Wednesday, November 27, 2013

Camera video performance comparisons

It occurred to me that I've used a number of cameras* for video and that it could be interesting to do a small comparison and some analysis of their video performance. So, without further ado, here we go.

I did my first real attempt at making a video in the beginning of 2011 with the Canon EOS 60D. The Canon EOS 60D was a very mixed affair for me. The ergonomics were great and the image quality often excellent, but sometimes the issues with back and front focus affected the image quality too much. Looking back at some pictures, the noise level is a bit higher than what you would expect from a APS-C camera. The video quality wasn't really as good as I had hoped: I felt the actual resolution was rather far from the nominal 1920x1080 of Full HD, something that was very obvious when pulling out still images from the video stream.

The move from an EOS system to a m43 system was a real winner for still photography (fhe fun factor, lack of back/front focus issues and light weight), but the video quality went down.

I liked the m43 format and again became more interested in video, and therefore aquired a Panasonic GH2, which was considered an excellent camera for video. With the hacked firmware the image quality was indeed great.

At the end of 2012 I became caught up in the GoPro Hero 3 Back Edition hype, like so many others, and snagged one of the first ones arriving to stores in Finland. I didn't use it much, since I found that I didn't like the POV angle much and the strong fisheye distortions really irritated me. Additionally, I found it rather clumsy to use and the form factor only suitable for the chest harness. I eventually sold it. Below is a short test video with it.

I still liked idea of having the possibility to do videos in bad weather, which the Panasonic GH2 definitely cannot handle, and bought a waterproof Panasonic FT4 cheap, after having seen the videos made by Joe. It turned out that the still image quality was rather lacking in less than perfect light, but the video quality was actually quite ok.

The Panasonic GH2, while being an excellent camera for video, never had the real fun factor for still photography for me, so when Rajala Camera cut the price on the Olympus EM-5, I came to buy one. It is an excellent and fun camera for still photography, but I didn't expect much from the video capabilities, due to the lacking PAL frame rates and the poor codec and low bitrates. I was pleasantly surprised after I made a little test video. It will certainly be usable for at least some glide camera stuff, because of its excellent image stabilizer.

The above videos do naturally not tell everything. Vimeo and YouTube do degrade the videos, and especially the GH2 is better with a higher bitrate (Vimeo gives maybe 10-15 Mbps, while the hacked GH2 records in 50-70 Mbps). Still, this might be of some interest to someone.

* Yes, I know, this is a lot of cameras, but I've sold the ones I don't use and generally have bought them on sale.

Sunday, November 24, 2013

Olympus EM-5

Regular readers might remember that I switched out my Canon EOS system, a 60D and some lenses, to a much lighter m43 system in August 2011. I never looked back and I have since complemented the system a little. My latest purchase is the Olympus EM-5, which was interesting when I bought the Panasonic GH2 in the beginning of 2012. At that point it was not yet available, and I would probably have bought the GH2 anyway, because of its superior video capabilities and lower price. The Panasonic GH2 was a good camera, especially on the video side, but for some reason it never had the same fun factor as the Olympus PEN cameras for me. Meanwhile, the Olympus EM-5 was continuously praised, and a month ago I noticed that Rajala Camera sold them for a rather low price. The temptation was great, but I did give it a few days of thought and meanwhile sold a lens that almost never got used. Finally I came up with enough good reasons to buy it and can give some first impressions after having used it for a month.

The camera is small and light, but still big enough for my hands. It has a good EVF, a big point for me, since I think a viewfinder is much superior to the display on the back of the camera. With the kit zoom, which qualitywise is ok, the setup is weatherproof enough to be used in the rain. According to the tests, the image quality is excellent, slightly worse than the best APS-C cameras from Sony and Nikon, but better than the Canon APS-C cameras. It has an excellent built-in image stabilizer, meaning that all lenses will be stabilized. Hence, on paper it almost seems like the optimal outdoor camera for someone like, possibly with the exception of the video capabilities.

With a small lens, the system is quite small.

The kit zoom is rather big, probably because it is sealed. The aperture is ok at the wide end of 12 mm (24 mm equivalent), but poor at the tele end of 50 mm. I would rather have a weatherproofed prime lens with a large aperture with a suitable wide angle, but seemingly no manufacturer believes that it is possible to sell such a combination.

In good light, the differences between cameras are not that big. The dynamic range does have some relevance, though, and here the EM-5 is actually quite good. Configurable highlight/shadow warnings makes it easy to keep the exposure right and there is some room to fix things in Lightroom. E.g. this picture would have had pretty much either a silhuette or a totally overexposed sky with a lesser camera.

The real test of the camera was the Vigu Survival course. You can see more pictures under the actual blog post, but below is an example of what is possible with it. I often find myself in situations where a tripod is just not possible to use, since there is no time for photography. You then have to work fast, and the EM-5 image stabilizer combined with a good lens certainly enables pictures that otherwise would be impossible. The below picture was taken handheld at ISO 3200, f1.8, 20 mm and 1/10 s. (ISO 6400 is still usable and the image stabilizer works with longer exposure times, though the success rate declines.)

A 100 percent crop of the above image, unsharpened. (Remember, this is pixel peeping.)

There was of course no doubt about the image quality in the camera, after all the positive reviews. The smaller sensor size is a compromise, though, and while it does not affect the image quality much, it does mean that it is more difficult to get a shallow DOF (depth-of-field). With the proper lenses it is still possible, though.

The video capabilities are lacking in two ways: There are no PAL frame rates and the codec is known for its low bitrate and inability to deal with complex scenes. My expectations were therefore not very high, so I decided to do a little test.

In this example, the video quality is quite ok, though the Vimeo compression made the video slightly jerky. I'll have to look into that later.

Finally, the maybe most important thing is that the Olympus EM-5 does have a high fun factor. I think I will be very satisfied with this camera.

Sunday, November 17, 2013

Vajosuo Beer Ride Edition V

November. Often just a full month of misery and darkness in Finland, with a few nicer moments, like the traditional beer ride to the Vajosuo mire. This was the fifth time we did it. The concept is simple: Pack your bike with gear to make it through the night and add some food and beer. Ride for a few hours to the shelter / fireplace in the evening, spend the night there and ride for a few hours back to the starting point the next morning through a different route. The difficulties are generally the technically very demanding trail, darkness, rain and once it was even rather cold with -20C temperatures and snow.

This time we had a record number of participants. Another shift in focus, at least for those who had participated before, was that the food had become more important than the beer. I had only two cans of beer, but more food instead. Everything was packed on my new bike and I rode without a backpack.

I arrived with Matti H and greenman a little late to the starting point a little after six in the evening. We encouraged the others to leave before us, since there was no question that we wouldn't catch up with them, but nothing happened, so we started a little late.

The duck boards were now and then covered with ice and with extremely slippery. Almost impossible to walk on and difficult to ride on.

Toni shows his style.

We soon noticed that the tail was nowhere visible and stopped to wait.

Full moon over the lake Savojärvi.

It didn't take long for the first participant to opt out. I followed him back to the parking place with Matti H and greenman. After that we could ride the boardwalks in our own pace and made good speed. After having gotten a little further this time, Pihvi came towards us with another casualty, whose rear derailleur was now useless after having fallen off the duckboards. I escorted him back with Pihvi, which resulted in us being back at our starting point again after one and a half hour of riding. We decided to take another slightly faster route with Pihvi and rode the trails and duckboards pretty efficiently.

Pihvi having passed one of the more exciting sections. I was very close to getting a full bath in the wet bog myself.

A short bit before Töykkälä I flatted my rear tire, something you don't want to do with a fatbike. It took almost twenty minutes to fix it, most of the time being pumping the air into the new tube. We decided to skip the last trail section and ride to the shelter on roads in order not to be too late. We rode for around three and a half hours for a total of around 22 km.

The fire was already burning hot when we arrived. Due to the gourmet nature of most of the food, there was no space for my food yet, so I heated some water for hot chocolate and warmed a sandwhich. My old Trangia coffee pan has now found new life for use with a fire, after having been unused for more than ten years.

JJ and Matti H had a vegetable stew in addition to pasta and big pieces of meat.

All sorts of food in foil, as well as some traditional sausages. I had one foil of vegetables and another one with salmon and blue cheese, which tasted great.

Activities, including beer and glorious plans for the future, around the fire.

Around two in the night it was time for the traditional and mandatory walk to the bird watching tower to look at the mire.

I only had the Olympus XZ1 pocket camera with me. With a real camera the moonlight would have made some real nice pictures possible.

My bivy spot...

and Pihvi's.

Morning again. The night had seen some hard wind, but I slept quite well and got up a little before eight.

Breakfast and coffee.

A few different bikes.

Toni's new 9:Zero:7 with 100 mm rims and Bud and Lou tires generated some interest.

Toni's bike loaded up. It looks really nice, but I do think it could have a little larger main triangle.

And mine.

Vajosuo in the morning.

We got started an hour late and two of us left the shelter to ride back on roads. The rest had some very difficult duckboards ahead of us. I rode most of them without problems, as did the Krampus man.

An alternative technique. The new duckboards still wait to be installed.

The sun doesn't rise high anymore.

JJ masters the duckboards.

A road section.

Filling the water at the Käärmeslähde spring. The water wasn't very clear and tasted mud this time. I didn't get sick, though, but I only drank a little of if.

We rode the Kangenmiekka trails and took a short sandwhich break at the Laaskorpi shelter on our way back.

The next section along the west border of the Vajosuo mire was the worst, but it went quite well. After that we ran out of time because of the late start in the morning and had to ride to the cars on road, all but JJ, who rode the final trails alone. We arrived at the cars a little over one in the afternoon after three hours and around 20 km of riding.

Thanks to everyone for participating. It was fun again.