A Bonanza Of Details About The Panasonic GH3

The GH3 stands a chance at being the video/photo hybrid camera to own. I have high hopes for inevitable tests of the sensor and its video capabilities. Still, there are some points that are available now that are cause for concern.

The Sensor

Thanks to a recent interview over at Imaging Resource, we have a large number of great data points. First, and I think most importantly, we now know that physically, the GH3 does not have a multi-aspect sensor. The designers wanted to make a large number of improvements to the sensor, and the resources required to do this essentially negated the larger sensor. They seemed to indicate that they still wanted the multi-aspect sensor, so we may yet see it again in the GH4. I, for one, don't really care about this feature. It's not an issue with larger lenses attached with adapters, but native m4/3 lenses vignette more severely on the 1.86x sensor. For photography, not much of an issue. For video, though, it can be annoying.


In that same question, they also dropped that the sensor is going to be an in-house Panasonic design. There had been a lot of questions about whether the sensor was going to be a Sony, as is found in the Olympus E-M5, but the engineers said that they were limited by the sensors available from the "sensor group." It is always possible that the sensor group was working with extant Sony sensors, which would help to explain the lack of the 1.86x sensor, but I think that this confirms that the sensor is a genuine Panasonic creation.

The engineers provide an excellent data point that provides some insight into the sensor's dynamic range. They say that it is an "extended dynamic range," and if the numbers they provided are even slightly accurate, that's going to be very true, and a huge upgrade to the GH2. The engineers say that the theoretical limit for the pixel saturation should be 45,000. The actual number will be lower, but even if it is half the limit, it will be good. We can use SensorGen for actual data, and the E-M5 has an actual limit of 25,000. The comparatively pathetic GH2 has only 11,000. What's going to be most important is an increase in the quantum efficiency, in which Panasonic is weak since the GH1, which was both objectively and subjectively the best sensor that Panny has made.

Unfortunately, they also revealed some distinctly non-marketing speak details when they said that the "sensor isn't specially designed, so it just scans normally." They talked about this in reference to the read rate being 1/10th of a second, which does mean that the sensor takes 1/10th of a second from the first pixel to the last pixel being read. Apparently, with standard sensors, the read rate is based on the pixel count, and that means that the GH3's rolling shutter is going to be no better than the GH2. Disappointing, to say the least.

Features

One place where the GH3 is not at all disappointing is in the feature department. The GH3 is so loaded with features as to be a little overwhelming. Thankfully, unlike many other companies, Panasonic's features are real features. You know. The kind that customers actually want and don't think that they want they read the sign when in the store.

The biggest one is the extensive connectivity features. Bluetooth and WiFi provide all manner of connections, including full video controls via a tablet or smartphone.



The biggest detail though is that usage of WiFi allows the potential for any and all kinds of connections, including XLR connections. The engineers in the aforementioned interview said that this isn't currently on the slate, but if it's possible, you can rest assured other companies and hackers will make it happen.

For me, the most important feature was always focus peaking. I was excited to hear that this would be coming to the GH3, but now, we hear otherwise. The GH3 will not have focus peaking, and that is a massive disappointment. The engineers were incredibly evasive on the question of this and zebra-striping to indicate overexposure.

They said that they have limited resources and must focus on many products. If this is indeed true, then Panasonic is still just as stupid as we've become accustomed to. The GH3 is the only Panasonic product about which anyone gives a rat's ass. It is the very definition of a halo product, and being affordable makes it the ideal halo product. That Panasonic isn't putting everything that they can behind it would be a display of magnificent stupidity.

Which is what makes me think that this reason is a lie. I suspect that Panasonic is holding off on things in an attempt to figure out how to squeeze more money from the situation. Their stupid pricing of their zoom lenses  illustrates this executive drive in action.

This is not the time for Panasonic to half-ass it. They need to go all-out. And if they don't, Sony will eventually pass them.

Trust Me, You Want The Blackmagic Cinema Camera

One River Media has produced a stunning example of the Blackmagic Cinema Camera's abilities. It's going up against the Canon 5D Mark III, which in comparison to most of today's digital cameras comes out looking pretty poorly, but even in the ways the the Canon is usually very strong, the BMCC comes out on top. In low-light performance, the BMCC does very well, and its dynamic range leaves the large-chipped Canon in the freaking dust. Truly, the stunning clarity of the sensor makes me want to write a movie and film it just so I have an excuse to use this camera.

Ladies and gentlemen, this is what innovation looks like.


Comparing the Cinema Camera & 5D Mk III from OneRiver Media on Vimeo.

Canon EOS 1DX Is Identical To 1DC. WTF?

As if we needed more evidence to support the assertion that Canon sucks, here comes this doozy of a factoid: the upcoming, thirteen thousand dollar, EOS 1D C cinema SLR camera is fucking identical to the already released EOS 1DX. The only difference is the firmware.

Why would Canon do something like this? Because they are an arrogant monstrosity that likes to rip people off. They are, quite literally, bolting $6,000 in profit on to every 1D C that they sell and telling anyone who is interested to bend over and take it. How in the bloody blue hell do they expect to compete in an increasingly competitive market with that sort of internal business philosophy?

I thought Panasonic was bad when they priced their new lenses so high. I thought Hasselblad was bad when they wrapped an NEX 7 in leather and tried to sell it for $5,000. This blows them both out of the water. The only other camera that even comes close is the Sigma SD1, and while its overprice was worse, it was at least a unique camera!

It's amazing how often technology companies try to pull this shit. There are two examples in the world of computing that are prominent in my mind. Intel, back in the day, released a series of 486 processors that were, in all ways, identical. They were all a bunch of different prices, though. Why was this? Because Intel intentionally crippled the cheaper ones. An industrious user could, somewhat easily, change the settings to "unlock" the actual potential of their new chip. Later on, the chips were actually different, but for a good year or so, they were absolutely identical.

Similarly, there has been an ongoing debate in the world of high-end graphics cards for some time. A graphics card that is good for running games usually costs between $100 and $400, with a few mega cards as much as $1,000. Workstation cards, on the other hand, cost upwards of $6,000. What's the difference between the two types? No one knows. Most people suspect that there is almost no difference. Understandably, many pros are pretty pissed off when they think that their $3,000 is available to gamers for $300.

As such, it's not surprising to find out that almost every single graphics pro whom I know is running a standard desktop card. People don't like to feel ripped off.

The world of pro-level graphics cards supports this sort of rip-off, I think, because competition is borderline non-existent. There are two companies: Nvidia and AMD, and that's pretty much it. Similarly, back during the 486 processor era, you had one processor company: Intel. Companies in positions such as that can afford to be arrogant. It usually breeds contempt, though, and today we have companies running in droves to mobile processors to get away from Intel, and almost every computer power-user that I know hates either AMD or Nvidia.

That is almost beside the point. The camera world is very competitive, and it is getting immensely more so. The world of pro-level imaging may have been receptive to this sort of cash grab in the past, but that is no longer true. When Fuji's APS-C cameras provide full-frame level performance from a small sensor, and Sony produced the NEX-7, capable of world-class images for around $1,000, and Panasonic's GH3 is more pro in may ways than Canon's $13,000 EOS 1D C(rap), when all of this is considered, Canon is essentially insulting us when they do this. They are saying "we think that you are stupid enough to buy whatever we give you."

Unfortunately for you, Canon, I'm not.

Hooooo-wee! Nikon D600 Tops The Charts!

Well, not quite. It's in third place over at DxOMark. Behind the D800 and D800E. So in the world of performance per dollar, it's #1. As I mentioned in an earlier post, this only further relegates the recently-announced Canon EOS 6D to "why the fuck do you even exist?" status. I'm still not the biggest fan of the D600 because they didn't drop the price to below $2,000, but this test result makes it much easier to like.

Yes, yes, it's only $100 more than that target, but I see the fact that they didn't go for the sub-$2,000 price as further evidence for the severe conservatism in place at Nikon. They were very aggressive with the feature set, which is nice, but that price sticks in my craw. And since this area of the market just got real competitive, real quick, it makes it all the more important to take bold steps in design and pricing.

Moreover, the price of the D800 makes this camera's price harder to see as, not reasonable, but sensible. When I can get the greatest camera ever (for the time being) for less than 50% more, why wouldn't I want and save up? It's smaller and a bit lighter, but in this size category, the lenses are going to negate any noticeable difference.

It must be recognized that Nikon is stepping up their game. I remember the shock and awe when Nikon priced the D3X into the Oort Cloud. The D800 and D600 indicate that they are never likely to do something that arrogant again, especially seeing as they lost nearly half of their market to Canon during the time that the D3X was king. We have a new king, and this time, he isn't handicapped with a ridiculous price.

Now let's see what Sony has up its sleeves.

First GH3 ISO Tests

And there you have it. The first tests of the GH3's ISO performance have hit the interpipes. They are good. I won't say spectacular because, first, they are only JPEGs and not RAW, and two, because even with that considered, they are only what we would expect.



Based on my "seat of the pants" analysis, the camera is a stop better than the GH2 at high-ISO. It also lags the E-M5 slightly. This is an admittedly limited analysis, but from what I'm seeing, I don't think there will be much to criticize. That doesn't mean there won't be anything to criticize.

At the minimum ISO of 200, there is some color banding in the shadow areas. And considering that this camera will be going up against the lower-priced Fuji XE-1, which is two full stops better, this is a point of serious consideration for anyone who's planning on buying into a system. I'd say that the ISO performance is good enough where Sony's APS-C cameras hold no specific advantage and if not for Sony's dedication to innovation, I'd say that m4/3 was the obvious choice.

All that said, the GH3 is a unique camera. No other company is even close to offering what this camera has. Based on these photos, I'd say that if you are only interested in still photography, then Sony or Fuji is the better bet. They simply have a fundamental advantage in the form of their sensors. But if you like shooting both stills and video, the GH3 is looking more and more like the camera for you.

More Info About The GH3 and Micro 4/3 In General

A few interviews with reps for both Oly and Panny have hit the webs and they provide some interesting, if not quite juicy, details.

First, Olympus is retarded. They denied any work on m4/3 versions of their Zuiko glass. Instead, they made some wan gesture toward a converter. I don't think that I can put to words how stupid this is. With Sony, the converter made perfect sense. Their Alpha lenses are a vibrant system, made only more vibrant with the exciting release of their SLT pro-level cameras. 4/3 does not have this advantage/limitation. 4/3 is dead. Drop the dead donkey.

But back to the good stuff, Panasonic. An interview over at EOSHD has given us some very cool bits of data. First, the sensor's readout time on the new GH3 is 100ms. I assume that means that it takes 1/10th of a second from the point the first pixel is read to the last pixel. This seems very slow to me. I don't know what other digital cinema cameras have, but considering what I've seen in rolling shutter tests from the likes of the Canon EOS 5D Mark III and the Sony NEX 7, this read-out time is among the best.

Interestingly, the reason for the low bitrate recording on the GH2 was to keep internal temperatures within Japanese regulations. I suspect that there is some truth to this, but I also suspect that the bigger reason was that Panasonic didn't want to step on the AF100's toes, which was stupid. There is a brief mention in the interview of the AF100, where the rep claims that it was once popular, but now isn't. No. It was never popular. There's a big difference.

That point about temperature, though, is interesting. Temperature in chips has been a point of focus in the computer world for years. People all around the photography community were suddenly introduced to the problem when cameras like the EOD 5D Mark II started overheating like crazy during extended shoots. As pointed out in the interview, one of the reasons is that to get the sensor to read out at an acceptable speed requires powah, and that means heat. This gives me hope that future hacks of the GH3, when combined with  suitable heat sinks, could effectively be overclocked to increase the sensor readout. It's too bad that we are at least a generation away from a global shutter, though, specifically because of these issues.

The only video point that is very disappointing is the 4-2-0 HDMI out. Yes, it's uncompressed, but 4-2-2 is important for pixel-peeping. It's noticeable. Not terribly so, mind you, but enough to be a niggle. It's gonna' make one bad-ass webcam, though.

Panasonic X 12-35mm Loses Some Dollar Signs

The Panasonic X 12-35mm, which has really only been available for a very short time, has already seen its first price drop. This doesn't come as much of a surprise, seeing as I've spent the last few days posting repeatedly about Panasonic and how the X lenses will likely be a failure.

So where are we? The 12-35mm now costs $1,099. To me, this speaks of a company that is gingerly stepping in unknown territory where they should be taking bold steps: drop it to $999. For one thing, I think that it is optically only worth about one thousand dollars. That's a personal judgment so you can accept it or not. My other point stands perfectly well, though: Panasonic needs market penetration. This will not happen with half-steps. This will not happen with greed. Attack the market. Attack the competition.

Do not be like Olympus. I have no idea how they bumbled into the 75mm and E-M5, considering that everything else they've done, and everything else that they have announced, has been incompetent. For example, in a recent interview, an Olympus rep denied any plans for m4/3 versions of their top-quality Zuiko lenses.

All that said, this price drop is an excellent idea. I don't know whether this is in response to poor market response or merely a strategic choice even before the market has a chance to react. Either way, this is a good showing on Panny's part. I think that they should drop this to $999, but $1,099 will be much more likely to experience significant uptake among pro-thusiasts.