December 18, 2014, 07:50:43 PM

Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - Policar

Pages: [1] 2 3 ... 30
1
How Canon can ignore the market when the technology is there is beyond me. They created an entirely new product and idea accidentally, but are now intentionally turning their back on it. IMO they will come to regret this decision with Sony, BlackMagic, and the likes dominating. Why would I purchase a C100 for twice the cost of a A7s? The two have offsetting features/outputs, but ultimately the Sony delivers far more value and image quality.

Then get the A7s.

Canon isn't trying to market toward the cutting edge of hobbyists, but toward wedding videographers, low end corporate, etc. That's why their mirrorless segment sucks for stills, too; they're focusing on products designed for professional shooters, who don't need that extra DR but do need that killer autofocus and optical finder and lens compatibility. When they throw that out, you get a sorry consumer product in the EOS M. Whereas the 7 DII looks fantastic.

That said, the A7s has a dreadful interface and ergonomics for shooting video, poor lens compatibility, a non-standard codec that's less industry-friendly (if arguably better) than on the C300 or C100, and Sony's SLOG 2 implementation is awful with over-saturated highlights that can't be fixed in post, odd skin tones, and too much dynamic range for an 8-bit wrapper. And the skew is just dreadful, not a nice combination for a lightweight camera. I find the image far inferior subjectively to the C300, but technically it is great and the low light knocks everyone's socks off.

I spent a lot of time with the F5 and while the specs blow the C300 away, the user experience was really poor. Bad timecode sync, ugly colors, magenta skin, ugly grain structure, SLOG 2 sucks (SLOG 3 is better so kudos to Sony there!) wheras WideDR and the current Canon Log color matrices are really nice and very pretty, etc. It was like driving a car with a lot of horsepower but terrible handling.

Canon's the market leader so they're focusing more on boring "pro" features that make the camera easy to use and with footage that's small, easy to ingest, and edit, and can even look ok straight out of the camera. WideDR looks great and has lots of DR; EOS standard intercuts with Canon dSLRS, which are still ubiquitous, even on network tv as b cameras (seriously). Sony has great specs but takes more work to make it work and it's not there when every MB and transcode costs your company serious time and money. For an enthusiast, seems like a no-brainer to get the Sony, the low light is incredible and if you avoid clippy saturated highlights and skew then the image is very competitive and better than Canon dSLRs for sure (if you can grade well). But for bread and butter shooting the C300 still owns the market segment and there are very good reasons why it does. The image is beautiful and really easy to post with and requires low end hardware and not much support to use. Great single operator camera, I see them all over. Canon is a lot like Apple: poor specs for the money, but great user experience and system loyalty for that reason. Arri even more so. The highest end cinema camera still has the lowest resolution sensor, but the Alexa is a dream. Curiously, Canon's lowest sensor density camera is also its most professional, the 1DX. Nikon's D4 is even lower-res.... Maybe not so curious. Pros pay for hard drives, they don't sell more photos if they have more resolution beyond the baseline needed.

Black Magic seems cool, but I find their products horribly bad for ergonomics, and the 4k has static noise like crazy and just a bad interface whereas the 2.5k aliases like nuts and also has a bad interface. The pocket camera seems cute, though incomplete. Cool studio camera for well-lit green screen shots, but not as flexible by any means as any of the competition.

But yes, Canon did abandon anyone who wants "cinema" video IQ in a dSLR for cheap (whereas Sony and Panasonic made some great strides here), and they did so to support their cinema line. The Mark III isn't terrible and the 70D introduced really cool AF, but the image didn't get dramatically better. No, it wasn't very nice for them to focus on professional gear instead, but so far it has paid off financially. To be honest, I also think the C300 is a great product, just pricy. Sharpest 1080p on the market (sharper than some 4k and probably as sharp as the Alexa's 2k tbh definitely sharper than Alexa 1080p), great colors, easy codec to handle in post. I still hope the C300 Mk II is strong enough that Canon isn't afraid to improve their image quality on their still cameras. The 5D Mark II still has a nice look, and Canon does great color processing. Would be cool if they improved the specs just a bit, I agree, but they will never again be cutting-edge, at least not intentionally. The 5D Mark II was a fluke. Also, 120fps... that would be sick.

2
Lik's photo is a tone-mapped clichéd mess;
Probably you are right.

Quote
Gursky's is extraordinary.
Please tell me why? I really (honestly) would like to understand.

Gursky reminds me a bit of David Fincher. You get this really technically perfect (imagine that photo except wall-sized and with the equivalent of like 400 megapixels of detail) cold, clinical image that's still beautiful. I don't think this is Gursky's best photo at all, but try framing up a shot of a river that has that clean a composition and that symmetry. You would never walk by that scene and see something so perfect (and I'm sure it's digitally manipulated, but I can't tell where...) but it also feels naturalistic. It's a very elegant, beautiful look at something more mundane.

Gursky is a little crazy and he gives you this very high camera angle usually looking down on something and it's very clinical and cold (and usually shot in 8X10) so of course it's where you put the camera and how you present the subject matter. I think Terry Richardson is a great photographer if you're trying to shoot everything from a predator's POV, for instance. Gursky's POV is very cold, superior, organized, and it's not easy to emulate, though it's trivially easy to identify his style.

Lik is a great brand-maker, but his photos are just overbaked landscapes. He's not incompetent, but his stuff wavers between decent and garish.

It's kind of silly that on one hand you have someone who's so cold and intellectual and on the other you have someone who's maudlin and basically the Thomas Kinkade of photographers. Kincade and Mondrian, maybe... But Gursky is better than Mondrian.

I'll take Vermeer or Velazquez. :) Dig the hell out of Gursky but he's not my personal favorite, just think he's very accomplished and does great stuff.

I do think Gursky's stuff looks much better LARGE, as it's meant to be, whereas Lik's photos have a nice thumbnail over saturated/over-tonemapped POP that looks great on Facebook.

3
I just don't understand the value put on this photo.  'Bella Luna' is hideous. But he does good work in general.  I could see it fetching 100's, maybe 1000's at most.  I guess it's all about marketing.

"Bella Luna" is hilarious. He didn't even use the right transfer mode when he superimposed the stock image of the moon. And the composition is poor... everything about it is awful, but hilariously awful.

This is just a fairly good photo that's been tone mapped a bit too aggressively and put into black and white to make it art. I keep forgetting about that one. This isn't nearly that bad.

4
Lik's photo is a tone-mapped clichéd mess; Gursky's is extraordinary.

5
I never found my old 70-200mm f2.8 L to match the test charts, so I traded up to a brand new 70-200mm f2.8 II IS, but it's... not that much better. It is better wide open, less "weird," and possessing less character, but it's not in the same category as the Sigma 18-35mm f1.8 for tack-sharpness wide open...

It does seem to have good IS (for stills, not video as much) and great autofocus and build quality (so the the original) so I can't say it's a bad lens at all. The IS alone really made the difference, at least for stills, but it's no "miracle" lens as the 200mm f2 and Otus and recent Sigma Art lenses appear to be.

6
Hideous tone-mapped mess, but clever technique with the thumb thing.

7
Introduce a lot of backlit clouds or street lights at night or even go deep in a forest with just a few rays of light or peer into a cave or balance between the inside of a house and the outside without much light.... and you're looking at valid 14+ EV scenes no problem.

In other words, you'd still be clipping or blocking with an Exmor sensor.  It'a a point that I've made before...~13.x stops of DR is better than ~11.x stops of DR, but the number of scenes that have more than ~11.x but less than ~13.x stops is far exceeded by the number of scenes with <11 or >14 stops.  In jrista's example with the a7R home interior shots, the windows still had blown highlights despite the greater DR of the Exmor sensor.

Shooting with the C300 and Alexa side by side, you soon see just how many scenes fit into the 11-15 stops of DR range. But that's for video. I still disagree for stills, a little extra detail can't hurt. Clipping isn't so bad in a very high contrast scene, but clipping later rather than sooner is best.

Canon's sensors are still fine for what I need, I don't see all the fuss, but I'm aware that there are those to whom a little extra shadow detail matters and try to respect that.

8
A stop or two off is a HUGE error if you're metering right. If you're getting that kind of error from metering, Canon really does have a problem! When I shot 135, my F4 could expose properly for Velvia, which has +/- 1/3 stop exposure latitude, when set correctly, though I'd still spot meter (with an external meter) to check.

You can't tell, but if you've been spot metering scenes for years (which any of us shooting MF and LF I'm guessing have) you can easily guess... and that scene has around 8-9EV of meaningful dynamic range (even if you exposed wrong it's only the white sign, maybe the reflections off the sand, or the black shirt in shadow–nothing important to have tons of texture in–that would clip) and the 10D should have no trouble with it even if you expose a stop wrong one side or the other.

What I meant is that the contrast on a print is 4-5 stops of contrast at most, usually 4. For the best print. So you either have to tone map (cheesy) or shoot a flat image if your scene has a lot of contrast. Or wait on good (flat) light. That's why Velvia had 4-5 stops of DR and made for the best landscapes. It's funny everyone wants all this dynamic range for landscapes when ideally landscapes were something you'd approach trying to get as flat in-camera as possible because you knew the final print would only have so much contrast. Or the zone system accommodated 10 stops of DR, but Ansel Adams' B&W prints would look a bit cheesy tone mapped that far in color, let's be honest, and not many of us have surpassed his technique even digitally! But a computer screen has a 1000:1 contrast ratio nearly (9-10 stops of DR) so you can shoot higher contrast material and still have it "pop" like slides on a lightbox with less tone mapping.


9

Shoot that scene on an Exmor and set the jpeg contrast too high and you'll get the same result. A 10D would cope with that.

You can screw up with all gear, that isn't a reason for or against anything. However, if you shoot this with a d750 and 13.9ev of dynamic range, my guess is that it's much easier (or possible at all) to keep the bright sand from clipping while preserving detail in the shadows - 3ev is a big, real world difference.

As for the 10d - you're correct, at iso 100 it has nearly the same dynamic range as a 5d2 or 70d (11ev) which tells us about how Canon sees the tradeoff resolution vs dr :->

Disclaimer: I totally love all my Canon gear and feel you can produce great pictures with it. Unless you shoot movement in harsh daylight and cannot bracket, that is.

That isn't a terribly challenging scene to expose. The contrast ratio of noon day sun is about 4-5 stops and the contrast ratio between the sand and darkest parts of the scene are 4-5 stops, too. So we're looking at 10 stops of contrast at most, likely less, with the exception of specular highlights kicking off bits of sand. I agree that a 10D could handle it.

I guess if you dramatically underexposed that shot it would look better on a Nikon sensor, but you'd have to be really incompetent.

Introduce a lot of backlit clouds or street lights at night or even go deep in a forest with just a few rays of light or peer into a cave or balance between the inside of a house and the outside without much light.... and you're looking at valid 14+ EV scenes no problem. Of course they would look better lit than tone mapped. :)

But that's why digital cinema cameras have gobs of DR (sorry, the Alexa's 14 stops eat the D800's 14s for a snack and even the C300 has more highlight detail than the 5D at least it's distributed to favor it) is because you're dealing with tricky situations without the potential to use strobes and without the time to wait on light and you're viewing on monitors/tvs/retroreflective movie screens, which have a lot more contrast than print.

But now that we see photos on our computer screens, I can see the need for a little more DR. Dealing with 10 stops of contrast on a very good screen vs 4-5 on the best prints available.

10
They're using JPEGs it looks like, and even admit that real-world testing is inferior to the 12.7 stops they found on the test wedge...

And yet they're more credible than a company measuring raw sensor data? Why? Oh right, you want to agree with them.

Who cares? If you need more DR get a Nikon, few people are regularly shooting scenes with 14+ EV.... it's hard to make that look any good on the print. Remember zone system is designed only for 10 stops and TBH Ansel Adam's work is pretty heavily baked by any standard! Those going much further go into some deep tone mapping nonsense more often than not.

I wouldn't trust this review if DR is the most important factor to you. If it is the most important factor, I'd ask why.

Not to quibble but it's not correct to equate the zones in the zone system to stops. The Zones represent tonal values and the change in the values between zones varies and are not logarithmic steps the way stops are. A scene with 6 stops of DR or a scene with 20 both map to the same ten zones. The zone system is about how you map the DR in the scene to the ten zones using expansion and contraction of the tonal scale.

It is all a bit "overbaked" by modern standards but they were doing the best they could with the crude tools they had available. The concept still has value however.

BTW in the Sierra Nevada 20 stops of DR in a scene is not uncommon. Hence "the Range of Light". Adams was compressing that to prehistoric wet plates and printing papers. It was an impressive technical achievement regardless of how you feel about his pictures artistic merit. Personally I'm not always a fan.

Not usually a nit-picker but its an important distinction. The histograms we all use today are a modern evolution of Adam's Zone System. When you move the sliders in lightroom to fill out the histogram, that is the zone method's modern equivalent.

A stop correlates with a zone, all this talk of steps, stops, EV, and zones sounds like a mess but really they're pretty interchangeable. Ok... a zone is a stop of reflective rather than incident light. A stop through the spot meter. :)

I agree that Ansel Adams overbaked, but he overbaked well. His work is among the best HDR I've seen. :)

I figure if you're printing on paper that can render 4-5 stops of contrast at best you'll run into trouble once your camera reaches for more than 5-6 (velvia, for instance, got it well). But since B&W reduces information significantly you can get away with a 10 stop zone system with real artistry and it still reads on print...

Monitors show a lot more contrast. I can see the need, but I don't need more than 12 stops unless I expose wrong.

That said, this figure is taken from JPEGs with noise reduction. It's not accurate. It doesn't matter to me, but if it matters to you how much way extra DR you have buried in shadows, Nikon does offer more.

But Ektar offers more than Velvia and yuck lol. Thought Portra is quite nice. :)

11
They're using JPEGs it looks like, and even admit that real-world testing is inferior to the 12.7 stops they found on the test wedge...

And yet they're more credible than a company measuring raw sensor data? Why? Oh right, you want to agree with them.

Who cares? If you need more DR get a Nikon, few people are regularly shooting scenes with 14+ EV.... it's hard to make that look any good on the print. Remember zone system is designed only for 10 stops and TBH Ansel Adam's work is pretty heavily baked by any standard! Those going much further go into some deep tone mapping nonsense more often than not.

I wouldn't trust this review if DR is the most important factor to you. If it is the most important factor, I'd ask why.

12
EOS Bodies - For Stills / Re: 6D photos on a 4K TV monitor
« on: November 14, 2014, 01:10:12 AM »
All things equal, yes, but the 4k display might not have the same contrast ratio, refresh rate, color gamut, calibration, etc.

So unless the pixels are really the big problem, and you're willing to risk (maybe lose, maybe not) contrast/gamut/whatever, I wouldn't go for 4k quite yet. But when you find the right panel, yes, it will help.

13
Canon General / Re: Canon Cinema EOS C300 Mark II Will be 4K [CR2]
« on: November 09, 2014, 04:41:25 PM »
This rumor make me worried for two reasons.

1) If they introduce 4K with the C300 Mark II, then it seem doubtful that 4k will find its way to the relatively cheaper models. I don't desperately need 4k, but it would nice see it in a Canon body that doesn't cost $12,000. So for example I wouldn't expect them to give the 5d Mark IV 4k, in order to not cannibalize the C300.

2) The idea of a shoulder mounted production camera is very interesting, but if they are targeting the same marked as the Arri Alexis, then it's not something that would be relevant to everyone here, if you don't work with high budget productions. And I think that is a damn shame. I love the idea of big sensor camera for documentaries that I just can take out of the box and on my shoulders, and then go out shooting.

It the same thing Sony are trying to do with the FS7, but I would love to see Canon do the same.  But of course this is just rumors, so I can still be pleasantly surprised.

Canon is trying to position itself like Arri (and Red... sort of) as an "industry leader" that doesn't need to compete with specs so much on image, trust among producers and DPs, and not on the basis of specs. Arri doesn't offer 4k (even at 60k); red doesn't offer it for under 15k realistically for 4k-ready package. Canon is at the low end of the industry in terms of specs, but also prices for a well-liked professional solution. The F5 didn't catch on great, but it's as expensive as the C300.

The C300 rents very well. I find the IQ from the camera awesome, I far prefer it to the F5 despite the F5's better specs.

The A7S and GH4 are not gaining traction among serious shooters due to reliability issues (even the F5 has poor timecode sync); the IQ is there and if all you're after is IQ just buy one! Canon won't cater to you, the same as Arri won't. Stop waiting. Canon is after the professional market, which demands less in terms of image quality and more in terms of reliability and conventional workflow (which is where Red stumbles and Arri does best, despite... only 2k on the body).

Listen, I actually don't care about IQ that much. As a video-journalist I want a new C300 that I can comfortably put on my shoulder without building a rig or the need of a external evf. Mix in broadcast friendly codecs, sell it at $8000, and then I am a happy camper.  If this rumor is true, then Canon will not make that camera.

It sounds like they're making that camera, it might just not be the C300.

Would be cool if it were.

14
Canon General / Re: Canon Cinema EOS C300 Mark II Will be 4K [CR2]
« on: November 09, 2014, 02:46:40 PM »
This rumor make me worried for two reasons.

1) If they introduce 4K with the C300 Mark II, then it seem doubtful that 4k will find its way to the relatively cheaper models. I don't desperately need 4k, but it would nice see it in a Canon body that doesn't cost $12,000. So for example I wouldn't expect them to give the 5d Mark IV 4k, in order to not cannibalize the C300.

2) The idea of a shoulder mounted production camera is very interesting, but if they are targeting the same marked as the Arri Alexis, then it's not something that would be relevant to everyone here, if you don't work with high budget productions. And I think that is a damn shame. I love the idea of big sensor camera for documentaries that I just can take out of the box and on my shoulders, and then go out shooting.

It the same thing Sony are trying to do with the FS7, but I would love to see Canon do the same.  But of course this is just rumors, so I can still be pleasantly surprised.

Canon is trying to position itself like Arri (and Red... sort of) as an "industry leader" that doesn't need to compete with specs so much on image, trust among producers and DPs, and not on the basis of specs. Arri doesn't offer 4k (even at 60k); red doesn't offer it for under 15k realistically for 4k-ready package. Canon is at the low end of the industry in terms of specs, but also prices for a well-liked professional solution. The F5 didn't catch on great, but it's as expensive as the C300.

The C300 rents very well. I find the IQ from the camera awesome, I far prefer it to the F5 despite the F5's better specs.

The A7S and GH4 are not gaining traction among serious shooters due to reliability issues (even the F5 has poor timecode sync); the IQ is there and if all you're after is IQ just buy one! Canon won't cater to you, the same as Arri won't. Stop waiting. Canon is after the professional market, which demands less in terms of image quality and more in terms of reliability and conventional workflow (which is where Red stumbles and Arri does best, despite... only 2k on the body).

15
Canon General / Re: Canon Cinema EOS C300 Mark II Will be 4K [CR2]
« on: November 07, 2014, 04:09:29 PM »
It'll be 4k. Hopefully 10 bit codec. Probably the same sensor.

Pages: [1] 2 3 ... 30