maxotics

Nice. "Raw requires less processing power than compression. It is just a dump of data to the card. So all you need is a fast card interface. Easy." Andrew, how can you, someone who has never built a piece of electronics or programmed anything say it is "Easy". Sometimes your ability to tell people they "talk out of their arse" is staggeringly fitting to someone else I know "LOG curves are not all problematic." Only if you shoot buildings and sunsets like you do.

deleted

In a strange way, Andrew, you're making the argument for the other side. If computer technology isn't powerful enough to put video editing in a tablet, or smart phone, why should it be powerful enough to read and write RAW data to memory cards (which would remove the need for problematic LOG curves). More powerful chips would mean faster WiFi connections, or multiple transmitters, etc. The camera companies have done amazing things with the electronics they have. The argument that they aren't innovating fast enough, or well enough, is difficult to buy when, as you point out, there are serious speed/power/battery obstacles in all imaging/computing technology. Steve Jobs was all about simplicity, aesthetics before function, and a wager that people are generally not to be trusted with advanced, open, OS's. He was right, but sadly so IMHO. BTW, after the above skirmish, I looked at Cantsin's videos on Vimeo. Very nice stuff!!! Whatever the merits of his arguments, he knows how to make film.

Having lived through the PC/MAC years I don't remember it as a slam-dunk. Keep in mind that Gates did the same thing with IBM. Hindsight for all this is 20/20. I see the opposite with modern cameras. What they've been able to achieve in the past 10 years is nothing short of astounding to me. When I get into film in 1970s, the technology had been more or less the same for 40 years, the high end being 16mm and low-end 8mm. Think about that, if the tech we were using today was petty much the same as what we were using in 1975! Yet, I have a Panasonic LX100 in my bag that shoots as well as, if not better, than the quarter-million dollar cameras in "Phantom Menace" 15 years ago. Better WiFi would be nice, but keep in mind that an app can't access the password to the wireless network and the only real solution is to build a second transmitter and that's not up to the camera companies (it's up to the smartphone makers). That it works at all, given the constraints, is to be applauded. The number one problem all cameras make is light sensitivity (as it was in film). Software can only do so much. Anyway, there is a rich ecosystem of camera software out there, though not from the camera makers. Building cameras and building software are two different skill-sets, IMHO. The biggest problem today is people don't spend the time to experiment and learn how to use the technology they have. Most people put whatever gadget they have on "auto." There needs to be more respect, love, and FUN with THE ART OF PHOTOGRAPHY/FILM. This blog does a great job at that, except when it looks at the glass as half-full and gets into arm-chair business discussions that are better left to the nit-wit junior financial analysts down in The City

Nice work!!!!

I want the 4K to get good 1080, so not an option for me. Thanks! I looked at some footage I shot last night with both the A7 and the LX100. I'm now on Dishe's bus. I lose resolution with the A7 but the colors and bokeh are out of this world. I'm going to sell my MFT gear (except for the BMPCC). Like dishe, I've always had a problem with Panny color... I've forgotten. The problem I have with the Pannys is there isn't a great upgrade path that uses the lenses I have for the Sony. The A7 with 55mm/1.8 is a killer setup. An A7S plus the new 90mm and I'd have a seriously long-shot combo (I also have the KM1 audio adapter and 10-18 for long shots with A6000). Anyway, I can use the lenses on an FS7 if I go full 'tard My friend's getting one. The video on the plain A7 is fine for me for now. This has been the direction I was heading until I did that test shoot last weekend. I've always likened sharpness to the "salt" or "sugar" of cinematography. It easily sways you away from more refined tastes! I succumbed. If I really need the sharpness later, I'll buy the A7rII or something. With Panny, I can't upgrade to be saturation and shallow DOF.

I'm trying to f__ you up in the head and you're doing it to me I haven't seen any good comparison of the A6000 to the A7S. If, as you say, the A7S has as good 1080 as the downsampled LX100/G7 then it would make the A7S very appealing as their prices drop in the used market. Anyway, if that's what you're seeing, to answer your original question, I'd go with the A7S and bag the Pannys. Even though it is more expensive, it would be worth it because MFT, IMHO, don't hold a candle photography/still wise to MFT and, as you know the colors are much richer and shallow DOF better. Hmmmm..... I was thinking MFT for video, but now you have me turned around!

Dishe, one of us is horribly wrong I was wrong the first time, I'm hoping to even the score! With the Panny Image app I can manually focus the LX100, With focus zoom. I can zoom the lens in an out. I can start/stop 4K recording. I can view the scene LIVE on the Android app! I can do NONE of those things, in VIDEO mode, with the A6000 or A7. Can you do that with the A7S? The A5100 doesn't have an EVF so isn't a fair comparison to the GH4. Put it next to an LX100. There's a lot of confusion on this blog, in my opinion, about what makes for contrast sharpness and and what makes for color sharpness. In low light, the GH4 will be detail sharper than the A7S, but color blurrier. Or, the A7S will be color sharper, but detail softer. I don't have an A7S, so you'd know better than I which the better trade-off is. Again, I LOVE both systems, Panny and Sony. You're comparing an old Panny camera, the GH2 with a new Sony A7S. As I mentioned above, if money is no object, I'd get some A7S's or IIs and call it a day. However, for the money, the Pannys deliver a fantastic value in video--if you look at their current 4K cameras!

I think it on topic because dishe seems to be going through, what boils down to me, MFT of APS-C. I've never had any interest in 4K as 4K. As I mentioned, 4K is now 'real' 1080. The A7S, with it's larger pixels REALLY confuses things because its fantastic saturated and moire-less image is very appealing. A problem I have with the LX100 is at 70mm it really isn't long enough to shoot over-the-shoulder shots (which I imagine you'll want too). It seems to me, two G7s with the 14-140mm would be a good approach. One can get two of those, with lenses for $2,000. It would cost $6,000 for similar A7S setups. The problem with dishe's goal, which I'm sure he understands all to well, it that it gets expensive. IMHO you should have double the camera bodies expense in lenses and the same amount in lights. So a $1,000 body should end up as $1,000+$2,000+$2,000 = $5,000. Yes, shooting out of my a__ here. But what I remind myself! I loved the colors you got out of the NX. It handled the clouds very nicely.

You're right, my experience is with the A7, NOT the A7S, and you're right the less megapixel count of the A7S should allow it to run longer. Sounds like others haven't had a problem with shoot length, so I'm simply wrong I have both Panny and Sony. I'm also trying to chose between them. I currently have an A7, A6000, LX100, G6, GM1, BMPCC (I've owned a GH4 but sold it, though might not have if I didn't get into that bullshit "Log" craze which was more harm than good in what I was doing). Anyway, the big picture is Sony has more saturated colors and (less noise), shallower DOF and better autofocus than the Pannys. The Pannys are sharper, smaller, easier to use and have better remote apps. For stills I love the Sonys and have never been able to embrace MFT. For video, Pannys are small, idiot-proof and have never let me down. Anyway, I'm trying to decide between the A6000 and G7. I believe these are the best bang-for-the-buck cameras going. I can afford the A7RII or A7S, but feel the extra money would be better spent on lighting or audio equipment. With the A6000 I'll get best autofocus (the thing is amazing), rich colors and shallow DOF. However, the G7's 4K downscaled to 1080 will give me perfect 1080 really (whereas the A6000 is still more like effective 720). When I look at the LX100 compared to the A6000 they are both super great cameras with different strengths. Right now I'm leaning towards the LX100 because the image is so sharp and clean (even though in photography I'm more of a color/dof guy). Here's a video where I use some of these cameras https://vimeo.com/145871944

Personally, I would never depend on the Sony sensors to run for hours, let alone tens-of-minutes. I shot some stand-up with an A7, and even with turning the camera off for a minute between comics, after about 4 of them, it cut out in the middle of the 5th. It isn't a battery issue. It's a large sensor uses a lot of heat. I believe it will be years, maybe never, before large sensor cameras can shoot video as long as the Panny MFT cameras. The GH2's don't have much re-sale cameras now, so you don't have much value there; just the way it is. However, they are excellent cameras for long-form events like concerts, lectures, etc., and still produce a marvelous image, sharper (or as sharp as) Sony because less space between the pixels. As you know the Panny cameras are pretty idiot proof and reliable. You should feel lucky you have two cameras to shoot those kind of events. Just create two bags, one for corporate events, and the other for "filmmaking"

Here's some test footage of the A6000 and the 10-18/f4. Sorry it isn't better. IMHO you can't do better at 15 to 24mm with any camera, stills or video than that combination. Maybe I shouldn't beat around the bush. I LOVE LOVE THAT LENS As for the pistol grip stabilizers. They work fine with small cameras. Mine is lent out to a DP who is at the low end of the technical spectrum and he's shot at least 4 shoots with it with no problem. He always shoots the A6000+35mm+wireless mic. Many of the problems people have with them is that they try to get many camera/lens combinations working with them. You're right in that you can't quickly adjust them. Again, if you always shoot a prime with a small camera on it, the pistol grip gimbals are fine. Sensor stabilizers don't match the dreamy feel you get with a gimbal. Completely different looks/feel. I know you know that. You can do lots of incognito stuff with the Nebula you can't do with the Ronin. Remember to post your first videos using you new combination, whatever it is!!!!!

Another setup to consider, all in one bag, is the A6000 + 35mm/1.8, 10-18mm and the wireless mic attachment ECM-W1M (which you can clip on someone, attach a lav, or just put on a table near the action). I've used it a lot on the Nebula 4000 gimbal. For 4K consider adding the RX10 IV to your bag. You can buy all that stuff, some used, at the price of one of those bodies alone. I don't know about the new A7s, but the A6000 focuses better than my A7. If you're not going to shoot stills, the A7s are overkill. Here's one sample https://vimeo.com/121428480

There's no easy solution. I've lately standardized on the Sony A7 because my eyes are old and I like the EVF, focus zoom, small size. However, with the 5DII I'd get faster menus, autofocus, ETTL flash, Magic Lantern RAW, no worries with Tilt-shift lenses, plus great Canon lenses to choose from.(not to mention a practically free 50mm, nifty-fifty). If you're going to use manual-focus glass, I'd go with the A7 with EVF If you need fast auto-focus for weddings, events, I'd go with the 5D2 There's a lot of Canon and Nikon bashing on this focus from people who don't take stills for a living The best solution is to steal your friends D800 !!! That's a camera! Another idea is to get a D600, though its video is horrible. If you can't afford lenses, get the camera that will work with your friends lenses so you both have options!

Very interesting! Thanks! Some further explanation may help some readers. Essentially, the CPU does the rendering and your GPU (graphic card) does the playback. If one looks at the data above they can see that they have little effect on each other. Because there's only so much parallel processing that can be done on a render, a faster CPU clock speed / data bus will shorten render times? Yes? So better to have 4 core at 4.5MHz, then 8-core at 3.2. On optimization, depends on what you want to optimize, right? If you're more interested in viewing 4K playback with effects, the more powerful GPU you have, the better. And it doesn't really matter if the GPU is above 50% because that's all it's doing, it shouldn't effect CPU business, yes? So it's less important if you have a slow CPU. If you want to optimize your render times, then a liquid cooled, RAID 0, over-clocked CPU is probably your best bet (though I have no direct experience). The latest i7 chipsets give maybe 5% improvement, from what I've read, so one shouldn't pay a premium for the absolute latest. Also, I think you may mean for maximum output quality as an intermediate clip, use Cineform YUV10bit. For final render, that should be up to the destination requirement? Finally, another point of your analysis may be that laptops simply do not have strong enough GPUs for heavy 4K editing yet. People may not realize even if they have the same number card, they're "M" versions, with less memory and power. So you may want to use a desktop. Again, very interesting! THANKS!

Why didn't you ask the manufacturers, or find some experts and ask them before writing the article? There are two components to perception, contrast and color. We also have two types of visual cells, cones (color, about 5%) and rods (contrast, about 95%). I believe the reason for the predominance of green sensels is that they most accurately capture contrast, or dynamic range, of light, regardless of color. I have NEVER met a normal person who said they saw a color difference between cameras. I feel I see it, but when I take time to color correct I can generally get one camera to match another one. The log profiles are major destroyers of color data. Anyway, each camera tweaks its JPG output to its taste--espcially in H.264 where so much is baked in that can't be baked out. In RAW, I believe there is little difference. If there was, they would probably come out through analysis with Imatest software.

I might be. But if so, you'd need to buy one of the expensive ones. A tilt-shit lens works by bending the light across the sensor. Sort of like a hose that sprays one side of water onto the sensor, and the rest off to the sides. The baffle in the adapter bounces the light back into the lens. Of course, these issues are meaningless for videographers.

Last I checked, you can't even get remote (WiFI) VIDEO viewing of the video image through the PlayMemory apps, like you can using Panasonic cameras. The Panasonic implementation is amazing. You can set focus on your smartphone when in video/photo mode. I haven't look at the SONY API lately, but my guess is that any video LUT application is not going to happen. Unlike Panasonic, SONY puts the app on many kinds of cameras, from action cams, camcorders to their mirrorless. So you have a lowest-common denominator thing going where they only provide API calls that work across all cameras. Also, video seems to be the lowest priority in the API functions. Here's their list https://developer.sony.com/develop/cameras/device-support/

I know. I didn't mean any offense either! When I'm in a bad mood about work, and haven't had much time to actually use my cameras, I get argumentative I shouldn't have made that last dig. Sorry! Gelaxstudio, the problem I had with an adapter and the Canon TS 24 II was that there was a lot of light flare hitting the sensor. I believe this is a general problem of all wide-angles on adapters with the A7. In this photo, I dodged it out best I can, but you can see it in the right of the image Another problem, I've read, is that the sensor is a bit more shiny than other sensors and creates issues. I also want to point out that though you can remove hot spots in post, the fact that you're getting them means the sensor is hotter than it should be and is actually degrading the performance of all the sensels. All this isn't to argue against getting one of these cameras! Just something to think about. Consider also that most of the battery power in the BMPCC goes to peltier cooling. The more work the sensor does, the hotter it gets. This makes me believe there will never be RAW on the A7RII. Again, I'd love to get that camera. But if I get into architectural photography again I'm getting a Nikon or Canon with a native TS lens.

Is that fair? I think my info is artistically correct too. Is it fair to imply I think your "mates" have errors and took images that "look like crap". It IS NOT just pixel peeping. I can't speak for other Fovean users, but for me I see the difference even in a web-image. The main problem anyone ever has with Fovean (including me) is the slow read/write times of Fovean sensor data, which leads to a slow camera/post processing. Since you have had experience, are you telling me that you would not trade any Bayer camera for a Sigma if you could get those images as fast and easily and at high ISO? Again, my main point is that for some photographers high megapixel counts are necessary to deliver them the kind of color fidelity they need for their work. Even if it's 1% of users who see the difference, does that diminish what they're trying to do? Isn't that what we strive for anyway, in Art, to become that top 1% for someone? As others have posted, and unfortunately, been ridiculed because their "1%" wasn't following some party-line, the problemw ith the A7RII for stills is the floating sensor. Sure, for hand-held work, that camera is amazing. For tripod work, I don't see it as better than the Canon 5DS or Nikon D810. It's technically difficult to keep gimbaled sensors aligned (slight focus errors) and dissipate heat (which creates hot pixels). Anyway, I was just supporting your original post that there are subtle differences in these cameras that should be considered for whatever work someone wants to do. I had no idea you weren't really looking for knowledge, but only giving it

Another benefit of high megapixels is improved color fidelity. Because of the arrangement of the bayer sensor RGRGRG.... line one, and GBGBGBGB line two, you always need at least 4 sensels/pixels to get one full-color pixel. A 24 megapixel camera is really 6 megapixels (6x4) for a discerning photographer. A 36 megapixel is 9 megapixels. Canon didn't come out with a 50 megapixel camera just to win some marketing war, or to solely make cropping easier. Photographers who care about color accuracy on the pixel level need all the sensels they can get. So it's more than just cropping that weakens the desirability of the A7S, at least for me. Yes, low light is great, but any hard line that goes through the image will pick up color distortions, faint colors where a red pixel picks up the line and interpolates around green and blue pixels that didn't--which also leads to moire and aliasing issues. There are some images in this article that show the effects I'm talking about. I'm comparing a Sigma/Foveon sensor camera, that doesn't use a Bayer method of color sampling (it uses a vertical sampling like film) compared to a bayer sensor. In the image, you can see how gray shading in a line on fabric picks up color aberrations. http://maxotics.com/2014/07/12/sigma-dp2-quattro-review/

Sorry, don't remember. I'm pretty sure I always shot 24fps.

My view is the 50D is better than the 7D where Magic Lantern is concerned because the 7D uses dual DIGIC processors which seems to cause some complications for ML devs. My reading on the ML forum is the 50D may even be better in some respects for ML RAW. The drawback of the 50D is that it doesn't have native H.264 WITH audio which you can get with the 7D. Of course, you can get H.264 with the 50D using ML, just no audio. Amazingly you can get HDMI out on the 50D too. You can now get a 64GB Lexar 1066x CF card for $90 which will store 64/4 = 16 minutes of RAW footage. ML Raw Viewer will then do a quick grade and write out ProRes, or DNGs, etc. I'm now seeing 50Ds for $300. So for $300 + 50mm "nifty-fifty" and a fast CF card for $500 you have a RAW camera package that will blow you away (blows me away). The 50D can only do 1568×882 resolution in non-crop mode. No PINK DOTS! And again, anyone thinking H.264 1080p (even 4K) can hold a candle to 1568x882 RAW in color depth/dynamic range has not seen for themselves Some moire in certain situations, but if not, BEAUTIFUL! You can get near 1080 in crop mode, so you'd want the 10-20mm lens. Here's some footage from 2 years ago when there was no ML Raw Viewer and computers still too slow. If I shot H.264 the skies would be blown out and the skin tones would have been over-saturated, a bit blotchy. This look isn't for everyone. Not saying H.264 is bad. Just a matter of taste.

Sold! Where do I send the $2,000.

A difficult question to be sure! It depends on what you define "detail" as. If you define it as color depth, then the 720p has RAW RGB pixels at 14bits, which when combined into a full-color pixel (like H.264), would be over 42bit color! What difficult for people to understand is we can can only see about 24bit color (I may be off in these, but general principle). So why would there be different color depth between H.264 and RAW processed into a viewable gamut? I've tried to explain it many ways on this forum, as have others, but don't seem to get anywhere. The best way to understand it is to do the following. COLOR DEPTH 1. Take a photo with the camera set to RAW+JPG of a scene where there is a fair amount of shadow detail and highlights (high dynamic range). 2. Pull the RAW into Photoshop and process until you get the best image that favors the shadows. Then do the same for the highlights. 3. Pull the JPG into Photoshop and process for both shadows and highlights. 4. Pick out the best AFTER processed photo for both details and highlights. The difference to me is night and day. I only shoot JPG or H.264 if I need to shoot a lot and can't afford the time to deal with RAW. RESOLUTION Again, a lot of people confuse image compression like JPG with video compression like H.264. They are similar, but also very different because video compression also compresses between frames. However, even when I shoot all "I", or separate JPGS for all frames, they still don't seem to have the "film" grain look I get from RAW. That is, I still don't feel I'm getting complete frames like RAW. I agree with Mercer, all the stuff I've shot with 720 RAW is more pleasing to me than any H.264 at 1080. However, 4K downsampled to 1080 gives it a run for the money because downsampling 4K removed de-bayering color artifacts (another long story). If you're shooting fashion I'd try to shoot RAW because biologically we're very sensitive to skin tones. Right now, in my photography stuff, I'm pulling my hair out and it's single RAW images! I guess I'd conclude and say 720p is great then h.264 1080p in color detail, but less in perceived sharpness.