PDA

View Full Version : Canon 7D, meet the Stouffer 4110C dynamic range test



Barry_Green
10-20-2009, 01:38 PM
After the Billups/ToNi shooting session, I decided I needed a better way to test for definitive dynamic range.

Enter the Stouffer 41-step transmission wedge, which is a calibrated strip designed to transmit light in calibrated 1/3-stop increments. Each step on the wedge is 1/3-stop denser than the step that came before it. So, every three steps = 1 stop. To calculate a camera's total dynamic range, you shoot the chart in front of a bright light source, and examine the steps to see where you can clearly distinguish between them, and wherever you can't, you take note of that step as the point where the dynamic range of the chart has now exceeded the camera's ability to resolve it. Whatever the number is, divide by three and that's the # of stops of dynamic range you're looking at (so, if it discerns 22 but not 23, you'd stop at 22, divide by 3, and end up with 7 1/3 stops of total dynamic range).

Much more repeatable than my prior test, and certainly less subjective in evaluation.

So -- I ran the 7D in every gamma mode, using contrast -2 to extend the dynamic range as far as possible. I also did one test with Highlight Tone Priority On vs. Off, and one test where I ran the same gamma curve through many of the ISO settings to see how dynamic range might change under higher ISOs.

In all these tests exposure was set so that the first step on the chart, step 1, the most translucent, was rendering on the waveform monitor as just barely darker than full clipped white.

So -- here are the results:
http://dvxuser.com/barry/Composite-Gammas-160.jpg


http://dvxuser.com/barry/Neutral-160-HTP-Off-On.jpg


http://dvxuser.com/barry/Composite-ISO-test.jpg


Draw your own conclusions. The way I see it, it looks like about 7 stops, maybe 7 1/3, in just about all the gamma curves. If you really dig into the shadows you might pull another 2/3 of a stop out, but I don't see how the most optimistic pixel peeper is going to be able to say the 7D is delivering more than 7 2/3 or maybe 8 stops.

Highlight Tone Priority certainly does some shadow-digging for you; without fishing around, the 160-iso strip stops at around step 21, and the HTP-enabled 160-iso strip stops at around 24. I don't know if it's actually added discernible dynamic range or if it's just pre-amplified the noise in the lower strips so that you can see them more discretely; you could probably extract the same range from the non-HTP strip in post. If you want maximum range without digging around in post though, maybe HTP is worth it. It definitely flattens the perceived contrast and boosts the brightness of the low-to-midtones.

As for the various ISOs, up to 1600 it doesn't really seem that there's any effect on dynamic range at all. They all get down to 21. But you definitely start losing at 3200 ISO, even though you can see strip 21, the noise has smeared the difference between 18 to 21 to a point where it's probably not "usable" dynamic range. At 6400 ISO the noise is contaminating all the way up to step 15, mushing to a blob of noise between 15 and 21 to the point where it's probably not usable.

So, there you go. I put it at about 7 1/3 stops usable range, and depending on how low the ISO is, you could probably post-boost it to 8.

polr
10-20-2009, 01:47 PM
Thanks Barry.
It was my understanding that using the "non-native" ISOs (320,640,etc.) produced less noise but also less dynamic range than the native ISOs (200,400,etc.), and my testing seems to confirm that.
Personally, if I use HTP, I will use a non-native ISO to counter some of the added noise, and if HTP is off then I'll stick with native ISOs.

Terry_Lasater
10-20-2009, 01:48 PM
Thanks for your tests, Barry. I appreciate all your insight and efforts.

Basically, I bought the 7D for both it's stills and video capabilities. For what it CAN do, I believe it to be a bargain.

J. Odoms
10-20-2009, 01:51 PM
hey Barry can you perform this same test on the 5D MKII

Kholi
10-20-2009, 01:57 PM
Thanks Barry.
It was my understanding that using the "non-native" ISOs (320,640,etc.) produced less noise but also less dynamic range than the native ISOs (200,400,etc.), and my testing seems to confirm that.
Personally, if I use HTP, I will use a non-native ISO to counter some of the added noise, and if HTP is off then I'll stick with native ISOs.


Wait, explain this if you haven't vanished already. LoL

Non-Native ISOs producing greater/extended dynamic range sounds off to me. Do you have any tests up? And how do non-native ISOs (adding more noise and odd cross hatched patterns to the image) combat more noise added by Highlight Tone Priority? I'm not saying you're wrong, I just haven't heard this one yet and looking for some insight.

Also, I am STILL hyper-curious to see if higher ISOs, native or non, retain more highlight detail than not. I guess it's time to go out to the Skate Park and run this theory.

In the end, I want to see how much can be pulled from shadows with the cleanest settings and from the highs in the same scenario.

commanderspike
10-20-2009, 02:07 PM
Now this is more like it.

Have you done this test on the 5D as well? Particularly interesting to see the ISO levels and how noise affects darker areas of the image. Would love to see how 7D and 5D compare at ISO 3200 and ISO 6400 on that chart. Dynamic range test between the two would valuable - but isn't dynamic range about colour as well, not just gradients from black to white? What if one colour channel has more dynamic range than another?

Also what screen and what software are you using to compare and compile the samples? Could gamma levels affect one's perception of the chart? 5D / 7D footage looks really dark on my Macbook until I transcode it. Were these samples taken directly from the H.264 footage?

polr
10-20-2009, 02:17 PM
Vanished?

Here's a post demonstrating what I'm talking about: http://forums.canonphotogroup.com/showthread.php?t=2541

It was a test done for the 5D, but seems to apply to the 7D as well based on my observations.

In short, native ISOs are 100, 200, 400, etc. Non-native ISOs 1/3rd of a stop directly above a native ISO (125, 250, etc) have higher noise, but non-native ISOs 1/3rd of a stop directly BELOW a native ISO (160, 320, etc) REDUCE noise at the cost of some dynamic range. This seems to hold true until about ISO 1600.

So, my reasoning is that if HTP adds what seems to be 1 ISO stop's worth of noise AND dynamic range, using a non-native ISO 1/3rd of a stop below a native ISO brings some of that noise back down (seemingly 1/3rd of a stop's worth) to an acceptable level, while keeping most of the benefit of the HTP. Basically, it's kind of an in-between step between HTP completely on or off.

Barry_Green
10-20-2009, 02:20 PM
hey Barry can you perform this same test on the 5D MKII
No, 'cause I don't have one... but DPReview did that test here:
http://www.dpreview.com/reviews/CanonEOS5DMarkII/page25.asp

Kholi
10-20-2009, 02:20 PM
Vanished?

Here's a post demonstrating what I'm talking about: http://forums.canonphotogroup.com/showthread.php?t=2541

It was a test done for the 5D, but seems to apply to the 7D as well based on my observations.

In short, native ISOs are 100, 200, 400, etc. Non-native ISOs 1/3rd of a stop directly above a native ISO (125, 250, etc) have higher noise, but non-native ISOs 1/3rd of a stop directly BELOW a native ISO (160, 320, etc) REDUCE noise at the cost of some dynamic range.

So, my reasoning is that if HTP adds what seems to be 1 ISO stop's worth of noise AND dynamic range, using a non-native ISO 1/3rd of a stop below a native ISO brings some of that noise back down (seemingly 1/3rd of a stop's worth) to an acceptable level, while keeping most of the benefit of the HTP. Basically, it's kind of an in-between step between HTP completely on or off.

LoL! Sometims when people ask questions, the original poster vamooses and we never get them answered. That's what I meant.

You're saying the native ISo's are 100, 200 and 400. That's much different than what everyone else is reporting: 160, 320, 640 which are much cleaner than anything else.

I'm confused now myself, because I certainly see 160 as noticably cleaner than 100.

Barry_Green
10-20-2009, 02:23 PM
but isn't dynamic range about colour as well, not just gradients from black to white? What if one colour channel has more dynamic range than another?

Also what screen and what software are you using to compare and compile the samples? Could gamma levels affect one's perception of the chart? 5D / 7D footage looks really dark on my Macbook until I transcode it. Were these samples taken directly from the H.264 footage? Dynamic range is about brightness, not color; that's why a wedge chart like the Stouffer is gray ND.

I don't use a screen to compare, I use a waveform monitor. And yes, if you're trying to evaluate something using a computer monitor whose gamma is off, you're pretty much at the mercy of whatever that display will show you.

And yes, I only ever work off the native h.264 footage.

Kholi
10-20-2009, 02:23 PM
Thread: http://www.dvxuser.com/V6/showthread.php?p=1764463#post1764463

Post:


ISO/Noise tests. 1080/24p, lens cap on. In FCP, native format, maxed the mids, applied same filter to every clip. Large video version is uploading.. Handheld, skewtastic footage to come..

http://farm4.static.flickr.com/3497/3974367271_a93c631447.jpghttp://farm3.static.flickr.com/2448/3974357139_5c68ba45a8.jpghttp://farm3.static.flickr.com/2461/3975111342_4bd784bb59.jpghttp://farm3.static.flickr.com/2675/3975101976_fd3948b533.jpghttp://farm3.static.flickr.com/2493/3975093594_ae9321103f.jpghttp://farm3.static.flickr.com/2570/3975085808_6b54f1c1e5.jpghttp://farm3.static.flickr.com/2605/3974313663_44e0832972.jpghttp://farm3.static.flickr.com/2635/3974305221_3c3a7394c1.jpghttp://farm3.static.flickr.com/2551/3974296195_c888c42194.jpghttp://farm3.static.flickr.com/2494/3974288047_15ac540f50.jpghttp://farm3.static.flickr.com/2659/3975043554_31ddf18350.jpghttp://farm3.static.flickr.com/2468/3974269769_17178325d7.jpghttp://farm3.static.flickr.com/2644/3975022920_7bd26f70a6.jpghttp://farm3.static.flickr.com/2564/3975011440_6eb4553704.jpghttp://farm4.static.flickr.com/3509/3974235497_380dfa68ec.jpghttp://farm3.static.flickr.com/2508/3974990346_65ea7b6468.jpghttp://farm4.static.flickr.com/3196/3974982104_318cf8d15d.jpghttp://farm4.static.flickr.com/3469/3974972786_10c3e549b1.jpghttp://farm4.static.flickr.com/3418/3974962262_5cb430458d.jpg

This only makes it more obvious for me... so what am I missing in conjunction with HTP??

USLatin
10-20-2009, 02:23 PM
Yea, it seems the native ISO's have a bit more dynamic range since the others are achieved by messing with them. Daniel explained all of that to me over at Reduser and he really knows his technical stuff. I've been shooting nothing but natives unless there is no way to get correct exposure at night but use one of the others.

It seems part of the reason why some have less noise is because they get the blacks crushed or cut completely... I can't remember the explanation, I have been meaning to make myself a cheat sheet though. Let me find that post.

polr
10-20-2009, 02:28 PM
I have read arguments on both sides, but feel the article I linked to makes the most convincing argument that the native ISOs are indeed 100, 200, etc. and that 160,320, etc have less noise because they are being "darkened" digitally. Of course, if anyone has a better interpretation of what might be going on, I'd love to hear it!

Barry_Green
10-20-2009, 02:33 PM
I don't know that anyone outside of Canon Japan knows what the "true" native ISO of the chip is. The idea that there could be multiple "native" ISOs is just bizarre to me. I've never heard of that before.

Traditionally you have a chip with a base native sensitivity, tuned to deliver the maximum dynamic range and sensitivity and least noise, and that's where it's set at. Then you gain up (or gain down). Gaining up will add noise and take away dynamic range off the top. Gaining down will clean up noise and take away dynamic range off the bottom. At least, that's how it works in any video camera I've ever tested.

So ... multiple "native" ISOs? Weird. And to have basically the same dynamic range across multiple stops? That's different too. Obviously the stills guys do it differently.

So whether they're "native" ISOs or not, there's no arguing that the 160/320/640 ISO grouping is cleaner in noise than the 100/200/400 groupings.

polr
10-20-2009, 02:38 PM
There is indeed no denying that 160/320/640 are cleaner, but if that's at the cost of dynamic range, which it seems to me to be, then that is worth looking into.

And regarding the whole native ISO question, perhaps there are steps which are done at a lower level (100, 200, 400) and others which are done at a higher, or later, level in the camera to add the extra steps (125,160,320, etc). Hardware vs. software? Just a thought..

USLatin
10-20-2009, 02:58 PM
Here, start with the last post on this page and read his posts in the next page or two: http://www.reduser.net/forum/showthread.php?t=34434&page=25

It seems there are only a few analogs and the rest are processed digitally.

There will be "multipe native ISOs" since noise is becoming so low through a range of sensitivities that you can get a rich and clean image on a range. However the tern "native" might not work for that. Native shoud be the absolute best, one. But a new term should be brought on for a range of ISOs that give you an image that is hard/"impossible" to argue with.

polr
10-20-2009, 03:11 PM
Thanks for that. The first post on this page seems to explain it best:
http://www.reduser.net/forum/showthread.php?t=34434&page=27



100, 200, 400, 800, 1600: normal amount of highlight headroom.
There is little difference from ISO 100 to 200, because most viewers can't tell the difference in photon shot noise, and the read noise drops by almost a full stop in ISO 200.
125, 250, 500, 1000: considered harmful. 1/3 stop less highlight headroom and 1/3 stop more shadow noise.
160, 320, 640, 1250: fine, as long as you are aware of the decreased highlight headroom. (Clips 1/3 highlights to get 1/3 more shadows.)
HTP should *always* be enabled when the ISO is higher than 1600.
There is a careful balance between increasing shadow detail through ISO (which clips highlights) or through HTP, the picture profile (e.g. contrast, tone curve), or Auto Lighting Optimizer. Generally, the higher you go in ISO, the more beneficial it is to use non-ISO methods to increase shadow detail. For example, ISO 3200+HTP (actually ISO 1600) is better than the real ISO 3200. ISO 800+HTP may be better than the real ISO 800, but it depends on how much highlight headroom you need.

Kholi
10-20-2009, 03:20 PM
Thanks for that. The first post on this page seems to explain it best:
http://www.reduser.net/forum/showthread.php?t=34434&page=27

Nice. Great info. I assume it applies to the 7D and video. Gonna try this out.

USLatin
10-20-2009, 03:42 PM
No problem Kholi. Thank Daniel over at Reduser, he keeps pouring detailed knowledge all over the forum.

I keep trying to remember the numbers. The easiest way for me is like this.

100 = best
125 = hell no
160 = 1/3 stop less highlight

Then 1600+ always use HTP.

I should write that on my hand.

Isaac_Brody
10-20-2009, 03:54 PM
Quick question, does this test take into account that quicktime color space issue that clips blacks much sooner? Or does that not make any difference?

Barry_Green
10-20-2009, 03:58 PM
I am unaware of what you're talking about; as a Windows user I rarely ever use quicktime, so if there's a clipping issue I don't know about it.

I was using the waveform monitor on the BT-LH80 when doing this test on the Canon, and blacks were rolling off slow and even, and my count of the chips on the live waveform was even a little bit lower than my count on the computer screen, so ... I don't think it was affected by any issue such as you describe?

Isaac_Brody
10-20-2009, 04:01 PM
I was referring to this:

http://cineform.blogspot.com/2009/01/correction-canon-5d-is-fine-here-is.html

Barry_Green
10-20-2009, 04:26 PM
I just processed one of the files through Cineform, brought it into Premiere and into Vegas, and compared the original against the cineform file.

On the AVCHD, with the GH1, there was no difference.

On the Canon quicktime movie vs. the cineform AVI, in Vegas, there was a difference, and it went the opposite way you'd expect. The original Quicktime file has a range of 0-109 on the waveform, but the cineform AVI was 0-100.

In Premiere, the Canon quicktime movie showed a range of about 5 to 100, and the Cineform file was about 4 to 102. So there was a difference, but backwards and to a much lesser degree than in Vegas.

Bizarre.

Sttratos
10-20-2009, 04:56 PM
Quick question, does this test take into account that quicktime color space issue that clips blacks much sooner? Or does that not make any difference?


Wasn't it a problem with quick time rather than with the recorded footage?
Is this what you are talking about?
http://prolost.com/imported-data/2009/1/22/quicktime-76-fixes-5d-movies.html

Isaac_Brody
10-20-2009, 05:38 PM
Yeah that's what I was referring too, looks like that's not an issue. Thanks Barry.

Sttratos
10-20-2009, 05:49 PM
Yeah that's what I was referring too, looks like that's not an issue. Thanks Barry.


You are now referring to what Barry said or to the link I posted as "what you were talking about"?

Isaac_Brody
10-20-2009, 05:53 PM
For some reason I can't access prolost on this computer. Not sure if it's what Barry was talking about.

Barry did you select Limit YUV to preserve dynamic range when using Cineform?

Kholi
10-20-2009, 07:25 PM
Well one thing's for sure, if 100 or 200 ISO or anything outside of the 160, 320, 640 and 1250 range are netting more highlight detail then I'm gonna have to pass.

100 ISO and 200 ISO look terrible to me, especially on a large of decent quality LCD Television (my prime qualifier, as it's most likely where people would see material if distributed).

160, 320, and 640 are clearly more pleasant, and I'm willing to bet just under exposing for highlights and bringing up shadows a pinch is worth more than shooting at off ISOs.

100 and 200 also introduce a VERY apparent pattern in the image. It's like horizontals and verticals, a cross hatch that my eyes pick up almost immediately. And, I haven't tested this yet but I'll make the statement so if anyone wants to get to testing before I do, they can:

I think Moire discoloration or Moire in general is exacerbated by non "native" ISO's. This from a quick test run on accident shooting a very finely patterned sweater. I haven't tested it, will try to tonight with more patterns but I'd love for someone else to give it a go.

Neutral, everything dialed down. See if it makes a difference or was I seeing things.

USLatin
10-20-2009, 07:47 PM
You forgot to post the link.


Well one thing's for sure, if 100 or 200 ISO or anything outside of the 160, 320, 640 and 1250 range are netting more highlight detail then I'm gonna have to pass.

I don't think I am understanding what you are saying. I know you aren't saying that you don't care to get more highlight detail. But I can't follow what you are aiming at.


I'm willing to bet just under exposing for highlights and bringing up shadows a pinch is worth more than shooting at off ISOs.

Yes, but if those shadows are clipped then there would be less to bring up. Now if you don't want them there anyway because of noise I understand.

Please let me know what you end up deciding is best when you do. I would really appreciate knowing what you are doing.

Kholi
10-20-2009, 07:52 PM
You forgot to post the link.



I don't think I am understanding what you are saying. I know you aren't saying that you don't care to get more highlight detail. But I can't follow what you are aiming at.



What link?

No that's exactly what I'm saying. I'm not sacrificing overall image fidelity for something I can fix with diffusion, a polarizer or adjusting exposure with ND's etc and bring up later in post. If the image is clean and properly exposed in the first place there's no need to worry about Highlight Detail (because you've exposed properly) and if you're worried about highlight detail there are many more things you can do aside from throwing your entire image under the bus.

Any ISo's around 160, 320 and 640 look horrendous to my eyes. It's not worth it.



Yes, but if those shadows are clipped then there would be less to bring up. Now if you don't want them there anyway because of noise I understand.A properly exposed and lit composition won't suffer from muddy or messed shadows. Again, native ISO's are clean enough (I've seen this with my own eyes) to be brought up a pinch without much degradation to overall image.


Please let me know what you end up deciding is best when you do. I would really appreciate knowing what you are doing.I'm gonna keep doing what I'm doing: Shooting 160 or 320 when I can, using NDs and Polarizers like I should be doing and controlling light to the best of my ability with the equipment available so that I don't have to worry about it later.

However, trying to correct a baked in pattern that presents itself-- along with chroma anomalies-- isn't a good idea to me. The milage of others will probably vary.

Barry_Green
10-20-2009, 08:04 PM
Barry did you select Limit YUV to preserve dynamic range when using Cineform?
There's no option like that in the windows version of NeoScene... ?

sblfilms
10-20-2009, 08:51 PM
Thanks for the time you've spent Barry. Definitely has me excited for the Mysterium-X as we have found the Canon is not impressive in DR or resolution yet can produce a very pleasing image.

Luckily by the time the S35 Scarlet actually ships (I'm guessing second half of 2010) I should have replenished my camera account after dropping a ton of cash on lenses, rails, HD LCDs, mics, etc (and all of that can be used with the Scarlet) in addition to the 7D.

I can definitely say I am loving the 7D and so are my clients.

Sttratos
10-20-2009, 09:19 PM
Any ISo's around 160, 320 and 640 look horrendous to my eyes. It's not worth it.

A properly exposed and lit composition won't suffer from muddy or messed shadows. Again, native ISO's are clean enough

So which ones are the so called native ISOs? 160, 320, 640 or 200, 400, 800? It has been said both ways in this thread.




100 and 200 also introduce a VERY apparent pattern in the image. It's like horizontals and verticals, a cross hatch that my eyes pick up almost immediately.

Wasn't this pattern in low ISOs problem supposed to be caused by defective cameras? I remember somebody posting about it and then posting that Canon had swap his camera for a new one because they said it was a defect?

ydgmdlu
10-20-2009, 09:47 PM
So which ones are the so called native ISOs? 160, 320, 640 or 200, 400, 800? It has been said both ways in this thread.
Forget about what is or is not "native." The terminology isn't important here. What's important to know is simply that 160, 320, 640, and 1250 have the lowest noise.

Kholi
10-20-2009, 09:52 PM
So which ones are the so called native ISOs? 160, 320, 640 or 200, 400, 800? It has been said both ways in this thread.



Wasn't this pattern in low ISOs problem supposed to be caused by defective cameras? I remember somebody posting about it and then posting that Canon had swap his camera for a new one because they said it was a defect?

It's no defect. I have two cameras and an mkii and all three exhibit the same thing in "off" iso values.

What the post above me says. I'll be sticking to those isos and doing some experiment with lighting optimizer and htp in post.

polr
10-20-2009, 10:07 PM
It's no defect. I have two cameras and an mkii and all three exhibit the same thing in "off" iso values.

What the post above me says. I'll be sticking to those isos and doing some experiment with lighting optimizer and htp in post.
Even if those ISOs are possibly simply clipping the shadows along with the noise?
Barry, you've been infinitely generous with your tests, and I almost feel bad asking for anything, but any chance you could test say, 320 vs 400? If the dynamic range is identical, then there will be no question that 320 (and 640,etc) is the better choice, but for now I am seeing too much conflicting information..

Kholi
10-20-2009, 10:37 PM
Even if those ISOs are possibly simply clipping the shadows along with the noise?
Barry, you've been infinitely generous with your tests, and I almost feel bad asking for anything, but any chance you could test say, 320 vs 400? If the dynamic range is identical, then there will be no question that 320 (and 640,etc) is the better choice, but for now I am seeing too much conflicting information..

Well, I'm not saying anyone else has to do it. Once you're educated on the subject you do have a choice: my choice will be to do the same thing I did with the HVX and HPX which is light according to the limitations to get the cleanest image possible. As far as I've been shooting, 320 ISO with nothing else enabled, indoors in proper light is excellent. Midtones etc. haven't looked unpleasant to my eyes. So it's just a choice.

Any other ISO outside of the primary three/four are just ugly. Especially with the added chroma noise in mids and darks.

Andrew Brinkhaus
10-20-2009, 11:01 PM
So based off Daniel Browning's post:

"160, 320, 640, 1250: fine, as long as you are aware of the decreased highlight headroom. (Clips 1/3 highlights to get 1/3 more shadows.)"

Is this with HTP on, or off? A big difference, right? What happens with these ISO's and HTP both on and off? Seems like the safest ISO's to use for the largest DR, or should one stick with 200, 400, 800, 1600ISO, etc.

Uwe Lansing
10-20-2009, 11:04 PM
This chart is interesting too:
http://www.dvinfo.net/forum/attachments/canon-eos-7d-hd/14582d1255984794-canon-7d-iso-wb-noise-tests-noisechart_quartersize.jpg


Found in this thread: http://www.dvinfo.net/forum/canon-eos-7d-hd/466073-canon-7d-iso-wb-noise-tests.html

Tim Joy
10-22-2009, 08:51 AM
Thanks for all your testing Barry.

What I would like to know is the amount of headroom before over exposure. For example- if I took a meter reading of a flat white surface and it gives me a stop of 5.6, and that's what I set the lens to, how much can I open up before over exposure?

Thanks!

Barry_Green
10-22-2009, 09:15 AM
Didn't test specifically for that. It would probably vary depending on the scene file you used, and whether you had the auto lighting optimizer on, and whether you had HTP on.

Tim Joy
10-22-2009, 09:40 AM
What's your opinion on those settings for best dynamic range/ low noise/ color? I seem to like portrait mode with everything else off.

It would be really nice to know how far it can go before over exposure, and, is it linear? The nice thing is, you can always snap a still pic and if you have the blinking over-exposure feature on, it will tell you if anything is over. I love how it takes a still with the exact same settings as the video. The GH1 was annoying because it didn't do that.

Barry_Green
10-22-2009, 09:51 AM
With any luck, Tramm Hudson will have a "Magic Lantern" out for the 7D soon, with real zebras, and that will put all this guesswork behind us and let us get back to shooting, and able to tell immediately what's too bright or threatening to go over...

mhood
10-22-2009, 10:03 AM
None of you find the Highlight Alert useful? Coming from a "waveform/vectorscope background", I'm having a hard time getting anything useful from the histogram.

USLatin
10-22-2009, 01:22 PM
Barry, so do you think we might get a live histogram? Was that enabled on the 5D? Or could it be too much processing for the camera?

Barry_Green
10-22-2009, 01:44 PM
In Magic Lantern? I thought ML added histogram and zebras and all that stuff. I don't really know, as I haven't used it, it's only available for the 5D right now and I've got a 7D...

mhood
10-22-2009, 01:51 PM
Highlight Alert is a zebra of sorts...it just takes clipped specular highlights to black. I get a lot more from it than I do a histogram. Any ideas where to go to learn how to read a histogram?
And, yea...live would be nice. Now, I'm taking a still and checking HA and HIST. I can also shoot a little and they both refresh when I pause a playback.

USLatin
10-22-2009, 03:42 PM
Yes, I am aware. I know what you are up to Barry, I am forum STALKING you!!! ;)

I too got a 7D. Haven't used a 5D at all.

Mhood, histograms are very basic, the X axis are the "luma" values from 0 to 255 to 100% or whatever. The Y axis I guess represents the amount of pixels that are in that "luma" range. Barry, please correct this if I am fudging it up. So you basically see where your image is. A graph that doesn't have any values towards the very right means you don't have anything close to clipping. A graph that is filled all the way to the edge means there is some stuff all the way at 255 or 100% or 105% or 110% or whatever the graph goes to.

Just google histograms and you'll find a gazillion explanations like this one for example: http://www.luminous-landscape.com/tutorials/understanding-series/understanding-histograms.shtml

The best tool IMO is the separate RGB histograms giving you one histogram per channel which also lets you know if you have any major issues with any one particular channel, but you can just look at the exposure histogram and look at the image in the screen and you are in very good shape already.

Barry_Green
10-22-2009, 03:48 PM
Separate histogram is better than composite, but I prefer a waveform monitor over a histogram, because a waveform monitor will not only tell you if you're clipping, it'll tell you where you're clipping (or crushing, or if your overall exposure is low, etc). In some ways a waveform monitor could be thought of as a 3-D histogram.

USLatin
10-22-2009, 04:26 PM
Yea, but there is no hope for one on a 7D, unfortunately.

That is what the Panny LCD's are for though. :)