release.

Piotr...

when you switched to 32-bit processing, did you check the Vegas WFM for clipping? Ordinarily, when I switch between 8-bit and 32-bit, I need to apply a correction, depending on the codec type. One can't arbitrarily turn on 32 bit without adjusting levels.

For example, .mxf files, I can drop a native .mxf file on the timeline in 8-bit and the levels look right, black is at IRE16 and white is at IRE235. If I switch to 32 bit, suddenly black is at 0(-8% IRE) and white is at 255 (108% IRE) These values are techncally blown out if you're looking at a studio RGB monitor. A Level correction converting computer RGB to studio RGB is needed to correct the images.

Bill,

Yes I am aware of how mfx decodes in 8bit vs 32bit Vegas projects, and belive me this is not the source of the problem. With 32bit, applying the Computer RGB to Studio RGB convertion is barely visible; in 8bit the difference is more apparent (right half of the grab below is converted to Studio RGB; while the superwhites are remapped, the blue "shadow" behind the trees is still there).

But believe me, Vegas is not the factor; in fact I only use it take grabs from timeline. The problem is clearly visible by just playing the file with WMP.

Guess I'm with Steve Thomas. Other than my initial foray into STD curves, which I abandoned very quickly, I don't use them much. I'll look at this some more to see if I can make my cam malfunction like this. In CINE mode, all the manual knee adjustments are disabled. In STD mode, auto knee is turned on. I'm not sure how this functions. Could it be that the camera processor pushed the knee out to an irrational setting to cause what you're seeing?

I must say that I just finished shooting a 240 sec short ...lots of gray skies, snow everywhere, a few patches of blue here and there. Difficult exposure situation for the camera meter with all that snow. Footage turned out stunningly. Lots of detail....no mosquito noise...none. I'm so happy with this camera.

http://www.dvinfo.net/gallery/showimage.php?i=848&c=2
http://www.dvinfo.net/gallery/showimage.php?i=849&c=2
http://www.dvinfo.net/gallery/showimage.php?i=850&c=2

So, after all, even Bill has agreed that the camera is malfunctioning.

I'd appreciate it very much indeed if anyone recreated this with his own unit, and put the result in this thread. Only then will I know for sure if it's only my camera that needs servicing, or a general firmware flaw that can hopefully be ironed out by a firmware update.

Yes, I 'll get around to it.
Bill, that last shot looks great man!
http://www.dvinfo.net/gallery/showimage.php?i=850&c=2

Thanx, Steve.
I used to live in Tempe. I ventured south to CG on many occasions.
Spring training is coming up soon, isn't it?

If the white areas in the sky are a result of being blown out and would have been blue. If the iris was closed a bit more. If the areas around the tree branches twigs etc was blocking light then those areas could possibly have a blue aura? It sounds logical to me.

The colour temperature is set wrong and is too blue, adding to this. I wonder how the image would look taken in camera at the correct temp? I corrected the colour temperature in AE and still there is a blue cast around the trees but also some areas of white. The problem is not uniform. I dont think the camera is at fault at this point. Its going to be interesting to follow this.

As we all know, you either expose for the sky and the foreground is underexposed, or expose for the foreground and blow out the sky. As to why this is happening in the camera, I do not own one so cannot say. But if you want your cake and to eat it too, you can always use a grad filter and you would be golden.

It is severly overexposed. The sky is clipped. One of the somewhat more expensive cameras that I use used to clip to purple. Wasn't really seen as a big issue although it got fixed. The Redone clips to pink if pushed.

The EX1 is more of a digital cinema camera than a video camera. As has been said by many, if you don't know what you're doing with it you can very easily get yourself into an ocean of grief. If you really want to tweak things then you're either going to have to learn a lot or have a monitor hooked up off the HD-SDI port. Perferably one with scopes built in as well. From what I've seen so far the composite SD port does not give you the true picture of what the camera is recording.

And yes, I know a few people who've been using all manner of video cameras for decades. They're having a hard time understanding how cameras such as the EX1 work and how to wrestle the best possible image out of them. This camera does go to 11, use it with caution.

Have a look at the histogram for the blue channel, look pretty badly clipped to me.

Just to jump in to the conversation...

Here's a grab of a test shoot I did yesterday.

I'm a little dark inside (the mantle) and totally blown out...especially in the window picture right.

I could have sworn the stuff I've shot with my XL-2 (my old cam) wouldn't have been so drastic...but this is a totally different camera. There are a lot more examples where there are these drastic brightness conditions, but also a lot of stuff that just blew me away with how stunning it was...and this was shooting 720p.

I love the camera so far, just think it will take some time for us to get to know each other!

I'm chalking this up to operator error since it was my first day out with the camera in a practical situation and wasn't equipped with filters.

After having read Bob's post, Piotr, I know you were using Bill's original TC setting which uses a standard gamma. Did you happen to try turning off the profiles and use the camera's default standard video settings?

That would of been interesting. Bill, I know you took great care in setting up the color channels. Maybe its possible that things are a bit hot with these setting which may drive the colors to clipping a bit early, which may be causing what Piotr has been concentrating on. But on the otherhand, it sounds like you did a decent job setting these up.

Bob, which image did you grab the blue channel histo from? I know he had a couple examples of intentionally over exposed to show what he is seeing.

Jamie,
As you said it was over exposed. Also, I'm not sure what you were focusing on?

To add to Bob's post, Adam Wilt's latest EX1 review he mentioned the following quote regarding EX1 highlights:

"I’ve found that the EX1’s knee does a fine job except when highlights are strongly colored. Saturated highlights show more hue shift and harsh clipping than I’d like. I’m exploring this further out of curiosity, but even if the knees were perfect I would still shoot with cine gammas, because I prefer the progressive compression to the look of a traditional knee."

The one from here http://www.dvinfo.net/conf/showpost....5&postcount=53

I incorrectly named my screen grab Image41, it's actually from image42!

The green channel looks pretty much the same, the red channel isn't clipped.
This is always the problem with cameras that go to 11, if you have anything other than a specular highlight then having any channel go into clipping wierd things will happen if you're not very careful. Even WB setting can affect the color shift as you go into clipping. Trying to fix these problems in post varies from extremely difficult to impossible. Shooting it right is the answer. If that scene was being shot on film there'd be serious lighting in use and the DP running around with a light meter. Sony gave us a cheap camera that gets us one notch closer to film. We need to learn what DPs have been doing for nearly a century.

Probably what the EX1 should have is false color metering but that could be difficult to implement and again you need to know what it's telling you.

What do you mean by the camera goes to 11?

Some funky skies have also been seen in footage comparing the EX1 to the HVX. If I recall, default STD curve used. I can not find the link, so I frame grabbed a problem shot.
http://www.siliconcine.net/temp/Blew_Skies.jpg
White balance also looks too blue, a point Bill raised earlier for consideration. My gut feeling is the STD curve and perhaps the knee is the cause, coupled with exposure.

I posted this before in another thread:
From the XDCam HD brochure (the 350 series), we can see how different the Standard curve compares to the Cine curves, and a lot more of your highlights would have shown a burnt out look with "Standard" curve.
http://www.siliconcine.net/temp/XDCAM_Gamna_Curves.jpg

The Cine Gamma Curves for the EX1, from the brochure and manual.
http://www.siliconcine.net/temp/EX1_Gamma_Curves.jpg

Yes I saw this comparison too and also remember this being brought up.

There's no doubt that the image is way cool. It should of been white balanced. I believe the standard gamma was used. I also agree that the image is over exposed.
This may be just how the standard gamma blows out its highlights when over exposed. I have never seen this since I live on the cine curves.

We'll have to perform some experiments.

BTW, I have this same comparison, but it is 1080P version (Not the 720P version). The EX1 in the same shot does not show this problem. I'm not sure if they just did a better job with exposure, or they switched on a cine curve.

It can record right upto hard digital clipping at 110%. Some NLEs will clip that at 100% unless told otherwise.

Very interesting, Chris. When I looked at the b&w step wedges with the WFM, I compared STD with Cine gamma curves; and, initially chose STD for TC1 because it exhibited a more (vertically) symetrical wfm pattern, as I would expect with the STD curve you show in the F350 series curves. Overexposure indicated a definite cutoff at 100%. I, subsequently, changed to the CINE gamma curve for TC2 because of the more gentle rolloff characteristics at the right end. Furthermore, Adam Wilt writes the following about the CINE2 preset:
"• CINE2 – “broadcast safe cine”. CINE1 rescaled with whites limited to 100%."

If I get a chance, tomorrow, I'll post a screen capture of the WFM displays for the STD vs Cine Gamma curves.

Steven, I also observed the 1080p footage did not have the same funky sky as the 720p. Information from that comparison was perhaps not clearly documented and published.

I hope the following info is not too OT from Piotr's initial post and request for answers to this issue, however, it may have some general relevance and may assist those interested in comparing STD and CINE gamma curves:

It has been documented that even the F900's Standard gamma curve suffers in shots requiring a wide dynamic range. Have a look at this Standard curve image and the Digital Praxis custom gamma curve graphs applied to the camera/scene thereafter. Steve Shaw has been generous to provide a lot of information on the Digital Praxis site.
http://www.digitalpraxis.net/sonycurves.htm

You can see that the Digital Praxis custom gamma curves for the F900 blast the F900's Standard curve, compressing the range, lifting the blacks, preserving the highlights, and very nicely done by DoP, Geoff Boyle. Some curves are good for record, edit and finish with no grading (REC709 preset), others require post grading, however, there is more detail in the image and the right S shaped curves filter applied in post will extract the most out of the image (Bill Raven's "art"). The caveat I would add is that the EX1 codec, even in HQ mode, may have limitations to exploit such extreme PP settings that one may try to dial into the EX1 menus, such as trying to emulating the Cineon Log curve, due to the codec's 8-bit depth and compression likely not being robust enough to correct such extreme grading in post. In that case, recording via HD-SDI to an uncompressed 10-bit 4:2:2 device may be the next step for some. Thanks to Bob Grant for putting some of this into perspective for me, with his experience with the SI-2K camera and edit suite based on the CineForm codec.

Bill, your Wfm screen captures and explanation would be much appreciated, in this quest for knowledge and exploiting the EX1's goodness, to find it's limit!

Steven, thanks for joining the discussion. Let me address a couple of things:

1. Some of my grabs were intentionally overexposed to show conditions where ALL sky goes white. Yes I was checking with Vegas Histogram, Waveform and RGB Parade scopes and confirm that even the image41 shows levels upd to 110% (for Green and Blue). And yet it's the image41' exposure that one would use to "balance" between slightly overexposing the sky and showing details in the foreground - the camera' histogram looked normal to me, and nothing else spelled the disaster, either (all looked OK apart from the WB, but that has already been said).

2. To address Steven's question above: yes, I did try the same scenery with STD3 but Bill's PP off (i.e.default factory setting); the camera behaved in the very same way (auto-exposure, zebra and histogram - wise), BUT the image I got was fine in that I could properly expose the foreground (the sky went all white then), or keep the sky colour (with the foregroud a bit dark, but watchable).

3. As I mentioned many times before, it is the combination of Bill PP's modified Hisat matrix with the standard STD3 gamma that is a no-no. Why? Well, suppose one is already aware of what may happen to partially blow-out sky; he may either expose for the sky (to keep it safe and show the clouds etc), or for the darker foregorund (e.g. to show the face of a person standing there). Unfortunately, with this PP/STD3 combination, one must go way down with the exposure (much lower than the histogram/zebra would suggest) to avoid the patchy sky like I showed - so much so that the foregrund becomes already way too dark, and - due to the PP high colour saturation levels - oversaturated and noisy (see image38).

So, that the sky is clipping doesn't surprize me, as I wanted to avoid underexposed and noisy foreground. If the way it doesn't clip behind the trees is normal for a camera of this (high) class, as Bob is suggesting - is fine with me, as well. However - and I have said it before, as well - using high-saturation PP with STD3 curve is dangerous in that the zebra/histofram and autoexposure of the camera offer no help in getting the right balance (if it possible at all).

Thanks everyone for their participation.

Jamie,

I guess what you're showing with your picture is normal, though can see some traces of the strange abrupt blow-out beginning. I don't know your setting used with the picture you posted, it would be good if you remembered!

For comparison's sake, here go couple more examples of mine (sorry for the WB and focus being off; I took them while walking):

The first two are also through-the-window looks like yours; all factory settings used (i.e. STD3, but no PP at all). I don't think the outside is too much overblown.

The bottom two is the "horror continued" - the first pictures recorded while walking outside through the door (with Auto-Iris on), right after having dialed in Bill's TrueColour PP (SDT3 with Hisat matrix); note that even though I changed the WB quickly (to another bad value, but that proves to be of no importance), those trees against the sky look almost as if there was hoar frost on them!

Go figure...

EDIT: An idea occured to me and I went outside to check immediately - and yes, one can actually see the artefacts behind the trees in the camera's LCD! It's enough to switch peaking off (when I recorded these tests, I had it on with high intensity; since I was on autofocus the trees were all red with peaking, which obscured what was happening with the clipping around them...Now at least I know it's possible to observe and avoid this while shooting.

Please do, Bill! I've been searching for some graphical comparison of the STD curves with Cine gammas (the latter are depicted in the EX1 pdf brochure), but couldn't find anything.

Have some bad news for you guys - even without Bill's TrueColour PP, with the standard gamma curve (STD3, all factory setting) the ugly phenomenon when at the verge of sky clipping is still there :(

And, which is even stranger, it's bluish NOT because of the blue sky in the background; today I got it again with the sky totally overcast (light gray).

Evidently malfunctioning Knee/highlight processing, or at least totally not idiot-proof - even with the out-of-the-box settings!

Again, please try to reproduce it with your units, cause - after being somewhat reassured with the discussion in this thread - I'm again suspecting my camera IS malfunctioning, after all.

PS: Bill, I'm taking back my statements that your TC PP is a "no-no" with STD3 - if anything, it can only make the phenomenon even uglier by making the artefacts more saturated...

Well, if this is an issue with the standard gamma auto knee, Sony will address it.
Having said that, I use cine gamma curves and they are not showing this issue.

I certainly hope so; nevertheless would appreciate somebody confirming it is not JUST my unit...

Anyway, here is the link to the short clip from which my last two grabs were taken (no PP, STD3 gamma, out-of-the-box settings); it shows how the artefact changes with iris opening/closing (original m2t, even with my Polsih comments in it; the format is SP 1080/50i as I was testing my DR60 drive - BTW, works fine with the EX1's i.LINK):

http://rapidshare.com/files/89637385...SP_108050i.m2t

As most of you don't understand my Polish commentary: at the clip end the iris is at F8, and only then the effect vanishes (but the rest of image is severely underexposed). Also, I deliberately put the trees slightly off focus, as it seems to magnify the artefacting.

Again, please somebody confirm it's not just my unit! TIA.

Piotr,

Please post a similar scene where you:

-Use one of the Cine Gammas
-Center your Histogram
-White balance
-Focus

Then shoot it again in the Std 3 (as you have been) and push it as you see fit to prompt the potential defect that you are seeing from the camera.

I think that should be very helpful in letting others help you determine the likelihood of a camera problem.

Best,
Randy

You want to put money on that guess? I'd like to beleive it but I wouldn't be sure. Not unless we make alot of noise.

WFM Readouts for Gamma Presets-Sony EX1
 

Here's some links to the results of the 8 EX1 gamma preset readings I observed on an HDRack Waveform Monitor. In all cases, I used the factory presets with only the GAMMA preset changed. Everything else was at the factory setting.

Here are the caveats...PLEASE be sure you understand them before you make judgements about the results!!

1-Factory settings except for Gamma preset selection. Auto knee ON in STD mode, disabled in Cine mode.
2-Homemade 11-step gray chart--NO guarantees that the super black and super white are at 0 and 100 IRE
3-Scene illuminated by single 600w halogen lamp, minor skews in WFM due to uneven illumination.
4-White balance set to 2700K
5-Exposure normalized for each plot such that the white point always falls on 100 IRE (except for CINE2, see note 6)
6-CINE2 plot intentionally (over-exposed) clipped. Cine2 gamma clips, by Sony design, at 92IRE. This also accounts for the similarities in the middle gray point amongst the STD gammas.
7-I include a jpeg of the step chart for reference. Don't EVEN begin to think you can make some reasonable judgements from this jpeg. There are so many errors due to color profiles, that you'll send yourself down the proverbial garden path. Suffice it to say that the steps are evenly divided and symmetrical.

I see absolutely no evidence of premature clipping, weird effects, with the exception of Cine2, which is specialized, by design. I am putting some faith in whatever algorithm Sony uses to control the auto knee in the std modes. It's conceivable that this circuit can be fooled under some lighting conditions.

Non-normalized Cine Curves http://www.dvinfo.net/gallery/showimage.php?i=851&c=2
Non-normalized STD Curves http://www.dvinfo.net/gallery/showimage.php?i=853&c=2