Bali Autoview and openzwave

Need to write this down before I forget. I spent a bit of time trying to figure out how to get my new Bali Autoview-powered shades to work with my python-openzwave based automation server.

I ordered mine through the Bali website, using their very weird integration. If you order the Autoview configuration, it ships with a separate remote that is pre-paired with the motor in the blinds. It takes a little work to undo this pairing and get it working with your own zwave controller.

Here were the steps I took:

  1. “network” reset on the motor side. Hold the “button” in on the top bar of the blinds until it flashes red. This makes the motor side forget about all pairings
  2. On the server side, get your controller into inclusion mode. I usually do this using ozwsh that comes as part of the python-openzwave package. You launch the tool, do a “cd controller”, then a “send add_device True”
  3. Once the controller is in inclusion mode, hold the button on the blinds until the light flashes green. After a second it should be included and it will confirm by moving the slightly up and back.
  4. You show now see the blinds as a node in your zwave network. The version of openzwave I’m using appears to not know about this device and only shows some basic command classes (dimmer). You can control the blinds using standard on/off and Level/Dim/Bright settings. “On” means the blinds go up, “off” means the blinds go down.

This is enough to get your zwave stick to control your blinds. The problem now is that small remote that shipped with the blinds is now unpaired. You need to get that to work. I tried the procedures listed in the included manual, but nothing worked. I then figured out that you have to first include the remote into your zwave network.

Here’s how I did that:

  1. On the remote, use a paper clip to push the button in the back and hold it in until the light on the front flashes red. This means a total reset and it’s forgotten all associations and inclusions.
  2. Set your zwave controller into inclusion mode
  3. Push the back-side button in again, and hold it, until the light on the front flashes green. This gets the remote onto your zwave network. (For me it shows up as a blank node)
  4. Once the remote is on the network, then the standard pairing described in the manual seems to work. On the remote you poke the back side button once to get it flashing, then you hold the button on the blinds in until it flashes green. These two then associate.

Now I can control the blinds through my home automation server / z-wave stick and also with the included remote.


Panasonic LX100 Review

I’ve been in the market for an enthusiast compact for a while now. I had my eyes set on the RX100mk3, but then out of nowhere, Panasonic announced the LX100. 4/3 sensor? check. Manual-ish controls? check. Fast autofocus? check. Built in EVF? check. Small size? check. I couldn’t resist, so as soon as the first couple reviews came out, I picked one up for myself. Below are my impressions after about two months of ownership.

Image Quality

Overall, the image quality is great. The sensor is on par with any of the other 4/3 sensors out there. Actually, I suspect most of the modern m43 bodies are based on the same Panasonic sensor (E-M1, GX7, GH4, LX100). While the output is a little different from my EM-1, in terms of overall quality, dynamic range, etc., it goes toe-to-toe with the EM1. Not terribly surprising given the shared sensor.
One thing that immediately turned me off was the default JPEG rendering. I think I’ve been spoiled by the E-M1’s JPEG rendering over the past year or so, but the LX100’s default JPEG setting is pretty terrible. Crappy noise reduction and what looks to me like toy-ish colors. At minimum you have to turn the JPEG noise reduction setting waaaay down. For the most part, I avoid it.

People often make fun of the filters on the camera, but I personally like them. Yes, you can re-create the effects in Lightroom, but it may take you some time. I like having them right on the camera, and especially being able to preview the shot in the selected style. I like Panasonic’s “Dynamic Monochrome” more than the Olympus equivalent, but I like Olympus’s retro style more than Panasonic’s. Go figure.


I think this was the big surprise for me. For my E-M1, I only have primes. I use the super-sharp Leica 25mm/1.4, the Olympus 45mm/1.8 and on occasion the original 20mm/1.7 — all three top-notch lenses.

As I started to shoot with the LX100, I tried to manage my own expectations. This is a compact zoom, with relatively large aperture. There’s no way it’s going to compete with the best primes for the system. I had read a few of the reviews talking about how corner sharpness suffered at some focal lengths.

But this lens has surpassed all my expectations. Yes, I bet if I peep, I can see some of the corner fall off, but the center is damn sharp. In fact, the Leica 25mm prime is only sharper if you really stop it down to it’s optimum f/4. For something so small and convenient, this is an awesome result. The net of it is that I can do most of my shooting without really worrying about the lens sharpness.

The long end of the zoom gets you about 75mm equivalent. While useful at times, I find 75mm to be a bit of an odd length and not that useful. If I had the choice, I would give up the 50-75mm part of the range for a smaller lens or wider aperture.

Speaking of aperture, one disappointment is that the maximum aperture decreases pretty rapidly as you zoom in. I really like to shoot at 35mm, and at that length, max aperture is f/2.3, which means you’ve lost almost a full stop from it’s maximum 1.7. f/2.3 means that I need ISO6400 to capture my kids at 1/160 indoors, which gives you a bit of grain. At 50mm equivalent, it’s down to f/2.7. This is where the Leica 25/1.4 still comes in


Of course the whole point of this thing is the small size. If you’re coming from a bigger m43 body or even larger system, the size does not disappoint. It will fit most large coat pockets, and will take up no space in any kind of bag. That being said, it is definitely not pants-pocketable. The lens barrel just sticks out too much, even when collapsed.

When I realized this, I initially hesitated. I really wanted rx100-levels of portability with the 4/3 sensor capability and better controls. However, as I’ve taken this camera anywhere, I’ve come to realize that it’s generally so small and light that I can just use the included strap and string it over my shoulder and forget about it. The camera is so small and light that it doesn’t get in the way of most light activities. For example, after a day at the park, when I get back in the car to drive home, the E-M1 would have to come off for me to drive. With the LX100, I just adjust the strap so that I wear it diagonally and then just sit down and drive.

Physical Controls

The other big draw of the LX100 are the full set of external dials for all the major controls (shutter speed dial, aperture ring around the lens barrel, EV comp dial). In general, I love having these and they make shooting with the camera lots of fun. Interestingly, when I tried a similar scheme over a year ago with the Fuji X-E2, I didn’t like it as much. But since then, I’ve learned to use the manual mode on my E-M1 a lot more. So I think this type of control scheme really requires a level of maturity with camera operation. If you’re still learning the ropes, the dials are going to require you to up your game and really think about all the parameters.

Even then, I still sometimes get tripped up by the EV comp dial. On a more DSLR-like control scheme like the E-M1, switching from shutter to aperture priority typically resents the EV comp value. No so on a LX100 (or any other body that employs similar control). Once you dial in -2 EV, you had better remember to switch it back when you’re on to the next shot. Other times, the EV comp dial gets knocked while the camera is in my bag, and next thing I know I’ve shot a few shots at -1 EV. Over time I’ve learned to be more vigilant about the dial, but still it is a source of stress.

The other slightly odd decision was to make the focus ring by default adjust the shutter speed. This is configurable in the menus but the behavior is really confusing until you notice it. It’s very easy to move the focus right every so slightly, especially when you’re going for the manual aperture dial, and by doing so, you’ll adjust the shutter speed a little bit off of it’s dial setting. In general, for a given shutter speed dial setting, you can fine-tune it using either the manual focus ring, or the rear wheel by up to 1 stop in each direction. While this is handy in that you can set “in-between” shutter speeds like 1/160, it also means that the current setting may not be exactly the same as the dial, which means you have to look at the screen if you want to be sure.


For the most part, the autofocus is amazingly fast. I was really surprised by this. I’ve never used a compact camera that focused this fast. In most cases, it’s faster than my E-M1 with 25mm/1.4 (not the fastest lens in terms of focusing). It also retains it’s AF speed in low light better than the E-M1. Basically no complaints here.

One of my torture tests for AF is to take shots of my kids as they go back and forth on the swings in the park. LX100 can handle this no problem. Just set the shutter to 1/1000, S-AF and fire away.

I haven’t done too much with the C-AF. In my limited use, it seems OK, but not amazing. I suspect it’s probably on par with the GH4. Panasonic’s face detection works quite well, and I prefer the UI to Olympus’. However, the one thing I’ve gotten used to on the Olympus is the ability to set it up so that by default you have your selected focus point that gets overridden automatically when a face is detected. On the LX100 (as other Panasonic camera), face detection is it’s own mode. So if you want to switch between Face detection and single point focus, it takes some menu navigation to do it.

Other Minor Observations and Nit-picking

One very nice feature is the nearly silent leaf shutter. I’ve played a bit with the Fuji X100s, but this shutter is even more silent. So much so that when you first start  laying with the camera, you can take a shot w/o realizing it. I’ve seen this happen to multiple people I’ve handed the camera to. They push the shutter, they see the screen blank for a second, but because I had auto image review turned off, they’re not sure if they took anything. The shutter barely makes a sound and registers no tangible vibration. The ultimate sealth camera.

One disappointment is a lack of built-in flash. There’s a small flash in the box, which is usable in a desperate situation, but I much prefer a built-in bounce-able flash like the one found on the a6000.

It’s very hard to chimp your RAW files on this camera. While this problem exists on any cameras, it’s particularly bad on the LX100. The jpeg preview that is rendered when you take a shot in RAW is way too low-res. You zoom in two levels to 4x, and that’s the limit. Any more and you just see stretched blur. The main problem is that the resolution is not enough to be able to confirm focus. I just have pray that I got it and check on the computer later.

I found the WI-FI feature ok when it works, but have had tons of problems getting it to connect at all.. especially when I’m out and about with no wifi networks around.  Perhaps an issue on the iphone side, but it’s bad enough that I haven’t been able to use the feature much, even in cases where I really wanted it. Perhaps I’m just doing  something wrong, but I’ve never had this much trouble from any of the apps from any of the brands.

I think one of the persistent complaints about the LX series has been the lens cap. The LX100 is no different. Included in the box is a standard snap on cap with a short leash chord you can use to attach it to the body. There’s just something about it that feels super onerous when you just want to get a quick shot. There is a petaled lens hood, but availability of the black one in the US seems poor, even months after launch. I may just go with a UV filter, but haven’t decided yet.
Other reviews have noted this as well, but on startup, it takes a long time for the lens to fully extend. It’s sad because the camera feels so snappy once it’s ready to go, but turning it on feels sloooow. On a related note, if you turn it on, and then go into playback mode and start looking at some pictures, the lens retracts. To get it to extend again, you have to half-press the shutter, but then again it takes it’s time to fully extend again. Kind of annoying when you’re out and you take a few shots of something, then want to review them in playback, and next thing you know the lens is back in so when you’re ready to take some more shots there’s a delay.

I haven’t seen it mentioned anywhere, but the default luminance setting of the rear LCD seems really off. Images look as if you did a +2EV exposure adjustment in Lightroom. You can tone it down in the menus, but this confused me initially. In combination with the fact that I often didn’t realize I had knocked the EV-comp dial to a -1EV position, I underexposed a bunch of shots. I’ve learned to try to keep an eye on the live histogram as much as possible.

Another minor nitpick. When you turn on both the leveling meter and the histogram on the live view, it tends to cover a large part of the image. When I handed the camera off to a friend when it was in this state, she asked me, “where’s the actual picture?” 😦

I haven’t used 4K photo mode that much, but when I have, it’s been pretty nice. It’s certainly a very interesting way to shoot, and I definitely plan to play with it more in the future. For the scenarios I tried it in, it seemed highly dependent on the C-AF performance during video. In other words, you get 30 frames per second to choose  from, but that doesn’t mean anything if your subject is moving out of the focus plane and the AF can’t keep up. Also, the in-camera UI to browse the video and pick out frames is kind of clunky and slow. It will work in a pinch, but it’s probably better to do it back on the computer, in which case, I’m not totally sure what software I’m supposed to use.

Lens stabilization seems reasonable. It’s not anywhere close to the E-M1, but it does let you get down to the 1/20 shutter speed or so w/o much issue.

Bottom Line

In most of my day-to-day, outdoor situations, the LX100 has replaced my E-M1 as my go-to camera. It’s smaller enough than the E-M1 that I will take it out in more situations. So much so that debating getting something bigger as my “big gun” (D750, A7ii, etc.)

I still turn to E-M1 + 25mm f/1.4 for indoor situations (mostly because of max aperture). I’d prefer if this combo focused as quickly as LX100 though. But those are often situations where I don’t need as much portability, so it’s starting to make sense to differentiate further and get something slightly larger for those situations.

Early 2016 Oculus/Lightroom PC build notes

The last time I built a PC was in late 2010. I was pretty sure it was going to be my last build ever. Apple products were slowly taking over my household, and the only reason for the build was to play Starcraft 2. I went with mid-range parts and ended up with a machine which served me well for 6 years, but had lots of little annoyances.

Fast forward to late 2015, several things had changed. I had a bit more space to work with, so I no longer needed to cram everything onto a laptop. As much as I love my 13 inch Macbook Pro, I was getting annoyed at it’s performance (mainly it’s lack of a quad core processor). While Lightroom is not the best at fully utilizing multiple cores, for the most common operations, it can at least decently scale up to 4 cores. As I was starting to shoot more at my son’s sports outings, I was getting very frustrated with the sluggish workflow.

The simplest solution would be pick up a 5k iMac or a Mac Pro. Initially that was the plan, but then Oculus announces the launch of the Rift. I’m in the camp of folks that believes VR is going to be one of the biggest shifts in computing experience we’ve seen in a long time. It’s an opportunity for anyone to jump in and do foundational work in defining the computing experience of the future. I had to get involved somehow. But that also means, for now, that I needed a PC.

After weeks of debating whether I actually could tolerate all the troubles of a PC, I decided to go for it. My goal was to build something that met the requirements for the Oculus Rift (both in terms of using the devince and developing experiences for it), as well as provided a good Lightroom/photo-editing workstation. A second, but also highly important requirement, was that this machine be as silent as possible. My last build had all kinds of fan / coil whine issues which annoyed me every time I used it. This time I would do all the research to make sure I would avoid all those problems.

Below is the build I ended up with along with some build notes and observations.


  • Intel i7 5820K CPU (LGA2011-v3 socket)
  • Asus X99-A / USB 3.1 motherboard
  • Kingston DDR4 4x8GB 2133Mhz Value RAM (KVR21N15D8/8)
  • EVGA GTX970 ACX 2.0+
  • Noctua NH-U12S CPU heatsink + fan
  • Samsung 950 Pro 512GB NVMe PCIe M.2 SSD
  • EVGA Supernova P2 750W Power supply
  • Fractal Define R5 White ATX case

Upgrade Path

  • Broadwell-E based CPU
  • More RAM
  • NVidia Pascal-based GPU
  • 4K or 5K monitor
  • Hard drives


  • CPU: If all you want to do is play Oculus games, then any of the LGA2011 CPU’s are overkill. You should either go Skylake (i.e. Core i7 6700K) or even a couple older generation which still meet the Oculus minimum spec. I went with the cheap 6 core because a) it supposedly overlocks pretty well, and b) the cores will become more useful as Lightroom improves on it’s multi-core support and also when I want to start developing software for Oculus. Also X99-based platforms have higher memory limits so I can easily go to 64GB, or move to a Xeon chip if I really feel the need.
  • Memory: From everything I’ve read, overclockable RAM seems hardly worth the effort. So I just go for baseline Kingston stuff, which I’ve always had good experiences with.
  • Graphics: The GTX 970 meets the minimum spec for oculus. This was intentional. I wanted basically the cheapest card that meets the spec, for 2 reasons: Nvidia was going to intro Pascal architecture-based GPUs in the next year (for which a large perf bump is expected) and I also really wanted support for DisplayPort 1.3 (which can properly drive 5K monitors w/o special hacks).

Build Notes

  • Overall: Pretty smooth build. Things have gotten way better since my last build, and I really appreciated the case having built-in cable management and a power suppy which allowed me to remove all the cables that I was not using.
  • Case Fans: One thing I was confused about was the fact that the included case fans only have 3 pin connectors, where as all the headers on my motherboard for driving/controlling those fans had 4 pins. After a bit of research, I realized that the 4 pin headers can accept the 3 pin connectors, and are in fact keyed in a way that forces you to align the pins correctly. In the BIOS these fans get correctly detected as voltage-driven rather than PWM-driven.
  • Coil Whine: Though I had done a ton of research. I wasn’t able to completely avoid it. I noticed that I get some at very high load (running Prime95) or during lots of network activity. Thankfully the Define R5 case has enough sound dampening so that I don’t hear it.
  • The fractal design-branded fans that came with a case have a slight rattle when they are running. You can only hear it with your ear right up against it, but the Noctua fan on the CPU cooler has no such rattle. I might invest in a couple Noctua case cans at some point to eliminate the rattle.
  • I chose a white case because I was a sick of the standard gamer PC look. While the Define case is indeed white, it has plastic panels and other features that still make it a far cry from a cleanly designed Apple product.
  • The EVGA power supply produces some noise from it’s fan, unless you turn on the “ECO mode” switch. My understanding is that in this mode it only turns on the fan when it gets to a certain temp, which it doesn’t reach most of the time.

Update (Jan 20, 2016)

  • The EVGA GTX 970 card comes only with full display ports. The Dell P2415Q comes with only a mini-DP to full-DP cable. So I had to get another cable. Apparently the 2415Q can be a big temperamental depending ont he cable you pick, I went with the Accell UltraAV (B142C-007B-2) cable which purports to support the full DP 1.2 spec bandwidth.
  • There’s something a little wonky with either the display or the graphics card. Once a week or so, the display gets into a rut where it seems to lose the sync on the signal once every few seconds. This manifests as the screen going black momentarily and then coming back, with random screen glitches in between. Whenever it happens, turning off the display and turning it back on seems to fix it. Par for the course for a $400 monitor? Maybe.
  • The display has an ever so slightly green tint on the left side. But you have to really look with a test pattern to see it. It’s subtle enough that in practice I never notice. Also probably what you get for $400. I will just have to wait until there are high end SST 5k displays to be had for reasonable prices
  • I installed the ASUS performance tweaking utility. It looked scary. When I had it up, Lightroom slowed down to a crawl. Not sure if it’s related, but I uninstalled most of the ASUS stuff immediately. Lightroom isn’t slow anymore. Again, I can’t prove cause and effect, but I’m not taking any chances. Seems like most of the software duplicates stuff in the BIOS anyways.

HP Stream Mini + Ubuntu 14.04.3 + Ubiquiti Unifi Controller

I’m a recent convert to the Unifi system. So far I’ve gone in for two AC AP’s and the USG router.

The one annoyance with unifi is having to run the controller. I’d been running it on-demand on my laptop, but wanted a better solution. Separately, I had also been thinking up about setting up a low power, linux server for some other home stuff. I decided to take the plunge after realizing that a small home server would also be a great place to run the unifi controller 24/7.

Hardware: HP Stream Mini 200-010

I was looking for something cheap, low power, reasonable cpu, and could run linux. I had vaguely heard about people hacking chromeboxes (desktop form factor of chromebooks) for this purpose, but upon researching found out that it required a bit of hacking and even some bios rom editing.

Then I came across the HP. I have mixed feelings about HP – I’ve used some of their good high end laptops, and workstations, but also a lot of junk – but figured this is a standard laptop-hardware-in-a-desktop build, so it’s hard to screw up. A few blog posts and youtube videos out there also provided empirical evidence that a smooth linux install was a not totally unreasonable expectation.

For those who are unaware of this product. It’s basically a windows-based chromebox competitor. It sells for ~$180, has 2GB of RAM, a haswell-class celeron, and 32GB ssd. RAM and SSD are easily upgradeable. And because it’s not a chromebox, it has a normal sane bios. Seemed worth a try.

Fast forward to today, I got the thing in the email. It took me a bit, but I did manage to get it deployed as a small server in a matter of a few hours. Here are my notes.


Backing up windows

In case the thing didn’t work out and I wanted to send it back, I needed a way to reset back to the factory windows install. This is done through HP’s “recovery manager” software, which has a “create recovery media” option. Plug in a USB stick when prompted and just wait a while, and voila!

Installing Ubuntu server

I used the instructions here to create a bootable installer for Ubuntu server. The instructions call for Ubuntu desktop, but I just downloaded the server iso for 14.04.3 (amd64 edition) and it worked fine.

When booting with the ubuntu drive, you really have to mash the escape key to get into the bios (the time window is really narrow). Once there, muck with your boot order so that the usb stick is higher than the “windows boot manager” entry in your UEFI sources, and you should be able to get it to boot into the ubuntu installer.

From there on, follow the standard Ubuntu installer. It’s been a few years since I personally had to do this so I had a couple minor points of confusion.

  1. I’m not sure if this behavior is only present in the server version, but it doesn’t attempt to configure the wifi chip during the install, which means if you’re not wired, then some of the network auto config stuff will fail. I just plugged it into the network for the initial install, and things progressed on from there.
  2. It seems partition schemes are more complicated in the UEFI world. I just sed the “guided use entire disk” option. I probably could have done something better.

Installing Unifi controller

I originally went to Unifi’s page to download their pre-packaged .deb for their controller, but then realized that their release notes have instructions for setting up an apt repository to get the controller. Sweet! Just followed the instructions there and got everything installed with only one minor complication.

I originally skipped using the separate mongodb source repository that is labeled as “optional” in the instructions. I then couldn’t get my browser to connect to the controller (see below), and so in the process of debugging, I tried to add the separate mongodb repo, and re-install. (The custom mongodb package has some file conflicts with Ubuntu’s native mongodb-clients package, so you have to do some muckery to get the custom version to install. In my case, I removed the custom mongodb source from my sources.list.d, then apt-get update, then apt-get remove mongodb-clients, then re-add the source, then apt-get install mongodb-10gen.)

It turns out that my connection issue had nothing to do with which mongodb I used (though I did not try the ubuntu native one). Rather, it was that I was connecting to http://hostname:8443 instead of https://hostname:8443. Unifi controller helpfully does absolutely nothing when you for get the s. It just closes the connection, leaving Safari utterly confused.

Sadly it took me a good hour to figure this out, but once I did, things have been smooth sailing. To get your new controller to adopt all your existing devices, you can use the backup feature in the controller to export then import the entire configuration. (As of Unifi 4.6.6, the backup features is in the settings dialog under “maintenance”)


In summary. Cheap hardware (under $200), smooth Ubuntu install, custom apt-get repo. Don’t forget https in your URL. Seems like it’s working well so far.

Update: memory upgrade

I decided to spring for a bit of extra memory to give me some headroom beyond the included 2GB. I ended up going with this 4GB Kingston module ($23 from amazon). Popped it in and seems to work fine.

A trip to Leica Store SF

It’s not often that the world’s largest Leica store (in terms of sq footage) opens up close to you, so last month, I dragged a couple of my coworkers/fellow enthusiasts to pay a visit.

As you can see from the photo of the Noctilux above, they had the full set of very expensive Leica gear, and they let us touch it!

Of particular interest to me were the new X Typ 113, and the Leica T.

I’ll start with the T. First impressions upon picking it up were that a) it’s heavy and b) it’s not as slippery as it looks. It looks nice in person, but to me it also somehow doesn’t look like a camera, and I can’t tell if that’s a good thing or not.

It had the 23mm mounted on it, so I walked around the store and took a few shots. AF wasn’t terrible, but definitely not fast. Images, as far as I could tell on the rear LCD (which is quite large) looked fine. Zooming in revealed the high quality of the prime lens.

After playing with it for a few minutes, navigating menus and the like, I got the distinct feeling that this is a v1 product (which it is). It makes some bold choices, some things work, but other things don’t. The thing that bothered me the most was the frame-rate of the UI. To review shots, you pull down from the top, but the animation that happens when you do that reminds you of 2010 android. Likewise for scrolling around the settings menu (which are visually very simple). It feels like they just put a really underpowered CPU in there to power the UI.

I found it somewhat ironic, since really if you think about it, Leica’s flagship M series is know for it’s feel. People choose the things because of the dials and precise manual focusing and the analog experience it gives you, despite the digital innards of the latest revisions. With the T, somehow it felt like this same attention to the feel of the camera didn’t extend into the firmware. The interface, while interesting and potentially even good, felt clunky because of it’s physics. If I’m going to spend two grand on a camera that is primarily operated through it’s touch screen, I expect to have a damn good touchscreen UI, and I don’t think the T’s UI lives up to that standard.

Maybe they’ll fix it in the T2. I do like the overall size and the possibility of being able to use real Leica lenses on an APS-C sized body, so I guess I’ll just have to stay tuned into what their doing.

On to the X. After playing with the T for a while, when I picked up the X, I almost threw it out of my hand. It is significantly lighter. It almost felt like a toy.

I immediately started checking out the behavior that Steve Huff talks about in his review. At close focusing distances, this camera stops down the lens to smaller apertures. At really close distances, it only opens up to f/2.8. I don’t know why but it just feels so misleading. There’s nothing I’ve seen in any of their marketing so far that has any indication of this behavior.

I didn’t get to play with the X as much, but the UI was much more traditional, much more straightforward camera. AF didn’t seem faster or slower than the T. Not terrible, but far from impressive.

After some time with both, strangely enough, I liked the X better. the T is interesting but needs some refinement. The X is much more traditional, and while it has that weird flaw, it’s still surprisingly light and probably takes some awesome images. Overall though, not sure that it’s worth anywhere close to the 2x price premium over, say, a Fuji X100S.

iOS 8 iPad wi-fi issues

So yea, if you’re googling to figure out some problem with your apple product, and it you end up on a Apple Support Communities thread, you’re wasting your time 99.99999% of the time. Except this time.

iOS 8.0.2 seems to make the wifi on my retina iPad mini quite sad. More than 10 feet away from my router? bam! no connection.

Turns out there’s a buried setting that seems to fix this. Settings > Privacy > Location services > System Services > Wi-Fi networking. I turned that off, and I was back in business immediately.

I guess that’s what I get for upgrading early.

Olympus E-M1 vs A6000: Follow up

It’s been just about 2 months since my last post comparing these two cameras. When I wrote the last post, I hadn’t decided which camera to keep (since they greatly overlap in terms of usage scenarios) but I’ve recently decided to keep the E-M1, and sell the A6000.

It really came down to the lens selection. But not in the way I expected.

You’ll see that in my previous post, I acknowledged that the E-mount has a smaller lens selection compared to micro 4/3. That’s obvious to anyone who follows this stuff. I claimed that as long as you have your standard focal lengths covered, the extra variety of lenses available to you didn’t matter much.

So why are the lenses my prime decision factor?

I had a chance to visit my dad, who now owns a Sony A7 with both the 35mm f/2.8 and the 55mm f/1.8. After playing with these lenses (both on the A7 and the A6K) it was clear to me that Sony is going to build the high end lenses only for their FF system. It makes sense. Price sensitive customers who come in and buy the $600 bodies are not going to be jumping all over $1000 lenses. And you can tell by looking at Sony’s APSC E-mount lens lineup. Most all of them (other than the CZ 16-70 and the CZ 24mm) are in the $200-500 range.

Don’t get me wrong. Those E-mount lenses aren’t bad lenses. Most of them are decent, especially for their price points. But reading all the reviews, you can tell that the only standout lens is the CZ 24mm (which is $900). I think this was paired with the NEX-7 which was the “high end” of the NEX system at the time. But the A6K doesn’t seem like the NEX-7 successor. That role has been taken on more by the A7 series.

Compare this with micro 4/3, where premium lenses are plentiful. Four constant-wide-aperture zooms (from both Oly and Pana). Lots of top-of-the-line primes. Not that I’m going to go out and binge on a bunch of $800 lenses, but it’s good to know that as I get more into the system, I do have the choice of investing more in my favorite focal lengths.

The conclusion seemed pretty clear. By buying into the mid-range a6000 body, I was also buying into the solidly mid-range lens lineup as well.

So while the E-mount lens lineup does have all the basic focal length covered, it hasn’t done it at the various price/quality points you’ll see in a more mature system. In the m43 system, you can find multiple price points for virtually any focal length, which makes it possible to go budget for lenses you might rarely use, but splurge on the 50mm prime that you use all the time.

So I’m going to stick with m43 for the time being. While I really wish the E-M1 had the CAF performance of the a6000, I’m sure I’ll manage.

Some related observations

Shopping for lenses for the A6000 really reminded me of when I owned the Nikon D70, one of their earlier APS-C dslr bodies. Nikon was (and still is) in the same position, where they sell both FF and APS-C systems. As a APS-C user, I waited and waited for them to sell a good APS-C 35mm equivalent prime, but to this day they still haven’t made one. The only choice is to get the FF 24mm/1.4 lens which is $2200 (!), or a slower f/2.8 lens. If you look at their APS-C lens line-up, it’s almost all zooms and super-zooms.

It seems that whenever companies end up supporting FF and APS-C systems, they come to the same realization that they should only make the real high-end stuff in their FF format.

It’s really only the manufacturers that only have smaller format bodies that are really willing to invest in producing the best lenses for smaller formats. Currently they are Panasonic + Olympus for m43, and Fuji for the X-mount*

* We’ll have to see how the Leica T system turns out. I’m optimistic since they’re starting with such high price points to begin with.

Sony A6000 vs Olympus E-M1: Analysis

It’s been almost 2 months since I got my A6000. Over that time I’ve had the opportunity to take it out on a few occasions, shoot with it around the house, and play with a few lenses. Below is a collection of observations and comparisons between the two cameras.


One of the challenges for me in comparing these two is that I have a very different set of lenses for each. For the last few years, I’ve been using three m43 primes: the 20mm panasonic, the 25mm pana leica, and the 45mm olympus. All three have their uses, but they are all fixed primes.

With the A6000, I got the Zeiss 16-70 f/4, and then later found a deal on the 55-210 f/4.5-6.3. I was also able to borrow the sony 50mm f/1.8 OSS for a couple days to try it out.

Shooting with primes vs zooms is obviously going to very different experiences. But frankly speaking, I’m tired of dealing with primes. Yes, their IQ is better especially when you get them in their sweet spot (using the 25mm leica at f/2.8 leads to some really cool shots with a huge amount of pop). But when I’m out and about with family, nobody wants to wait for dad to switch lenses. The sheer amount of time I was fiddling around with lenses was making me miss opportunities for shots.

At the end of the first two months with the Zeiss 16-70, I can basically say that in terms sharpness, all of the m43 primes I have are better. But the difference is pretty incremental, and I still prefer having a versatile zoom over having to bring multiple lenses and a bag. The Zeiss 16-70 also has pretty nice color rendition and some of the Zeiss look.

The Sony 50mm prime has center sharpness at least that is on par with the m43 lenses. It arguably degrades a bit more in the corners but that affects approximately zero percent of my shots.

The one big question still in my head is how would shooting with the Zeiss zoom compare with the Olympus 12-40mm f/2.8. The pro-grade Oly lens gets rave reviews, and I suspect it probably performs better than the Zeiss. I’ve only had the chance to try one for a few minutes at a friends house, but it seemed pretty nice, if not a little large.


The A6000’s claim to fame is its fast focusing system with 179 on-imager phase detection sensors. It took me a quite bit of fiddling with the camera to learn how to use it well, but once you figure out, it’s really quite impressive.

The scenario that most impressed me was one that still seems to throw even the E-M1.. a kid swinging back and forth on a swing at relatively close distance. On all my previous mirrorless cameras I’ve tried (gh2, xe1, xe2, em1), I feel like this scenario nets me on the order of 1-2 useable shots per 10 or so that I took. So imagine my surprise when on the first try with the a6000, I get about 7-8 in focus shots per 10 on average (under good outdoor light). I was really quite impressed.

Here’s a bunch of 100% crops of shots I took in this manner (I just took screenshots of 1:1 zoom in LR, so not maybe the absolute best raw conversion). Also I took these at around 1/800, I suspect that might have been slightly slow since I was at such close distance. Some of them are not what I would call super pin sharp, but remember that this is also a 24mp sensor so the magnification factor is huge. At any normal viewing size, all of these look totally fine.

And for all these ones that I got in focus, I only had to filter out one bad shot!

There are other things I prefer about the A6000’s focusing system. For one, it’s 3×3 zone focus UI is way better. The E-M1 can also do a 3×3 matrix focusing, but the problem is when it locks, it only highlights one of the 9 squares – I’m guessing the one it used to actually obtain the lock. Some of the other squares might also be acceptably in focus but there’s no way to tell. The Sony, on the other hand, will highlight all focused squares at once. So if you’re using a 3×3 grid, 5 of the cells might light up all at the same time. Typically, when your subject is likely to not fit in just one cell, this type of UI gives you a much better sense of whether the camera picked the right area of focus.

Another area that I noticed a difference is face detection. Simply put, the E-M1’s face detection system is finicky. I can’t tell you how many times it seems to lose the face just as I’m about to push the shutter and I end up with a totally misfocused shot because the behavior is to default back to wherever your focus point was (which is usually in the center for me). I didn’t do an in-depth comparison, but I’m pretty sure the GH2’s system was more reliable. The A6000, for me, has been the best of the three. It’s remarkably good at keeping track of faces even as my kids look down or to the side. Most crucially though, it doesn’t seem to be indecisive like the E-M1.

The one major downside with the A6000’s focusing system is it’s performance in low light. Especially in S-AF mode, it really does slow down compared to when your’e in a good light situation. But it’s also not like it’s terrible. I would say it’s still better than the 20mm/1.8 panasonic on the E-M1, though admittedly that’s a low bar. Perhaps a more relatable statement is that, I’ve been shooting with the 16-70 at f4, 1/160 and iso 6400 indoors in pretty bad lighting, snapping pics of the kids doing stuff, and it’s been pretty useable. The hit rate definitely goes down but again, it’s not unusable.

For lowlight, the E-M1 definitely has the advantage, especially when I had the olympus 45mm on the body (the 25mm is not the fastest focuser by m43 standards, though it is not bad either).

The TL:DR in terms of focusing is this: In good light, I prefer the behavior of the A6000 on many different levels. In low light, the EM1 is faster, but it’s not a night and day difference. The A6000 is still usable with a little patience.

Form Factor and Physical Design

The A6000 is amazingly light! With the battery and card, the body weighs in at only 344g, which is lighter than the already hollow-feeling X-E2. The E-M1 weights in at 497g with the battery, so there’s already a huge difference there. The net of it is that the E-M1 + the 25mm prime feels about the same as the A6000 with the 16-70 zoom, even though the zoom lens is significantly bigger and heavier than the prime lens.

The A6000 has more of the traditional rangefinder shape, with the flat top, where as the E-M1 has the famous OM-D hump. Personal preference, but I like the more slender rangefinder shape. It’s more minimalist for one thing, but the other is that I find myself using the rear LCD to compose much of the time, which means that having less of the space dedicated to the EVF is not a big deal for me.

Both bodies have substantial grips. I can’t say one is definitely better than the other. Both are sufficient. I think the A6000 one is slightly smaller, but in practice it feels ok because the body is so much lighter.

In terms of overall look, the E-M1 to me looks more Pro with a hint of Gundam. The A6000 looks a little less gadgety than the average Sony product, but still hasn’t shaken that look completely. Having the black model helps in that regard though. It definitely doesn’t scream camera like, say the A7 or the Fuji designs.

The Sony’s card slot is right next to the battery, which sucks. On the plus side, it’s power switch is right by the shutter, which is good. Most of the Sony’s controls are actually on the right hand side of the body, which makes a lot of things easy to adjust with one hand.

The E-M1 has a dedicated card slot door, but the power switch is on the top left side, requiring two hands to flip the switch, which means I always forget to shut it off when walking around.

Image Quality, Metering, and Color

There are a ton of things that go into what we perceive as good image quailty. Sensor, lens, lighting, subject, etc. I’ll just talk about the sensor here. DXOMark gives the A6000 sensor an 82, and the E-M1 sensor a 73, which seems about right.

In good light, the A6000 sensor kicks ass. You get 24 million very good pixels. I didn’t realize how much I would appreciate having 24MP over 16MP – it gives you just that much more ability to crop and zoom. The one downside is that it makes my brand new 13 inch retina mbp work its ass off. Time to get a Mac pro I guess.

In low light, I think the most interesting observation is that between the E-M1 and the A6000, the per-pixel noise levels seem to be about the same. So if you look at images at, say ISO6400, and zoom them both to 1:1 the noise profiles look quite similar. But with the sony being 24mp, that means at any normal viewing size, you’re getting more pixels being down sampled into the same amount of space, so the overall noise profile looks better, and that’s born out in the dxo results as well.

In terms of metering, both are quite proficient, but each has its quirks. The E-M1 seems to prefer brighter exposures than the A6000 for the same scene. One thing that drives me bonkers with the E-M1 is that how aggressively it spot-meters a face if it one is found in the scene. I’ve had many occasions where the face was in a shadow, causing the rest of the scene to blow out. And there doesn’t seem to be any way to shut this off. A6000 does this too to some extent, but it’s not nearly as prominent.

In terms of color, the Sony and Olympus color profiles are traditionally different, but the bigger thing I noticed is in how each camera handles white balance. The E-M1 is very aggressive at trying to automatically correct the white balance. You’ll often see the color shift smoothly over a few seconds when pointing it at a new scene. This is especially true if you turn the “don’t auto WB for warm scenes” setting off. Sometimes this works quite well on the E-M1, saving me work in lightroom, but other times it’s too aggressive, making my kids faces look a little ghosty. The A6000 does not seem nearly as aggressive, but I do end up with a bunch of yellow or greenish shots from indoor lighting. Overall, I’d say the E-M1 system works slightly better.. in terms that I end up tweaking it less than shots I get from the Sony.

In terms of the built-in filters, I definitely like the Olympus ones better. Somehow the Sony ones feel just a tad cheesier, but in general I only use the B+W filters anyways. For some reason I find it very fun to be able to use the camera to see in black and white. Neither beat the Fuji film modes, though.


If you get this far, it probably sounds like I have a strong preference towards the A6000. But when you start to look at general handling, you start to see where the E-M1 justifies it’s price.

For one thing, operating the E-M1 is just plain faster. Everything from turning the camera on, reviewing images, zooming in during review, twiddling settings, all that stuff feels a bit faster on the E-M1. Perhaps it’s the fact that it only has 16MP images to deal with vs the 24MP for the Sony. Or maybe it just has a better processor inside. On the A6000, reviewing an image at 100% is painfully slow. It feels like there’s almost a whole second between when you hit the magnifier button and when you see the zoomed image. Then, once you’re zoomed, you can use one of the dials to navigate back and forth between images, but each ones takes the same delay to show up on the screen. This is particularly painful when you’re trying to go back through your stream of 11fps photos to see if you the shot you wanted in focus. I’ve experienced no such delays on the E-M1.

The buffer depth is pretty good on both cameras. Compare to cameras of even just one generation ago, buffer depth seems to be a disappearing issue. That being said, with the A6000 @ 11fps it’s pretty easy to max it out. On the E-M1, I’m not sure I’ve ever hit the buffer limit in normal use. Both cameras let you do very little while images are getting written out.

I can’t say the menu systems are significantly better on one or the other. Both are clunky to navigate, and I wish these guys would just find some real UEX designer to re-think their menu systems. On the E-M1, the item for formatting the card is pretty close to the default state of the menu, which is handy, but some other settings are frustratingly deep, and the in-camera documentation is terrible.

Both cameras have lots of customizable buttons, though the E-M1 has more. The E-M1 has a sort of “everything” menu that lets you get to all the photographic settings pretty easily. The Sony has the Fn button which lets you program shortcuts to 10 settings, covers what I need 99% of the time. The Olympus also has the 2×2 switch lever, which changes what the front and back dials do w/o going into the menu system. I thought this would be useful, but I mostly find it confusing. I often flip the switch w/o realizing, and then get really confused when my dials don’t do what they usually do.

Both cameras have twin control dials, but they feel much better on the E-M1, which has a dial in the front under the shutter, and one on the back by your right thumb – the typical DSLR layout. Sony on the other hand, has one dial next to the mode dial, and the circular jog dial servers as the second dial. In absolute terms, I prefer the DSLR layout, but I’ve found the Sony layout is not too bad once you get used to it. The difference likely goes away after using one or the other for a long time.

One often cited disadvantage of the A6000 is the lack of the touch screen. Honestly, it doesn’t bother me that much. The only thing I ever used a touch screen for on any camera that supported is for setting the focus point. But even when it does work, I tend to keep it off, otherwise I find myself moving the focus point through accidental touches. So while it’s a minor inconvenience on the A6000, I personally don’t consider it a crucial difference. One thing I should note is that it seems some reviews claim that the focus point is hard to change on the A6000, and I’m not sure what people are talking about. When your’e in “flexible point” mode, all it takes is a press of the center button inside the circular dial, and you’re in the mode to move the focus point. A few clicks in one direction or another, and another click of the central button and your’e done. It’s certainly not really any worse than the E-M1 if you’re not using the touch screen.

In terms of the LCD and EVF. I find the E-M1 LCD easier to see. It’s bigger, glossier, and feels like a clearer image. Same for the EVF.. which is also more magnified compared to the Sony. The Sony’s LCD feels a tad small, and it’s got a weird aspect ratio that doesn’t get filled up when you’re looking at thumbnails. I suppose it’s shaped for video recording.

There is one handling quirk on the Olympus that does drive me insane. If you’re in review mode, looking at the back LCD, and you move your eye up to the view finder to get a better look, the camera goes out or review mode. You then have to fidget and hit the review mode button to get back there. How annoying! None of the other camera’s I’ve used have this behavior (Sony, Panasonic, Fuji).

One I really like about the A6000 is the pop up flash. It’s small, tilt-able, and strong enough to bounce off the ceiling a medium size room. I’m not a big flash person (yet), but I find this one useful enough that I’ll often use it with the kids at home.

The last big difference I’ll mention here is around the stabilization. As all the Olympus reviews talk about, the E-M1’s built in stabilization is simply awesome. I took the camera out to shoot fireworks handheld, and actually got a bunch of pretty good shots at 1-second exposures. The 16-70 lens I got for the Sony does have OSS, and it does a reasonable job, but the E-M1 is just simply insane. It’s also apparent when you take videos with each camera. E-M1 sometimes looks like you’re on a much fancier stabilizer of some sort.


These are things that aren’t that big of a deal at the end of the day, but didn’t really fit cleanly into one of the other sections.

  • A6000 allows you to charge over USB. This turns out to be handier than I expected. Namely, it allows me to charge the camera in the car, or with my portable battery pack that I use for my phones. On the flip side, the camera doesn’t come with an external charge, which is lame.
  • Video quality is way better on the A6000. The PDAF-based continuous focusing really helps track subjects. It’s also got a lot more options for bitrate and framerates. Videos of my kids never looked this good.
  • I like the iPhone app for the Olympus a bit better. I expect both to evolve significantly over time.
  • Getting sharp shots on the E-M1 requires you to remember to use the “0-second anti shock” mode which engages the electronic first curtain shutter to kill any blur caused by the shutter. I don’t know why this isn’t enabled by default, and critically, on the E-M1 it seems you can’t use this in burst modes. A6000 has not such issues.
  • Totally personal preference, but I like 3:2 framing of the Sony over the 4:3 framing of m43. Probably because I learned on 35mm film and then used a Nikon D70 APS-C for such a long time.
  • E-M1 can do in-camera RAW development, which comes in handy from time to time.


Which one will I keep? At the time of this writing I still have both. I can honestly say though that I’m leaning towards the Sony. When I do finally decide, I’ll write some more thoughts about why I ended up where I did.

One thing I have to constantly remind myself is that the A6000 is only $600 retail compared to the E-M1’s $1400. The price differential is unbelievable considering how well these two bodies compare.

Sony A6000 Index

A ongoing collection of videos / reviews / and other info about the Sony A6000

Online stores with reviews

Review Articles

  • DPReview’s first impressions

  • DPReview’s final review

    Gold award

    The a6000 is a really solid camera and is better than its predecessor in almost every respect – anyone worried about Sony cutting corners to hit a lower price point should be reassured that it has done so intelligently. The result is a really capable camera, offering impressive image quality in both Raw and JPEG modes, plus class-leading video features. That it then does so in a small, convenient, well featured and competitively priced package is what really seals the deal for us. It’s strong in all areas and outstanding in some.

    If you plan to shoot video as well as stills, this is the camera to buy in this class. If your focus is purely stills, things are a little less clear-cut, and the decision should probably come down to lenses: Sony’s approach to its E-mount lens lineup doesn’t seem as enthusiast orientated as those of Olympus or Fujifilm (with a bit more focus on consumer zooms than accessible primes), and the power zoom included here offers lots in terms of convenience and flexibility, but falls short on ultimate image quality. As such, more than usual, the a6000’s full potential isn’t truly realized without buying additional lenses but, if you’re willing to make that commitment or can appreciate the a6000 kit as a really canny all-rounder, then you’ll end up with a camera that’s difficult to beat.

  • Ongoing report at

    Regarding AF performance compared to a Nikon D4S

    I won’t belabor this, but I know many of you are curious, so I’ll say that my success rate with the Nikon D4S was much, much higher. Its Dynamic AF area modes are much better for multi-player sports; it nails the first image nearly always; and on steady-state motion like bicycles, it almost never misses. But remember that the D4S would easily best a $650 Nikon DSLR, too. That the Sony A6000 can deliver 4-6 sharp images per second on a reasonably regular basis is very impressive for a mirrorless ILC camera costing 1/10th the price of the D4S.

  • First impressions at itsnotthecamera

  • Cameralabs review

    88% (recommended)

    The Sony A6000 is a good solid interchangeable lens camera with one major advantage over rivals: effective continuous AF which works pretty much anywhere on the frame for stills or movies. If you’re shooting stills, it also works at the top continuous speed of 11fps, and if you’re filming movies you can even fine-tune the AF speed and response time. This takes the A6000 beyond what any other mirrorless or DSLR camera offers at this mid-range price point, shooting it to the top of your shortlist if you’re after a camera that can confidently shoot or film sports or even just your kids running around.

  • Techradar Review

    4.5 stars (out of 5)

    “Sony has come within a whisker of creating the perfect CSC in the shape of the A6000, with just a few niggles stopping it being very good indeed.”

  • Gizmodo review

  • Review at
  • DXOMark sensor review
  • Review by Douglas Fung

    This isn’t a perfect camera, and as is normal for a Sony camera, it does feel at times that it leans more towards “gadgety consumer electronic product” than “serious enthusiast photography.” THis also isn’t helped by Sony’s rather sparse documentation. However, quirks and shortcomings aside, there is a lot to like about the A6000. You have to make compromises between photographic quality and portability, but the A6000 manages those compromises competently.

  • Initial review at Weeder

    If I had to rate the Sony A6000 right now, I’d probably give it 4/5 stars. I’ll admit, it has everything going for it and really deserves 5/5 stars for most types of photographers who are just getting into photography and/or may have never even known what a traditional SLR camera was like to operate. However I still get that “I know what I’m missing” feeling a little too much for me to personally give it more than 4/5 stars. Too many things could be dramatically improved with just a little firmware update, or increased (software-based) customizability.

    This therefore gives me a lot of hope for the future. Considering that the Sony A6000 costs less than the Nikon D5300, and 1/3 of the cost of a Sony a7, I’d say that for anyone interested, this camera is a winner.

  • David Pogue looks at A6000 and LeicaT

    In my month of using it, there were several occasions when I had only a fraction of a second to capture an image. There were no second chances. And the A6000, with its blistering focus speed and beautifully designed controls, nailed it each time.

  • Comparison with A7R at BeforeTheCoffee (Has some nice samples of AF-C performance)

    For those looking for a compact interchangeable lens camera with good quality performance, the A6000 is a bargain. It performs superbly with auto focus in low light and continuous focus of a moving object. The A6000 sensor does not perform as well as the A7R’s sensor but if you’re NOT pushing the envelop of high ISO and dynamic range, images taken with the A6000 will be very pleasing. As I stated in the beginning, the A6000 is speed priority and the A7® is sensor priority. For me personally I own both for the best combination; great image quality with the A7R and Fast Hybrid auto focusing with the A6000.

  • Review at Steve Huff

    Bottom Line? If I offered an “Editor’s Choice” the A6000 would take that title easily. Highly Recommended for those looking for a quality camera with DSLR quality and speed without the size. Also good to note is that during the review period I never had a mis focused shot or any issues at all with the camera. When I snapped I knew what would come out of the camera would be fantastic.

  • Sony Alpha Rumors’ review roundup

Forum Threads


CIPA camera industry stats for 2013

If you obsessively follow camera sites like I do, you will have seen the recent big headlines about the 40% drop in camera unit shipments from camera companies in 2013 compared to 2012. 40% is certainly a huge number, but I wanted to try and unpack that a bit.

The first thing I looked at was how the overall drop breaks down into the various categories tracked by CIPA. Here’s what I found:

Built-in lens camera shipment: -44% vs 2012
SLR shipments: -15% vs 2012
Mirrorless shipments: -16% vs 2012

Clearly, the bulk of the drop is coming from built-in-lens cameras (compacts, point and shoot, etc.). When you see that the “built-in lens” segment makes up 73% of all shipments, you can see how the overall drop can seem so large.

But is this really surprising? We know the iPhone and other smartphones are taking over the low end. We’ve heard all the companies say they’re dropping their low-end lines.

Interestingly, the data seems to also indicate the cost that went into production, so you can get some sense of profits in each category. I don’t know if this is actually how to interpret the data (I don’t see any docs on the rest of the CIPA site about this), but I compared the Yen amounts in the “shipment” column with that in the “production” column and divided by the number of units shipped to calculate a per-unit-shipped profit.

Assuming the methodology is legit, it looks like per-unit-shipped profit went up for the “built-in” category by almost 23% (this is after normalizing for the approx. 18% drop in the value of the YEN between 2013 and 2012) I obviously can’t conclude from just looking at this data, but it does seem like the companies did not over-produce these lower-end units, and actually probably lowered production on the lowest end models and are improving the profit margin on the units they do sell.

In the SLR category, year-over-year profit / shipment seems about flat (2% drop) after currency normalization.

Most interestingly, in the mirrorless category, the per-unit profit is up 50%  vs 2012. Perhaps this is an indication of the shift to higher-end mirrorless? We saw lots of FUJI models, high-end Panasonics, and even the Sony FF cameras appear last year, though I’m not sure those have enough volume to explain the uptick. 

If you look at absolute profit numbers in each segment, it falls out that:

“built in lens” category: -27% vs last year
SLR category: -16% vs last year
mirrorless category: +25% vs last year

And if you look at profit share by category:

“built-in” lens: 30% vs 35% in 2012
SLR: 56% vs 56% in 2012 (basically no change)
mirrorless: 14% vs 9% in 2012 (big gain)

Again, I don’t know if I’m interpreting the profit data correctly. But if I am, the data seems to support an overall decrease in volume, mostly coming the built-in category, and a shift to more profitable devices in both the “built in” an mirrorless categories.