Gallery of my favourite panoramas

While I have over 30 galleries of panoramic photos up on the web, a while ago I decided to generate some pages of favourites as an introduction to the photography. I’m way behind on putting up galleries from recent trips to Israel, Jordan, Russia and various other places, but in the meantime you can enjoy these three galleries:

My Best Panoramas — favourites from around the world

Burning Man Sampler — different sorts of shots from each year of Burning Man

Giant Black Rock City Shots — Each year I shoot a super-large shot of the whole of Black Rock City. This shows this shot for each year.

As always, I recommend you put your browser in full-screen mode (F11 in Firefox) to get the full width when clicking on the panos.

How to properly use widescreen mode for TV viewing

As some people come to my article about TVs in widescreen and stretch mode looking for advice, here is some basic advice on how to get it right.

Generally problems only occur when you have an ordinary analog connection from your video source (DVD player, Cable box, satellite box) to your TV. Analog connections are things like composite video (one wire with RCA plug, plus 2 RCA plug wires for audio), S-video (the video cable is a round plug with 4 pins in it) or component video (3 RCA plug wires, plus the audio.) When you have an analog connection, the source box can’t tell whether your TV is widescreen (16 by 9) or regular (4 by 3). If you have a digital TV connected to an antenna, it will get it right automatically, but if you have a digital TV converter box, you need to set up that converter box.

That usually means your source box will have a menu item, somewhere in the setup section, where you can tell it what sort of TV you are connected to. Most of them will default to an old 4:3 TV size, so you need to go into the menus on the source box, and look for where you can set the size or “aspect ratio” of the TV. This is not done on the TV’s setup, but on the setup of the DVD player or set top box.

If your source box is connected to the TV via a digital connection — which means DVI, HDMI or even VGA — it should be able to figure out the size of your TV on its own. Normally you should not need to play with things. However, if it’s doing it wrong, you may need to go into it and play around.

The other control — the one you want to avoid using — is the one on your TV remote. It will be marked with a name like “size” or “p. size” or “aspect ratio” or “zoom” or even “mode.” It lets you change what your TV does with the signal. You should very rarely have to use this setting if you have a widescreen TV and you set up the source boxes to know you have a widescreen TV.

The one exception will be if you have a widescreen TV, and you watch a program on a 4:3 channel (which means most non-HD channels) which was widescreen. Then you get what’s called a letterbox inside the pillarbox (the bars on left and right.) The result is your image is a smaller rectangle in the middle of the screen. If you are far away from your TV, you can blow this up by setting your TV to a “zoom” mode, expanding the box. Some TVs come with two zoom settings. I tend to prefer the partial zoom that does not fill the screen, as the full zoom tends to “overscan” and lose a little bit of the image.

Some TVs have an “auto” mode. In the auto mode, they try to detect this small box, and automatically zoom. If yours works that’s great, but it may not work, and it may get confused as commercials change the aspect ratio.

Normally it would be highly unusual to have to use the 16:9 or stretch mode, because again, you should have told your source box that you have a widescreen TV. This mode normally would only be used with a DVD player which is unable to be told that the TV is widescreen. It is the stretch mode that you often found used on widescreen TVs of people who hate the bars on the sides so much that they would rather see everything distorted and fat.

Hitler tries a DMCA takedown

New Update, April 2010: Yes, even this parody video has been taken down though the YouTube Content-ID takedown system — just as my version of Hitler says he is going to do at the end. I filed a dispute, and it seems that now you can watch it again on YouTube, at least until Constantin responds as well as on Vimeo. I have a new post about the takedown with more details. In addition, YouTube issued an official statement to which I responded.

Unless you’ve been under a rock, you have probably seen a parody clip that puts new subtitles on a scene of Hitler ranting and raving from the 2004 German movie Downfall (Der Untergang). Some of these videos have gathered millions of views, with Hitler complaining about how he’s been banned from X-box live, or nobody wants to go to Burning Man, or his new camera sucks. The phenomenon even rated a New York Times article.

It eventually spawned meta-parodies, where Hitler would rant about how many Hitler videos were out on the internet, or how they sucked. I’ve seen at least 4 of these. Remarkably, one of them, called Hitler is a Meme was pulled from YouTube by the studio, presumably using a DMCA takedown. A few others have also been pulled, though many remain intact. (More on that later.)

Of course, I had to do my own. I hope, even if you’ve seen a score of these, that this one will still give you some laughs. If you are familiar with the issues of DRM, DMCA takedowns, and copyright wars, I can assure you based on the reviews of others that you will enjoy this quite a bit. Of course, as it criticises YouTube as well as the studio, I have put it on YouTube. But somehow I don’t think they would be willing to try a takedown — not on so obvious a fair use as this one, not on the chairman of the most noted legal foundation in the field. But it’s fun to dare them.

(Shortly I may also provide the video in some higher quality locations. I do recommend you click on the “HQ” button if you have bandwidth.)  read more »

Making of the Video, Legally

On ultralight vehicles vs. large mass transit vehicles

One of the questions raised by the numbers which show that U.S. transit does not compete well on energy-efficiency was how transit can fare so poorly. Our intuition, as well as what we are taught, makes us feel that a shared vehicle must be more efficient than a private vehicle. And indeed a well-shared vehicle certainly is better than a solo driver in one of todays oversized cars and light trucks.

But this is a consequence of many factors, and surprisingly, shared transportation is not an inherent winner. Let’s consider why.

We have tended to build our transit on large, heavy vehicles. This is necessary to have large capacities at rush hour, and to use fewer drivers. But a transit system must serve the public at all times if it is to be effectively. If you ride the transit, you need to know you can get back, and at other than rush hour, without a hugely long wait. The right answer would be to use big vehicles at rush hour and small ones in the off-peak hours, but no transit agency is willing to pay for multiple sets of vehicles. The right answer is to use half-size vehicles twice as often, but again, no agency wants to pay for this or to double the number of drivers. It’s not a cost-effective use of capital or the operating budget, they judge.


The urban vehicle of the future, as I predict it, is a small, one-person vehicle which resembles a modern electric tricycle with fiberglass shell. It will be fancier than that, with nicer seat, better suspension and other amenities, but chances are it only has to weigh very little. Quite possibly it will weigh less than the passenger — 100 to 200lbs.

Transit vehicles weigh a lot. A city bus comes in around 30,000 lbs. At its average load of 9 passengers, that’s over 3,000lbs of bus per passenger. Even full-up with 60 people (standing room) it’s 500lbs per passenger — better than a modern car with its average of 1.5 people, but still much worse than the ultralight.  read more »

Can airports do paging as well as a restaurant?

I have a lot of peeves about airports, like almost everybody. One of them is the constant flow of public address announcements. They make it hard to read, work or concentrate for many people. Certainly it’s hard to sleep. It’s often even hard to have a phone call with the announcements in the background.

One solution to this is the premium airline lounges. These are announcement-free, but you must watch the screens regularly to track any changes. And of course they cost a lot of money, and may be far from your gate.

Some airlines have also improved things by putting up screens at the gates that list the status of standby passengers and people waiting for upgrades. This also saves them a lot of questions at the gate, which is good.

But it’s not enough. Yet, even in a cheap restaurant, they often have a solution. They give you a special pager programmed to summon you when your table or food is ready. It vibrates (never beeps) and they are designed to stack on top of one another for recharging.

Airports could do a lot better. Yes, they could hand you an electronic pager instead of/in addition to a boarding pass. This could be used to signal you anywhere in the airport. It could have an active RFID to allow you to walk though an automatic gate onto the plane with no need for even a gate agent, depositing the pager as you board.

Each pager could also know where it is in the airport. Thus a signal could go out about the start of boarding, and if your pager is not at the gate, it could tell the airline where you are. If you’re in the security line, it might tell you to show the pager to somebody who can get you through faster (though of course if you make this a regular thing that has other downsides.)  read more »

Electronic panorama head with rotation sensor

In my quest for the idea panorama head, I have recently written up some design notes and reviews. I found that the automatic head I tried, the beta version of the Gigapan turned out to be too slow for my tastes. I can shoot by hand much more quickly.

Manual pano heads either come with a smooth turning rotator with markers, or with a detent system that offers click-stops at intervals, like 15, 20 or 30 degrees. Having click-stops is great in theory — easy to turn, much less chance of error, more exact positioning. But it turns out to have its problems.

First, unless you shoot with just one lens, no one interval is perfect. I used to shoot all my large panos with a 10 degree interval which most detent systems didn’t even want to support. Your best compromise is to pick a series of focal lengths that are multiples. So if you shoot with say a 50mm and near-25mm lens, you can use a 15 degree interval, and just go 2-clicks for 30 degrees and so on. (It’s not quite this simple, you need more overlap at the wider focal lengths.)

Changing the click stops is a pain on some rotators — it involves taking apart the rotator, which is too much no matter how easy they make that. The new Nodal Ninja rotators and some others use a fat rotator with a series of pins. This is good, but the rotator alone is $200.

Click stops have another downside. You want them to be firm, but when they are, the “click” sets up vibrations in the assembly, which has a long lever arm, especially if there is a telephoto lens. Depending on the assembly it can take a few seconds for those vibrations to die down.

So here’s a proposal that might be a winner: electronic click stops. The rotator ring would have fine sensor marks on it, which would be read by a standard index photosensor. This would be hooked up to an inexpensive microcontroller. The microcontroller in turn would have a small piezo speaker and/or a couple of LEDs. The speaker would issue a beep when the camera was in the right place, and also issue a sub-tone which changes as you get close to the right spot — a “warmer/colder” signal to let you find it quickly. LEDs could blink faster and faster as you get warmer, and go solid when on the right spot. They would also warn you if you drifted too far from the spot before shooting.

Now this alone would be quite useful, and of course, fully general as it could handle any interval desired. Two more things are needed — a way to set the interval, and optionally a way to ease the taking of the photos.

To set the interval, you might first reset the device by giving it a quick spin of 360 degrees. It would give a distinctive beep when ready. Then you would look through the viewfinder and move the desired interval. Your interval would be set. If doing a multi-row you would have 2 sensors for angle, and you would do this twice. You could have a button for this, but I am interested in avoiding buttons. Now you would be ready to shoot. It would give a special signal after you had shot 360 degrees or the width of the first row in a multi-row. Other modes could be set with other large motions of the rotator, such as moving it back and forth 2 times quickly, or other highly atypical rotations.

(If you want buttons, an interesting way to do this is to have an IR sensor and to accept controls from other remotes, such as a universal TV remote set to a Sony TV, or some other tiny remote control which is readily available. Then you can have all the buttons and modes you want.)

We might need to have one button (for on/off) and since off could be a long press-and-hold, the button could also be used for interval setting and panorama starting.

The next issue is automatic shooting or shot detection. The sensor, since it will be finely tuned, will be able to tell when you’ve stopped at the proper stop. When all movement ceases, it could take the shot without you pressing the shutter using a bunch of methods. It might also be useful to have you manually control the shutter, but via a button on the panohead rather than the camera’s own shutter or cable release. First of all, this would let the head know you had taken the shot, so it could warn you about any shot that was missing. It could also know if you bumped the head or moved it during any shot — when doing long exposures there is a risk of doing this, especially if you are too eager for the next shot. Secondly, you should always be using a cable release anyway, so building one into the pano head makes some sense. However, this need not be included in the simplest form of the product.

One very cheap way of having the pano head fire the shutter is infrared. Many cameras, though sadly not all, will let you control the shutter with infrared. Digital SLRs stopped doing this for a while, but now Canon at least has reversed things and supports infrared remote on the 5D Mark II. I think we can expect to see more of this in future. Another way is with a custom cable into the camera’s cable release port. The non-standard connectors, such as the Canon N3, can now be bought but this does mean having various connector adapters available, and plugging them in.

A third way is via USB. This is cheap and the connector is standard, but not all cameras will fire via USB. Fortunately more and more microcontroller chipsets are getting USB built in. The libgphoto2 open source library will control a lot of cameras. Of course, if you have a fancy controller, you can do much more with USB, such as figure out the field of view of the camera from EXIF but that’s beyond the scope of a simple system like this.

The fourth way is a shutter servo, again beyond the scope of a small system like this. In addition, all these methods beg more UI, and that means more buttons and even eventually a screen if an LED and speaker can’t tell you all you need. However, in this case what’s called for is a button which you can use to fire the shutter, and which you can press and hold before starting a pano to ask for auto firing.

The parts cost of all this is quite small, especially in any bulk. Cheaper than a machined detent system, in fact. In smaller volumes, a pre-assembled microcontroller board could be used, such as the Arduino or its clones. The only custom part might be the optical rotary encoder disk, but a number of vendors make these in various sizes.

I’ve talked about this system being cheap but in fact it has another big advantage, which is it can be small. It’s also not out of the question that it could be retrofitted onto existing pano heads, as just about everybody is already carrying a ballhead or pan/tilt head. For retrofit, one would glue an index mark tape around the outside of your existing head near where it turns, and mount the sensor and other equipment on the other part. The result is a panohead that weighs nothing because you are already carrying it.

Update: I am working on even more sophisticated plans than this which could generate a panohead which is the strongest, smallest, fastest, most versatile and lightest all at the same time — and among the less expensive too. But I would probably want some partners if I were to manufacture it.

Allow a refund on prepaid cell service for tourists

Years ago I asked that they let me buy a SIM card in the airport arrivals area and now they often do. I also started a forum here on the best company to buy a prepaid SIM from in each country which has a fair bit of traffic.

And indeed, I have been doing that, because often roaming rates remain obscene. I dropped my Canadian SIM when Sprint offered a plan with 20 cent/minute roaming in Canada that I can turn on for $3/month — this was comparable to the prepaid price I was getting, and prepaid had the “overhang balance” problem I will discuss below. But I’ve gotten or been loaned local SIMs in several countries to good use — especially when both I and my travel companion have one so we can use our cell phones as radios.

But a few problems exist with getting a local SIM. First, you have to get one. The cheapest place to do this is usually the local cell phone shops that can be found in most urban shopping areas. If you plan ahead, you can get one mailed to your hotel, though the companies that do this which aim at tourists always overcharge — perhaps enough to wipe out your savings if your call volume is modest or your stay short. The ideal SIM is near-free, and can be found where you enter the country.

Next, you must fill the account. Almost everywhere, you must use prepaid cards bought for cash from the shops, as they will not let you fill, or refill, with an out-of-country credit card, for supposed security reasons. This is annoying because you don’t want to have a large balance remaining (overhang) on your prepaid account when you leave the country, unless you will be back before it expires. (Sometimes you can use it up in other countries with obscene roaming rates, but often not even then.) But you also don’t want to have to risk running out of minutes in the middle of a call.

The answer: Let me put a fat balance on my prepaid account, and let me refund all or most of it when I am done — ideally back to my credit card when I leave.

The cell company loses that wasted balance, sure, but instead, I am prone to use the phone more if I have a large enough balance and a good enough rate that I don’t have to worry. I will use it like a local. This would be a good competitive edge that would make the difference if I were buying a SIM. You could offer this only to people from out of the country but I see no reason not to offer it to local users too.

Yes, processing the refund has a small cost. If you insist, don’t refund the last few bucks to cover your costs. Or alternately, let me do “minutes transfer” to other prepaid users. Then I could meet somebody (or go into a shop) and transfer the minutes and get cash for them.

Of course, it would also be nice if you would let me just buy a monthly plan deal for just one month, with no contract. Cell companies seem loathe to do this, but T-Mobile in the USA has just started doing exactly that with their flex-pay. In that system you pre-pay for one month of any monthly plan, and if you think you will use more than the minutes on that plan, you can put money into an overage account. But you can’t get it back, so that’s one modification to add. But frankly I would probably never go over the monthly plan I bought in a typical trip.

The remaining big headache is data. Getting a prepaid plan with data at a decent price (or any data at all) can be hard. Those from the USA are used to unlimited data, which they resist selling in many countries. Those from the USA who bring their phones overseas and forget to turn off roaming data often find nightmare bills of many thousands of dollars. The world has to figure this out. Still, those who are used to fancy network PDA phones often find themselves literally lost without their Google Maps Mobile or their e-Mail. We need a way to roam data selectively, letting some apps use limited data budgets but preventing others if we can’t get a decent data price plan. E-mail apps can go into low-data move (never download attachments or long messages automatically, just imap headers) and less frequent checking. If one is careful, one can get something decent at the $2/megabyte (or $10/megabyte) crazy prices for mobile data roaming.

Oh yeah, and think about doing 2-SIM offers to tourists, who often arrive in pairs. Especially if they include cheap mobile-to-mobile calling in the pair.

Virgin America Airways and on-demand ordering

Yesterday I took my first flight on Virgin America airways, on the IAD-SFO run. Virgin offered a tremendous price (about $130 one way) but it’s worth examining how they have made use of technology on their planes. Mostly I usually end up on United, which is by far the largest carrier at SFO. Because of this, I fly enough on it to earn status, and that it turn provides a seat in their Economy Plus section which has more legroom, priority boarding and in theory, an empty middle if there are empty middles. This is 90% of the value of the status — the other main value, ability to upgrade, is hard to actually make use of because business class is usually full. The extra legroom is surprisingly pleasant, even for a widebody individual like myself who would much prefer extra width if I had a choice.

Other than Economy Plus (and some very nice business class on some of the long-haul planes,) United is falling behind other airlines. It would be hard to recommend an ordinary coach seat. The one big amenity that more and more other airlines are providing is power in coach, in particular 115v AC power which is more flexible than the older 15vdc “Empower” system United uses in business class. The main downside of the 115v connections is they tend to be mounted under the seats, making them hard to get to. Air Canada has put them in the personal video panels. Virgin placed them under the seats but high and forward enough to be reached (if you knew what you were looking for) but also so close as to make wall-warts bump against your legs. Virgin also offered USB jacks down under the seats, also hard to get to. Even if you don’t want to put 115v up higher, USB charging jacks are better placed in the video console/seatback I think.

American Airlines has a mix of DC and AC power, but still makes it available in coach. Continental has put EmPower on some planes in the front half of coach, but some newer planes have AC power all the way through coach.  read more »

Two sources on energy, and using heat

I was reminded yesterday, after posting more on the cost-effectiveness of energy sources, to point out an interesting new book on the economics of energy. The book is Sustainable Energy With the Hot Air by David MacKay, a physics professor from Cambridge University. What’s important about the book is that he pays hard attention to the numbers, and demonstrates that certain types of alternative energy are likely to never make sense, while others are more promising.

I only have a few faults to pick with the book, and he’s not unaware of them. He decides to express energy in the odd unit of “kilowatt-hours per day” as he feels this will make numbers more manageable to the reader. Of course with time in the numerator and denominator, it’s a bit strange to the scientist in me. (It’s the same as about 42 watts.) In a world where we often see people say “kilowatt” when they mean “kilowatt-hour” I suppose one deserves credit for using a correct, if strange unit.

My real quibble is over his decision to measure energy usage at the tank, so that an electric car’s energy usage is measured in the battery, while a gasoline car is measured in the fuel tank. Today we burn fuel to make electricity, and so electric cars actually consume 3 times the energy they put in the batteries. That’s a big factor. MacKay argues that since future energy sources (such as solar) might generate electricity without burning fuel, that this is a fair way to look at it. This is indeed possible but I think it is necessary to look at it both ways — how efficient the vehicles are today (and will be if we still generate electricity from heat) and how they might be in the future. Generating electricity from heat does complicate the math of energy in ways that people can’t agree on, so I understand his temptation.

Yesterday I was also pointed out to a solar power site called SolarBuzz. This is a pro-solar-panel site, and is rare in that it seems to do its math right. I haven’t looked at all the numbers, and I am surprised wthat with the numbers they show that they are such boosters. Their charts of payback times all focus on power costs from 20 to 50 cents/kwh. Those costs are found in Europe, and in the tiers of California, but the U.S. national average is closer to 10 cents, where there is no payback. They also use 5% for their interest rate, a low rate that is only found in strange economic times such as these — but justifiable in a chart today.  read more »

Pure heating is highly wasteful

Is being evangelical about solar the right course?

The earlier post on whether solar gives the best bang per buck in greening our electricity ran into some opposition, as I expected. Let me consider some of the objections and issues.

As a recap, I put forward that if we are going to use our money and time to attain greener electricity, what matters is how many MWH we take off the “dirty” grid (particularly coal plant output.) I measured various ways to do that, both green generation and conservation (which do the exact same thing in terms of grid offset) and worked out their cost, the MWH they take off the grid and thus the cost per MWH. Solar PV fares poorly. Converting incandescent bulbs to fluorescent in your own home or even other people’s homes fares best.

A big part of the blame lies on the fact that crystalline silicon is an expensive way to make solar cells. It is, however, quite common since many PV plants started with technology from semiconductor fabrication.

Evangelical green

One frequent objection is that purchasing expensive solar panels today encourages the market for solar panels, and in particular better solar panels. Indeed, panel makers are generally selling all they can make. Many hope that this demand will encourage financing for the companies who will deliver panels at prices that make sense and compete with other green energy.

I call this being “evangelical green.” Leading by example, and through encouraging markets. While I understand the logic, I am not sure I accept the argument.  read more »