This is what I hope to be the start of a new series of logs.. one that I have done and stopped years ago. Friends would remember what they get when they finger me. :-) I hope to update this on a monthly basis, at the least -- I don't want to be too ambitious. Feel free to email me at [email protected] to give me your comments.
Lately, I have been very complacent when shooting photos. With my compact camera, I rarely even look through the viewfinder any more. Have I mastered the 35mm focal length?
With my SLR, I do look, but I seemed not to care about my exposure. I seldom spot meters and determine the exposure. I just guess-timate, and use the exposure. No counter-check.
What have my photography become? Do I simply care-not, or is it that I depend on my gut feeling, or is it that I have internalised all these?
I have reached a point where I have to make a decision. No turning back after this. To do it, or not. Thankfully, this is not the first time I am faced with this situation, so I am not as anxious as before. I felt wiser and much more worldy experienced than before. First choice rarely work out. Finally, I learnt failure, how to cope with it, and overcome it. Once I do, I don't fear failure any more. A quote read, "We have nothing to fear but fear itself." Now I understand what it actually felt like.
From failure comes experience. And there is no need to worry too much about failing. There is always another day. Still, not having fear doesn't mean I am not anxious. I am! It is a very exciting time.
I consider it a myth that C++ is less efficient than C. Having worked with C++ for three solid years, and making it work in an embedded system with 4 MB memory, I know that C++ is just as efficient as C. C++ has a guiding principle that you pay for what you use. We will follow this as our guide.
We look at the features that are heavy and require run-time support first.
Exceptions. Exceptions are harder to use than they look. And they require too much run-time support in the form of hidden code generation.
STL. I suppose some programmers think that a C++ program written without using STL is not C++. What I feel is that these programmers should brush up their algorithms. I have no practical experience with the STL, so I am biased against it. What I fear is that you may need to link in much of the STL even if you use a small portion of it. This is not STL specific, but a common problem for framework-based libraries.
Templates. How about user-defined templates? Templates are okay as long as you can control it. A simplistic view is to think of templates as macro subsitution. C++ compilers support explicit template instantiation. Use it to control code bloat.
Polymorphism. You often need to use dynamic object allocation to use polymorphism. This can introduce some overhead and inefficiency. Dynamic binding is slower, but function pointers are often used anyway.
Operator overloading. C++ allows you to overload operators. This makes the code look more concise, but it may lead to worse code generation unless you are very careful.
Objects. Is encapsulation free? Almost. However, there are many hidden traps for you to fall into. Constructors and destructors are implicitly called, so watch their weight. They are chained as well, so watch the inheritence tree. Attributes are objects too? Oh, that's another layer of fat. The default behaviour is pass-by-copy, so you may pass things around inefficiently without realising it.
Programming is not entirely about coding. There is no better way to see how important a software architecture is, than to see what its absence will do to the program. All programs have an architecture, whether by accident or design.
(To be continued.)
In this era of Giga-everything, worrying about efficiency seems to be.. an inefficient use of time. However, for those of us who are still stuck in the Mega era, efficiency is something to worry about. Take a look at a typical desktop application. How much resource does it need? How much does it want?
While working on an application where resources are unconstrained is liberating, I like the challenge of programming for a resource limited device. These days, it means either a handheld device or an embedded system. And this brings me to my main topic.
Embedded systems are not what they used to be. Sure, the old 5 MHz 8-bit CPU and 64k memory embedded system are still with us. But at the higher end, they resemble a PC more than an embedded system. Take for example, a Pentium 233 MHz, 32 MB memory and 64 MB CompactFlash. This may be 5 years ago desktop technology, but it is pretty impressive for an embedded system.
The constraints of the embedded system determines the OS and language you can use. For devices with 16 kB of memory and less, writing in the CPU's native assembly language is the best bet. Compilers always introduce translation inefficiencies.
And due to the simplicity of the device, there need not be any OS at all. Why is an OS not needed? An OS has a multitasking scheduler, provides services and hardware abstraction. In an extremely small system, only the first is needed.
From 16 kB to 64 kB, you can start considering using a High Level Language (HLL). The most obvious choice is C, due to its transparency in translation -- there is an almost one-to-one translatioon. Can you use C++? I have no idea, because I haven't tried. I believe you can, but you have to restrict yourself to a very restricted subset, where it is more like C with Classes than C++. OS remains the same. Something simple will do.
From 64 kB to 1 MB, you have an almost free hand. However, from this point, you will probably be putting a lot of nice-to-have features and using third-party libraries, and code start to bloat pretty fast.
This time ten years ago, I begun my undergrad studies. I looked back the past six years -- since I got to know computers -- and asked myself, have I accomplished anything all these time? What have I done that is worthy of mention? Is there anything that has withstood the test of time? Looking back at the lost years, I have to say no, sadly.
Today, the answer is the same. What will my answer be, in ten years time. The same? It better not be!
The young can afford to be dreamy, reckless and carefree. After a while, you got to pick and choose your destiny. The rest is hard work and luck. You only got a few chances to get it to work.
My destiny is programming. I like computers and programming. Many people think of nothing about studying for a degree and then never apply the knowledge in their job. The paper chase. The job requires a degree, so I get a degree. You get the feeling the years are useless, wasted; spent just to get the degree.
Not me. University courses can be dry, but this is due to the poor educational system. Education is tough, but it can be fun. Nothing motivates like enthusiasm, and the desire.. to learn. Computer related skills is something I picked up over the years, refine, accumulate and never really obsolete. All fifteen years of them. Knowledge begats knowledge. As your base becomes stronger, it becomes easier to learn new things.
But programming what?
It is my dream to translate binary code between different ISA (Instruction Set Architecture). No emulators. Real binary translation. The output code runs natively, at full speed. No Virtual Machine. No Hardware Abstraction Layer.
Will I be able to do it? Well, I am trying. As it is now, I am slowing assembling the necessary skill sets. It is just lucky that I have always been interested in CPUs, OS, compiler translation and languages.
Web photo critiques don't work for me. Who is doing the judging? What is his skill level and credentials? God knows. This is why I have not submitted any photos for critique. I have viewed some, but the low level of disclosure makes it hard for me to judge if the work is worth the effort. I mean, is the photo a snapshot, pre-planned, or a lucky shot? What is the photographer's thought and intention, and did it come out as expected? If not, why not? Where did he get the inspiration from? Did he gain any experience out of it? Would he have done anything different, now that he got this shot? How would he improve it, if he had the chance?
My style is documentary style, to shoot everyday life as it unfolds, as a passive observer. As most of everyday life is mundane, it is hard to get one single outstanding shot that stands on its own. Also, I only like to shoot people I know (or beautiful girls or famous personalities), so they only have meaning within this context. Show the photos to strangers, and they are unable to connect to the event or person.
As far as I can, I only judge the technical aspects of a photo. However, digital images are especially hard to comment, in this aspect. Digital processing is an essential step. But, everyone does it differently. How to judge fairly? Has the image been cropped and enlarged? Has the image been digitally manipulated? It can be hard to tell. Worse, the screen's low dpi make it impossible to see fine details.
The ability to think ahead is an essential skill to good photos. Think, how do you nap a good photo? Were you in the right place at the right time? Were you there by accident? If so, how many times do you get lucky? Reduce your dependency on luck, and work on anticipation.
Many people depended on zoom lens. So did I. But my experience with the cheap 50/1.7 lens was so good that I learnt how to work with it. With single focal length lenses, you often need to change position or switch lenses. This is where you need to lookahead. Plot a course of movement, determine strategic spots, and batch up your shots for each lens to minimise lens changing.
Anticipation works for film too. Do you find the 36 shots per roll too restrictive? Even before you leave home, you must budget enough film already. On site, you must plan your shots well. Always lookahead to see how many shots you need. If you don't have enough, shoot quickly and change to a new roll. It helps if you keep track of "second-rate" photos along the way, so you can shoot them to fill the gap later.
A lot of people think that shooting negatives and then developing them into prints is expensive. It is, but it is not as expensive as you think.
Once you got your camera and lens, it is the running cost that worries you. Let's look at every aspect.
All AF cameras need a battery. For film SLRs, most last at least 10 rolls. Metering and film advance use up little battery. If you use the camera just for these two functions, a single set of battery can easily last 35 rolls. If you use AF, and the lens has a heavy AF focus group, it can consume a lot of battery. With AF, the battery life drops to 20+ rolls. But the worse is yet to come. The built-in flash saps the battery dry quickly. Besides, there is an annoying delay while the flash recharges before you can shoot again.
If you buy in boxes of 10, one 2CR5 costs S$5, and one CR123A costs S$2.80. Clearly, camera battery is not a problem.
How about a program flash? A set of 4 AA batteries is cheap. It will last you 6 to 10 rolls, where almost every shot needs a flash (wedding). This is assuming no flash modifiers and that the flash is used for fill-in. Always carry a spare set.
Now, film. Do you shoot less than you could have because you thought each shot costs you S$0.40? Never shoot two or more identical shots with different settings? Never rewind before using every single frame? Never "wasted" the last few shots on experiments? You should.
Let's look at the numbers first. Negative film costs S$3 to S$4. The Fuji Superia 800 film goes for S$6.50. If you can use ISO 400 films, use it. You save S$3 per roll, and you get better quality as well. Depending how much you shoot, you may be able to justify a faster lens just based on this.
Processing varies from S$12 to S$16. I now use the Fuji Frontier, at S$0.35 per print, S$0.40 for reprints. A full 36-frame roll costs S$16.10, including of the S$3 developing charge. Here, I believe that you get what you pay, hence I am willing to spend more. Processing is one area where amateurs lose out to professionals. Due to their large volume and constant business, they can get a better rate. Want to venture to part-time photography? Negotiate a better rate with your favourite lab first.
With slides, you pay more upfront. S$7 for a run-of-the-mill Fuji Sensia II 100, to S$9.50 for the sharp and ultra-saturated Fuji Velvia 50 to S$12 for the fast Fuji Provia 400. Processing is cheap, at S$4 to S$6.50, unmounted. Mounting costs some S$3 more.
Now, is this expensive? Basically, one roll of film will cost you a total of S$20. Shoot 5 rolls in a month, and you only spend S$100. Is this too much to spend on your hobby?
Now, film is cheap, but not so cheap that you can shoot without thinking. I'm sure you are horrified when pros mention they shoot 50+ rolls per day. Should you emulate them? Well, pros are pros. They need to do their job, which is to turn in a good shot. For them, taking 5 additional shots is insurance. For an amateur, more than anything, this speaks of your skills, or the lack of.
For action, I would shoot two or three shots at most. There is always another shot; different, but equally good. For static composition, how many "best" guesses do you have? A 5 stop bracket? Better learn to meter properly instead. In short, make every shot count. Do not be afraid to try out effects. Whether it works or not, you have learnt something.
A chain is only as strong as its weakest link. In a long chain, you have to ensure all the links are sufficiently strong. In the photography context, this is the old slides vs negatives debate.
Pros and serious amateurs favour slide film for many reasons. Among others, slide films are saturated in color, sharp and WYSIWYG. You can view the slides directly, either on a lightbox or a projected.
Negative films, however, need to be printed. An enlarger lens is used, making it easier to lose additional quality. Also, the colors can be adjusted until there is no resemblence to the actual scene. This is why people say if you are serious about photography, you use slide film. I disagree.
Good labs will do a good job. They are consistent. You will be able to tell the difference between different films and lenses. With digital-based printing, like the Fuji Frontier, you get sharper prints, more color saturation and slightly more lattitude.
Do not just use a lab just because it is cheap. The lab is part of the imaging chain, and any compromise here compromises the image quality. Like all things, you get what you pay for. A good lens is better than a bad lens. If you can't see it on print, try changing your lab first. Perhaps the lab's enlarger lens masked the quality.
How do you find out if a lab is worth your time? Simple, by elimination. Get people's recommendation first. Hopefully you can see their prints, and based on their camera equipment and skills, you can judge whether you want to give the lab a try. Select a few representative negatives for printing. Do it for a few labs. It is then easy to do a side-by-side comparison.
Lastly, the handling is important. Make sure the negatives come back clean, unscratched, untouched and hopefully uncurled.
Which is the best manual camera, the Nikon FM2, the FE2, the FM3A, or none of them? Recently, I had to choose a Nikon camera -- any Nikon camera, but it must be a Nikon camera. Since I only want to use one specific manual focus lens, I am not above choosing a manual camera. As you know, Nikon's AF and MF cameras use the same F-mount.
I want a cheap body, since this is only a supplement to my main system. However, I want a full featured camera. Nikkon still produce the FM10, FE10 and FM3A. The first two are relatively affordable, but they fall off the mark. FM3A is very good, a result of combining the best of FM2 and FE2. However, it is very expensive, at S$960. Therefore, it is the secondhand market for me.
How about AF camera bodies? The F55 and F65 are, well, they don't fit my needs. The F80 is pretty good, both specs and handling wise. F100 is an overkill and is simply too expensive. Discontinued models like the F70 -- are you joking?, and the F90 -- no thankks.
So only the FM2, FE2 and F80 are left. What about the F series? F3 and F4 are quite suitable candidates, but they are expensive. Besides, they are much bigger.
I ruled out the F80 first. I only need manual focus -- that lens I want is a manual focus lens. Besides, it is not compatible with MF lenses. I forsee building an arsenal of MF lenses, to save cost.
So it is the FM2 against the FE2. They are very close. Both have 0.86x viewfinder magnification, 1/250 flash sync speed, 1/4000 shutter speed, user-changable focusing screen, usable without battery, timer-based MLU and DOF-preview. These are very impressive specs, even today.
After combing several websites and scrutinising reviews, after all the dust is settled, there is one clear winner: the FE2. There is almost no contest at all. The FE2 has a needle metering system, aperture-priority, TTL flash and has shutter speeds up to 8 seconds. On the downside, the needle metering system is impossible to view in the dark, there are only two usable mechnical speeds (1/250 and bulb). FE2 wins the first round.
Next, FE2 vs the FM3A. Haven't I ruled out the FM3A due to cost? Yes, but I wanted to see how much better the FM3A is. If it is much better, I might get it after all.
The FM3A is better. It has a brighter focusing screen (K3 vs K2). It has fully mechnical shutter speeds. It allows -1 flash compensation. Its AEL button is better placed. Finally, it has a film window. On the downside, the shutter speed dial has up to 1 second only, and the viewfinder magnification is 0.8x. FE2 is badly scratched in the second round, but there is hope.
Since the FM2 and FE2 shares the same accessories, and the FM3A shares the same accessories as FM2, it doesn't take a genius to see that the FE2 can use the FM3A's accessories too! Therefore, the K3 focusing screen and the FM3A filmback are usable on the FE2.
Therefore, second round, I rule in favour of the FE2.
After getting the FE2, while I did overpay for it quite a bit -- being very excited, I immediately set out on my task of upgrading it. Getting the E3 focusing screen is not a problem (E3 is the grid-line focusing screen). It is a standard accessory for the FM3A. Costed S$31.50. However, it came with no instructions whether any exposure compensation is necessary. No one seemed to know as well. Testing it out for myself, by changing between the K2 and the E3 focusing screens, indicated that a +1/3 exposure compensation is needed.
The FM3A filmback proved to be a little more difficult. No one carried it. Nor did they know if it is for sale. Finally, one knowledgable salesman said this is not an accessory, but a replacement item, so I want it, I have to get it from the Nikon service center directly. Good, a lead! So I made my way down to the service center. No luck, the lady at the counter was skeptical, but asked me to check with their repair section. I called, to find another skeptical voice on the other end. But I was able to convince her using the FM3A-MF16-FE2 link. She agreed to reserve one filmback for me. So I went back one day to try it. I had to re-convince the counter guy, and I got my filmback. The film pressume plate slipped on the filmback effortlessly, and the filmback fit the camera perfectly. S$62. Mission accomplished.
Now I am in business.
Do you find it difficult to get good results using flash? First, do you already have an external flash unit, a fast (f/2.8) lens, and use medium speed film (ISO 400)? If so, you should be able to achieve good results already. If not, you should upgrade to this level first.
If you still get bad results, do a reality check first. Are you asking for the impossible, doing the insurmountable or just plain clueless? What do you expect? Do you know what to expect?
When it comes to using flash, I only have one technique. Use a wide angle lens. Set the camera to f/3.5 1/30. Turn the flash head to bounce off the ceiling. This is it. Perfect every time. Try it. If it works, try to reason why it works. If it doesn't, something is wrong!
For longer lenses, I use a different setting and bounce angle, but the premise is to be able to bounce flash. This automatically improves your flash shots significantly without needing you to do anything special. I rarely shoot direct. I always try to bounce flash, even if it looks impossible. To me, as long as there is a ceiling, I will try to bounce. In fact, if I really cannot bounce flash, I usually shoot without flash.
Other than bounce flash, it is good to use a wide aperture and a low shutter speed. This allows the ambient light to show. It helps to use a faster film too. I have such a good experience with the Fuji Superia 400 that I have not used ISO 800 films for my night shots.
Flash photography can get very bulky very quickly. You will be tempted to get flash modifiers first. They are the most convenient way to modify your flash. But most of them are next to useless other than sucking up battery. Others are fragile that falls apart at the first touch. Before you believe in them, do a direct comparison.
If you are shooting a controlled environment, you can be more fanciful. Use off-the-camera flash, additional flashes, studio flashes, filters and so on. However, I am not interested this currently. I am interested in the art of making the flash disappear, meaning, using the flash but making it look like part of the existing lighting. Bounced flash is the easiest way to accomplish it. Off-the-camera flash is more complicated, but it can work very well too.
Suppose the previous article didn't apply to you. After all, you are not using a superzoom. Now, let's say you have been using a standard zoom lens for some time and want to upgrade. Everyone recommends a fixed f/2.8 lens for better optics. You still need a zoom. Let's see, a constant f/2.8 zoom lens.. no matter how you look at it, it is going to leave a hole in your wallet. Not to mention big and heavy and impossible not to scream "I am taking your photo". Therefore, before you buy it, it is worth answering yourself some questions.
1. Is f/2.8 usable? Most people think that a lens has to be stopped down about two f-stops for optimal performance. True, but not all lenses are like that. Good lenses already give good results wide open. Not so good lenses are soft wide open. So, the question is, does the lens give acceptable results at f/2.8? If not, the f/2.8 is wasted.
2. Another thing related to edge sharpness is light falloff. Is it noticeable? When does it start to disappear?
3. What about weight and size? Constant f/2.8 zoom lenses are big and heavy.
4. Zoom lenses have more distortion. It is quite usual to find the zoom with barrel distortion on the short end and pin-cushioning on the long end.
5. Contrast. Glass reduce contrast. Multi-coating helps, but ultimately, there is no escape from it.
6. Flare. More glass means higher chance of flare.
Is there a better way? Of course! Ever considered a single focal length lens? What, a lens that cannot zoom? Surely this must be as backward as lenses that cannot auto-focus!
But, single focal length lenses are unmatched in several ways. They are smaller and lighter. And they are always faster. Got a f/1.4 zoom lens? And weighs less than 300 grams? The 50/1.4 lens fits the bill. And the 50mm lens is always the cheapest lens in the non-zoom lens arsenal.
Some people say they do not need wide apertures. They prefer to stop down for DOF. Sure. What focal length are you talking about? It is possible to stop down for more DOF for 50mm and shorter. Longer than that, it is only a matter of blurness. DOF varies with distance too. For a tight head-and-shoulder shot, you may need f/5.6. But for a loose three-quarter or full body shot, even f/2 suffices.
I have done newspaper tests for my lenses, with slides, of course. Edge sharpness improve on stopping down, due to curvature of field. Contrast improve significantly within one f-stop. There may be visible light falloff, but it also disappears within one f-stop. All in all, the lenses are usable wide open. I am talking about f/1.4 and f/2 here.
All lenses are compromise. Optical quality, Size/weight, price; choose any two.
Beginners love the Superzooms: zoom lenses with zoom range starting from wide angle and ending in telephoto. Such lenses have zoom range from 28-200, 24-200, 28-300, and even 50-500.
I must confess my first lens was a Tamron Superzoom 28-200/3.5-5.6 lens. No, don't throw stones at me. I didn't know better. But now that I do, I advise everyone who cares to listen on the alternatives.
What's wrong with superzoom lenses? Well, we can ignore their build quality. They are plasticy, although that's no fault of theirs. Noticed that they are priced cheaply? Well, that's a fairly reliable indication of their overall quality. I was surprised to be able to buy the Tamron lens new at S$380 (Oct 1998). What a bargain for a lens! Today, I have a S$380 paperweight. What a bargain for a paperweight? Superzooms are targeted at the low-end of the market, so cost is very important. As a result, the construction and glass quality goes out of the window.
One of the most important parameters in buying a lens is, guess what, its zoom range. The more the merrier. This is especially attractive to someone who upgrades from a P&S, where a longer zoom range is advertised as a highend feature. Every beginner wants convenience and hates missing opportunities, so a huge zoom range will help. But they didn't realise these:
1. Slow. I am not talking about AF speed, but the aperture. The Tamron 28-200, for example, is a f/3.5 to f/5.6 lens. 28/3.5 is slow, but still pretty decent, as most consumer range lenses start with this too. On the long end, it is 200/5.6. This is pretty slow.
What happens when a lens is slow? Believe it or not, you will miss shooting opportunities. Compare f/2.8 vs f/5.6. You can shoot in four times less light. You can use f/2.8 1/125 vs f/5.6 1/30. At 100mm, you can handhold the first and be acceptably sharp. For the second, you will get much more blurred shots, even if you braced yourself.
Use a tripod? This is often not an option for street and people photography. How about flash? Ever wonder why indoors or in "low" light shots never look like the real thing? Could it be the flash? But then you can't turn it off. Otherwise nothing will come out.
One thing lens maker never advertise for variable-aperture lenses is when the lenses change aperture. We often assume the change from f/2.8 to f/3.5 to f/4 to f/4.5 to f/5.6 is proportional to the zoom range. Not always. More often than not, the lens go to f/5.6 all too quickly.
2. Selective focus. This is the art of drawing attention to your subject. It works simply by having your subject in focus, while everything else is blurred. Now, why should I do this? I want everything to be in focus! But you should give it a try. You might like it. I do. But ever tried it on a superzoom? Can't get it to work, right?
The reason is because of the aperture. Among other things, depth-of-field (DOF) depends on the subject distance, focal length and aperture. 28/3.5 has a lot of DOF. 200/5.6 is somewhat better, but you still don't have a really blur background.
3. Manual focus. You bought an AF camera because it can auto focus, right? You know it can manual focus too. Do you use it? Why not? Perhaps it is because the camera's AF is good enough. Or perhaps it is too hard to manual focus.
To manual focus well, you need both a suitable camera body and lens. Yes, the camera body matters too. So go get a body with big and bright viewfinder first. But after that, the lens matters. Try manual focusing. Can you tell when your subject is in focus? Do not cheat by looking at the in-focus indicator! It's very difficult, isn't it? You give up and go back to AF.
It's because of the lens. Smaller aperture means more DOF. Since SLRs are viewed through the lens, the lens have direct impact on the viewing quality. A smaller aperture lens will make the viewfinder darker. You also see much more DOF, because it is f/4 instead of f/1.4.
In my experience, there is not much improvement in brightness once the aperture is at f/2.8. You'll never see the difference in good light. Try indoors. There is a difference. f/2 vs f/2.8 is very subtle. f/1.4 is also only marginally better than f/2.8.
But the DOF is noticeable. f/1.4 and f/2.8 is quite different. If you use one of the f/1.4 or f/2 lenses, you will finally understand why you need DOF-preview. The background that is totally blurred at f/1.4 is not at f/5.6.
The small DOF makes it easy to manual focus. You can tell by eye when your subject is in and out of focus.
Lenses still have some way to go before they are perfect for manual focus, though. AF lenses are optimised for AF, so they often have a short focus throw. This makes it difficult to focus precisely. Also, the focusing rings are often too small and they rotate in AF.
4. AF performance. The wide aperture also helps the camera to AF better and surer in poor light. No more AF hunting!
5. Close focusing. How close can your lens focus? Usually not close enough. Sometimes, you need to take a close up, but you have limited space to back up. Now, most zooms have higher magnification ratio (1:4) than single focal length lenses (1:8). Unfortunately, it is usually at the long end. The problem is that the perspective is different. And, you need to back up a long way to capture a wide area.
It is quite easy to estimate how close a single focal length lens can focus. It is usually very close to the focal length. For example, a 35 mm lens may have a 25 or 30 cm minimum focusing distance; a 50 mm lens may be 45 cm.
6. Optical quality. This is the ultimate test of a lens. After all, the output is what other people judge your skills by. Not content with being slow, the lens must be stopped down for "better" performance. But it never really get better. The lens is always soft, the colors bland. There is always light falloff. And the lens is very prone to flare.
Now, it is absolutely true you can take good photos with it. But ever wondered how much better it could have been? You won't realise it until it hits you in the face. I was happy with my Tamron 28-200 lens. The photos are sharp. And colorful, if I use ISO 100 film. Thankfully, my ignorance was short. The first roll I got back with my 50/1.7 lens. They were.. not sharper at all. But the colors.. the 50/1.7 totally blows the Tamron out of the water. And I am using a ISO 400 film. It is only tens of rolls later when I got used to the quality, that I looked at my Tamron shots and exclaimed, "Oh my god, it's so soft!"
A poor lens is always soft and renders details poorer. You may not realise it until you compare two identical shots side by side. And oh, if you still cannot honestly tell a difference, it's time to change lab. Yes, labs make a difference.
Color is also very obvious. A good lens gives good colors and has a high contrast range.
There are sweet spots, of course. A zoom lens performs best at some focal lengths and at some aperture. But you bought a zoom lens presumably because it can zoom, right?
7. Size/weight. Due to its plasticy construction, the lens is "light". However, it remains big. It takes a big filter too. 72mm is not uncommon. You should compare the price of a 55mm and a 72mm filter sometime.
Everyone is, or has, considered getting a digital camera. I have too, and have decided it is not for me yet. Now, I just evaluated based on just the camera. I haven't considered the supporting infrastructure yet. When you do, it is even more not worth it.
1. Startup cost. You need a PC. Not familiar with one? Better be. A PC works well most of the time. However, when things go wrong, do you know how to troubleshoot? You need a color inkjet printer, preferably one with 6-ink and gives photo quality. Depending on your morality and processing needs, you may need to buy a more full-featured imaging package. Lastly, you need a CDRW drive to archive your images.
Most likely you already have a PC. But if you bought it before last year, you may need to upgrade the CPU, the memory, the video card, the hard disk, get a CDRW drive and a bigger monitor. hmm, this is a new PC, isn't it?
2. Time. You will spend time downloading, doing digital touch-ups, then printing. Then you will be dissatisfied and do it all over again.
3. Running cost. You need ink and photo quality paper. The ink will finish faster than you think. And you need to throw away the whole 3-ink cartridge even if only one ink is exhausted. The Premium Photo Paper gives premium results. But you pay premium price for it, of course.
4. Life span. Is the printout light-fast? How about water-fast? Do you really believe the marketing department when they say the image won't fade for 100 years? I'll wait for it to withstand the test of time.
5. Photo printout. Suppose you want a real photo. Go to a lab and ask for the pricing. You will be shocked by the prices. Well, it is not so bad now; slides-to-print is even more expensive. But the handling fees still apply. You can print in bulk to amortise the cost. And being such a good customer, they will probably give you a discount too. We are talking about 100 images at the minimum. Do you have so many to print? How long do you take to accumulate 100 images you want to print?
6. Media readability. Let's say you store your images in TIFF or JPEG format. These formats should be around for a very long time. Even if they are not, you can always write your own image reader from their specs. However, ever considered that the media itself cannot be read? It may just happen. DVD, which superseds CD, is not compatible to it. However, DVD players read CDs for backward compatibility. This is done using a separate laser head. Don't think that one day DVD players will not be backward compatible?
The music industry has obsoleted several media formats. However, if you are desperate enough, you can still read them. How about the IT industry? The 1.44 MB floppy media is finally dying. You may have problems reading the Zip disk in twenty years time. Still got a 5-1/4" 360 kB disk? Even if you have the hardware, can you plug it in? Do you have the drivers for the OS? Yes now, but try considering the future.
It is worse if you store your files in a proprietary format. You need to find the OS to run the application. Can you?
I used Java for the last year and C++ for three years before that. While serving notice, I had nothing to do. On two occasions, I decided to give a primer on topics I am familiar with. One of the topics is C++. My (ex-) colleagues in the UOB project were all Java developers. They either knew only Java, or has lost touch with previous languages. This should not be the case.
Personally, I am biased against Java, but that's me.
One things about C++ I mentioned to them is the many faces of C++. C++ is a fairly large and very flexible language. You can choose to use subsets of it, and you still can say you are using C++.
1. Improved C. C++ is largely compatible with ANSI C, so much so that most of ANSI C can be compiled using a C++ compiler without change! But is this C++? You decide.
2. Classes. C++ allows you to group attributes and methods in a class. This facilities encapsulation.
3. Class hierarchy. You use inheritence and polymorphism. This makes it easy to manipulate devices or messages using a common interface.
4. Exception handling. You use C++ exceptions for error handling.
5. Template. You either use your own template or more likely, use the STL.
These are not mutually exclusive. You can have any combination at any one time. Why do I list these as different faces? I have little experience in using exception handling and templates. I tried, but gave up because they required too much resources. I have read sample codes in books, and frankly, I can't tell they are C++. I know they are, but I don't know what they mean.
So, when you meet someone who says he know C++, make sure he is referring to the same C++ as you are.
Today marks the last day my (ex-) colleagues work in Central Plaza, Tiong Bahru. They are relocating to Tampines. It all came about after UOB (United Overseas Bank) absorbed OUB (Overseas Union Bank). UOB is consolidating operations. They are moving to a OUB office in Tampines.
On Thursday, I visited the office for the last time, to retrieve my tripod -- my last belonging there. There, I saw the corridors filled with stacked boxes, shredded papers, empty cubicles and people packing up. I was brought back one year ago. First, a short history.
I worked in my (ex-) company since I graduated. I joined one of their projects, which later resulted me being seconded to the client as a "Consultant" ten months later. My company certainly received a Consultant's fees. I got my regular pay as usual. Guess what kind of work I produced? This was a source of major contention between me and my client at times. Two years passed. All along, I resisted all calls for me to quit, because I like the work. However, liking alone does not pay the bills, so there were times when I was depressed, tired, burnt-out or just plain rebellious.
Time for contract review. The client did not want to renew. Why keep paying a Consultant's pay for a Software Developer? They offered me a job, but I decided not not to take it up. The time to leave is always very sad. In my two-and-a-half-years there, I have made many friends, offended more than a share of them, joined them in many functions and events, and even witnessed one of their most troubled product launch. I have broken most rules, including ones I have no idea about; said or done the most inappropriate things.
It took me quite a while to pack up. My stuff filled three of my colleague's cubicles, not counting my own space in the lab. It took three trolley trips to load all my stuff into a colleague's car. I took a photo of all my stuff before I packed them up. When I showed it to my colleague, he understood why no one used his cubicle in the six months he was away.
Coincidentally, the client was moving out of its premise too. Soon after when I revisited the office, the corridors were filled with boxes and the atmosphere is.. different; you can feel it. Also, the empty cubicles, they give you an empty feeling. You feel alienated because you can't relate to the office anymore. This was the exact same feeling I felt when I stepped into UOB's office.
To experience it again, I felt so sad. And, I felt even sadder that I may never see some of them again. Then again, I realised that even if I meet them outside, the feeling will not be the same. The work relationship, the office context, the daily encounters and so-on, these will be what I miss, not the person per-se.
I joined the UOB project in mid-July last year. This means I only know most of my new colleagues for only one year. It certainly felt longer! To think that in less than one year all of us would be going our separate ways. Also, I hadn't expected to befriend the UOB staff so well. After all, they were the clients.
All good things come to an end. Enjoy and cherish it while you can. Once I read a qoute that basically said, do not miss the journey for the destination. I didn't realised what it felt like until my friends' graduation (I graduated one year later). Nothing stays the same forever. People change. Projects end. Enjoy the present.
For one, I will miss the photo opportunities that my (ex-) colleagues have given me. :-)
I just joined a new company, the result of an involuntary -- but long expected -- retrenchment. No retrenchment benefits. Nope, none at all. It is my assessment that I have given the company more than what they have given me. I am not very calculative in nature, so I never expected to be properly recognised or rewarded. However, this is the final straw for me. I should have known better. How blind and stubborn I have been! I am reminded by a quote one of my professors once posted in the DISCS (Department of IS/CS) BBS: "Fool me once, boo on you. Fool me twice, boo on me."
But I am side-tracking. The intention of this column is to talk about my next PC. We are currently witnessing a technological revolution in progress, if you are still not aware. Technology has improved by leaps and bounds. It is jaw dropping if you think about it. Electricity was discovered only two hundred years ago. We could only fly at the turn of the 20th century (1904). We only have RADAR in the middle of World War II. Man left Earth for the first time only forty years ago.
More down-to-Earth. The first computer, ENAIC, was constructed after WW II and filled one whole room. Despite this, it has less computation power and memory of any electronic calculator today. Just in 1990, CPUs were cruising at a mere 33 MHz, struggling to achieve even 5 MIPS. 16 MB memory runs every bloatware and bloat-OS you can throw at it, if you can find a motherboard to support it, and the cash to buy it. And, most hard disks are smaller than CD-ROM (650 MB). Monitors are 14" (12.5" viewable). Cutting edge video resolution is 800 x 600, 64k colors. PCs were just starting to get "multimedia". No music. No movie. You want to print something? A 24-pin dot-matrix printer doing two-pass prints very nice letters, thank you very much.
You must be shaking your head, wondering how the computers can be used. But they are very much usable. I will leave this to another column. Now, you must be wondering what people say about 1980-vintage computers to in 1990! I wonder too. :-)
My new company really knows how to pamper its R&D Engineers. On my first day, I got a brand new PC, my company's own brand workstation. Specs: 2.4 GHz Pentium 4, 512 MB RD-RAM, 34 GB SCSI hard disk, dual-head Matrox Millenium G450, 21" flat-screen CRT monitor. Other features are nothing to shout about: integrated soundcard, built-in speakers and an ordinary CD-ROM drive.
Impressive? And, to top it off, other than reading email, I don't know what to do with it. The real development work is done on a Unix workstation, again my company's own brand. Specs are nothing to shout about, except that it also comes with a 21" CRT monitor.
The point is, the hardware has become much more powerful than we, and our software, can make use of it -- with the exception of authoring movies. Therefore, other than raw speed, power and sheer storage, we need to look at how we can make our PC work for us, instead of the other way round.
Earlier I mentioned a revolution. What I meant was, PC is not just desktop PC any more. Notebooks are already a well-established category. In recent years, we get the incredibly shrinking cellphone, and finally, palmtops are small, light, powerful and does not eat battery two at a time.
When you were young, have you ever looked at the walkman and wondered how small it can get? Those ancient days? For the younger ones, subsitute the Discman. Today, MP3 players are no bigger than radio receivers.
Photography is not exempted. Digital photography is making a huge impact. Film is.. retreating. The writing is in the wall.
To add to all these, technologies are converging. Everyone doesn't really know how the pieces fit together, so they are just experimenting by combining two or more functions in the same device. A palmtop that is a cellphone, a cellphone that can take picture, a camera that can play music; you name it, it's there.
But, until directions are clearer, I am not so ambitious. I am not a gadget guy, so I don't have any of the devices mentioned above. Also, most digital devices are more for keeping-up-with-the-Jones than real productivity. Predictably, my next PC will be a notebook. I am very satisfied with my current notebook, the Acer TravelMate 611TXV. After using it, I have very clear preferences:
1. Adequate processing power and storage space. This is not so hard to achieve. Try finding one notebook slower than 700 MHz and has less than 10 GB hard disk space. Although the 611TXV is capable of 850 MHz, I run it at 700 MHz. I wish the Intel SpeedStep allows finer control. It is still pretty fast.
2. Weight. 2.3 Kg is the max for me. I almost never detach the DVD drive, but in case you do, check that a weight-saver is available.
3. Screen. I have grown very used to the 14.1" screen, with a 1024 x 768 resolution. A 13.3" screen, at the same resolution, may be usable. A larger screen is good because you get a higher resolution: 1280 x 1024 or even higher. I run my desktops at 1152 x 864 minimum.
Small notebooks are unusable for another reason: the keyboard. I bought an external keyboard and mouse almost immediately after I bought my notebook. I used the mouse a lot, but I never used the keyboard. The internal keyboard was wide enough for me. I can type at 70% of full speed using it. Good enough.
4. Video Card. The 611TXV uses the Intel 82815 Graphics Controller with (non-dedicated?) 11 MB video RAM. It is adequate for desktop applications and DVD playback. However, it is kind of slow for modern 3D games. Luckily I don't play many games. Speaking of this, the notebook gets extremely hot for games and DVD playback. I always use a fan to cool the notebook.
5. SPDIF output. I always have a fetish for DVD playback using a notebook. A notebook is actually very suitable as a HTPC -- Home Theatre PC. The reason is because it is very silent. Did you realised how loud your regular desktop PC is, with all the fans inside whinning at full speed? Not to mention the 12x or 16x DVD drive that spins loud enough to spoil all movie experience.
SPDIF is the missing link. Without it, notebooks only support a stereo line output. Now that external "soundcards" are available (see the Creative Audigy), this is not an issue any more.
6. S-Video or component output. This is a plus, since it allows easy connection to regular TVs. Real HTPC output to projectors. These output display in feet, not inches. All projectors accept VGA input, which all notebooks mate well with their VGA output. The only problem is that even an entry-level LCD projector costs S$3.5k, and it works only in the dark.
7. Interfaces. It is a good thing that the bulky parallel and serial ports are finally phased out. Future connectors are USB or FireWire. Even the PS/2 port can be left out. On the other hand, a Compact Flash slot would be very useful.
What else? You tell me.
A colleague noticed that I put my notebook in my unpadded bag and asked if this won't scratch the notebook. My answer to her was, yes, it would. However, I didn't really give her the full answer. Here it is. Some thoughts before buying a notebook.
If you are in the market for a notebook, stop. Are you putting your priorities right? If you treat it like buying a desktop PC, by focusing on the specs, you will get a desktop notebook, not a true notebook. The reason is very simple: features, size/weight, price, choose any two.
Once, in my previous company, I was able to borrow a shining new notebook for use at home. It was full featured, very powerful and complete with a 14.1" LCD. It is even able to play audio CD without powering the notebook! Once, I played a CD waiting for the rain to stop on the way to work -- I was not in a hurry; after all, I was only 45 minutes late. It is quite an experience to hear music emitting from your bag. AM radio quality. But cute.
Back to the notebook. You always need to add at least 300 grams more for the AC adapter. After a while, you begin to notice the weight. When does it start to become heavy? For me, it starts at 2.5 Kg. This rules out the majority of the notebooks out there. You can't really tell by lifting the notebook up. You need to put it in a bag, sling it over your shoulders and carry it around for several days. Then you can tell.
When it was time to return the notebook to the company, I had grown too fond of using a notebook anywhere -- and I mean anywhere. I watched DVDs on bus, MRT, on a projector in a friend's office; surfed net in the living room, bedroom, sofa, lying down. You are no longer constrained by the office walls. If you really want, you can go to a cafe to finish up your presentation or work. I can maintain one continuous computer session day after day, in office and home. It will redefine the way you use the computer.
I started looking for a notebook to call my own. Due to my direct experience, weight is an important consideration.
Of all the notebooks available then, the Acer TravelMate 611TXV is a strong contender. It is full featured and sports a 14.1" screen. Also, it is the lightest notebook in the 14.1" screen category. It weighs 2.26 Kg with a DVD drive, 2.03 Kg with a weight-saver.
Downsides: it uses an external USB floppy drive, but who uses floppy these days? It needs a separate cable for parallel and serial port access. This cable is supplied, but it is cumbersome. However, I print to a network printer, so I don't need it. I also don't have any serial devices. Mouse and keyboard are either PS/2 or USB. It also has a pretty dim LCD, but I only realised it after the fact. Well, nothing is perfect. It also has no S-Video output, so my dreams of connecting to the TV went up with the smoke. Lastly, the audio-out has very high noise.
I bought the notebook exactly one year ago (give a few more day's time). I have no complaints about it so far. It runs perfectly fine. I installed Windows XP when it came out. This really allowed me to use one session for weeks at a time. Memory was never enough, so I bought one additional strip of 128MB SD-RAM. It is interesting to note that the shop accidentially gave me a 256MB strip. I didn't realise it myself until I booted the notebook and wondered how I got 384 MB of memory. The price was S$33 vs S$150. Guess what I did? (This will give you a hint about how much you know about me.)
I use my regular bag instead of the supplied notebook bag. This was my intention all along. The supplied bag does have very good padding, but it adds weight -- a lot of weight. Besides the bag, people tend to carry too many accessories around. For me, I don't even carry the AC adapter. I bought a spare, so one at home and one in office. It costs a lot (S$170), but the saving in weight is worth it. There is a cheap and low-tech method: build your own AC adapter. Yes, it can be done. Just buy off-the-shelf parts, match the output voltage, amperes and polarity. Not too difficult if you are from Engineering or are interested in electrical/electronic stuff.
Using an unpadded bag has its risk. I was worried about knocks and falls, so I was very careful. Beyond the physical safety of the notebook is my data in the hard disk! Backup is very important, but I confess doing it less than enough.
I dropped my bag a total of three times and had several hard knocks. In the first drop, the battery-catch broke and there were some dents. As they say, the first scratch is always the most painful. After this, I was not too concerned with the physical appearance. But I still remain careful, of course. A few times I was careless and I put stuff in my bag that scratched the top cover. Ouch!
My bag is a regular soft sling bag. It was bought in my student days and has been through sun and rain with me for the past six years. Besides the notebook, I put in my water bottle, which is those S$1 mineral water bottle, beside the notebook. I also put my compact camera either above or below the water bottle. The whole bag remains very slim. No one would believe a notebook is inside.
I am also able to put a couple of books parallel to the notebook. This makes the bag very heavy. The wider the bag is, the heavier it becomes in proportion to the actual weight. I suspect it is due to the sling angle. Currently, I carry a Nikon FE2 with an unmounted Nikkor AF 35/2 lens in the bag. If I really want, I can unzip the bag to make it even wider. A very versatile bag. However, it is because the 611TXV is only 3 cm thick that I have so much space left.
The 611TXV has served me well. Not even in its first year anniversary, Acer has came out with the 620 series and the 630 series. Besides a more powerful CPU, they also have a larger screen of 15", a decent video controller and better power management. Unfortunately, they are also heavier. I am keeping a lookout for a new notebook; looking at large screen (15", >=1280x1024), with a good video controller. But weight is always on my mind.
The Acer TravelMate 611TXV. No regrets at all.
The problem for beginners is that, you rarely know your usage style until you have actually used a camera for some time.
And the wrong choice of camera body can actually hurt your "advancement" more than you think. Of course, if you use the P mode all the time, you don't miss much.
Here's what I'll miss:
1. Bright viewfinder, high viewfinder (0.8x) magnification and frame coverage.
This, together with a wide aperture lens, allows for easier manual focus. If you find MF too difficult, perhaps you have the wrong choice of camera and lens!
Entry-level models usually use mirror for their viewfinder instead of a real pentaprism. This is much darker. This is especially noticable indoors. Even a fast lens with f/2 aperture can't save you here.
The viewfinder magnification refers to the 36x24mm frame. I believe there used to be 1x viewfinders, but now the highest is the FM2/FE2, at 0.86x. Most entry-level cameras, in contrast, offers less than 0.7x! Very few AF cameras go above 0.75x.
A related issue is the viewfinder frame coverage. Most entry-level cameras give below 90%, most semi-pro give 92%, and only the very high-end gives 97% or 100%. I would do a test to see what it actually means. You will be surprised how much you don't see in the viewfinder.
2. Interchangable focusing screen.
Even if you don't use it now, this tells you whether the camera is positioned towards the serious or niche photographers. The most common optional focusing screen is the grid screen. This is useful for architectural shots, where straight lines must remain straight.
Sometimes, the camera comes with a focusing screen that is very bright, but at the expense of contrast. It is extremely hard to manual focus, even with a f/1.4 lens. Be sure it can be changed if you want to manual focus.
Be sure to find out if the focusing screen is user changeable. Otherwise you may be without a camera for several days.
3. Vertical grip.
Do you take many vertical shots? If not, perhaps it is due to the awkward holding position? If there is a vertical grip, see if there is more than the shutter-release button on it. At the least, it should replicate all the controls on the backcover. It may or may not take extra batteries. And, the real test is if it can be detached just like that and the camera remains operational.
4. High flash sync (1/200).
This allows higher shutter speed even in bright light. Now you can make the background blur! HSS works as well, but at much reduced power.
5. Depth-of-field review.
Absolutely essential for macro and portraiture shots, especially if you have a fast lens. At f/1.4 or f/2, the background is always blur. You will have a surprise if you set the aperture at f/5.6 or f/8. Always check first.
6. Spot meter.
Always using matrix metering? Don't understand 18% gray? Don't know how to determine the exposure? Perhaps it is because you never have a spot meter. Or if you do, how easy is it to switch between spot and matrix?
7. Auto-exposure lock (AEL).
Absolutely essential when using aperature-priority mode. Without this, you need to go manual mode much more often. Best used with spot meter.
8. Flash exposure compensation.
Essential to control the ambient-to-flash ratio in TTL flash mode. Of course there is a workaround, but do you really want to use it?
(I seldom change my flash exp. comp. setting, but if you do, make sure it is easily accessible.)
9. IR AF illuminator.
Good to have if you AF in low light a lot. Otherwise the lens will hunt a lot. Don't blame the lens.
10. Noise.
Want to be stealth? Check out the AF noise, mirror slap noise and the film advance noise first. And don't forget to check the film rewind noise!
AF noise can be eliminated totally by going manual-focus, which unfortunately is not considered by many people at all. In-lens motors are usually much more silent, so Canon's USM and Nikon's AFS win here. Minolta is coming out with their own version, called SSM.
Ironically, semi-pro cameras usually have a noiser mirror slap than lower-end and pro models. This makes them totally useless in quiet environment.
Some cameras have very quiet AF and mirror slap, but an annoying film advance noise. This defeats the whole purpose. I advise checking this with a roll of (unwanted) film to see the noise is acceptable. Entry level cameras usually have loud film advance noise.
Lastly, if you ever need to change roll during a wedding or graduation ceremony, you will realise film rewind noise is a killer. This must be relatively silent, because it takes 18 seconds to rewind (at 2 fps). If it is noisy, 18 seconds is an eternity. Pro cameras can afford to be noisy, because they can rewind very quickly. A EOS 1v can shoot at 9 fps, so it can rewind in 4 seconds.
11. Low shutter lag.
How fast does the camera take the photo after you fully pressed the shutter release button? Ever notice the lag? If no, good. It is easier to tell if you compare two cameras side-by-side.
12. Short mirror blackout time.
SLRs flip the mirror up the take the shot. Entry level cameras are usually slower in flipping the mirror back, so it will take you longer to get ready for the next shot. This is somewhat related to the motor drive speed. Cameras with 1 fps have very long mirror blackout time. But even entry-level cameras are now 2 fps. These are much better.
To test, set a high enough shutter speed, say 1/125, look through the viewfinder and press the shutter release button. See if the mirror blackout time interrupts your concentration on the scene. Note that the mirror blackout time is constant, so it is noticeable even if you set a low shutter speed like 1/15.
A few other nice touches:
Compact cameras, also called Point-n-shoots (P&S). If you ever need to buy one, here's my opinion.
It is unfortunate that most people, including me in the past, are specifications crazy. Also, the bigger the zoom range, the better. What a mistake!
Perhaps you think that P&S can never give results comparable to
SLRs. But perhaps you are wrong -- if you choose the right P&S. There
are good P&S. Some are more expensive than the others. In the low-end
of the market, there are two affordable ones: the Yashica T5 and the Olympus
Mju II. Here's what I think of them, in a nutshell.
I shot some 30 rolls over the past 2-1/2 years (since Nov 1999) before I lost it in Mar 2002. All I have left is the pouch, which I reuse for the Mju II.
The waist-level viewfinder is the reason why I chose the T5 over the Mju II. It turned out that I never used the feature much. The 35mm focal length is relatively wide that you can estimate the scene pretty well. However, the spot AF sensor makes it difficult to shoot blindly. Thus, I started using the waist-level viewfinder to check that my subject is within the focus sensor.
The shutter-lag is very annoying. I lost many shots because of it. Either the composition has changed, due to the dynamic nature of the scene, or the person's expression has changed.
It is relatively silent. You can get shot after shot without people knowing -- if you don't look through the viewfinder. Unfortunately, the spot AF sensor makes this difficult. I've got one truckload of out-of-focused photos to show for it.
The T5 has a reasonably wide aperture of f/3.5. "Reasonably" is relative. f/3.5 is wider than all other P&S, especially the zoom P&S. However, it is not much wider. Most P&S start at 35mm f/3.8 or f/4.5. It offers no control of the aperture. However, you can predict its behaviour pretty well. Good light, it stops down. Poor light, it opens up. It is not possible to force it to use a wide aperture for protraiture and a small one for landscape though. One thing about the lens is that it has very noticable light-falloff wide open. How noticable? Well, it's visible on prints, that's how bad it is. (Prints are cropped about 5% on each side.)
The T5 is also pretty big. I thought it was small, until I tried to walk with it in my pocket. Not natural at all. However, it feels very light for its size.
People complain that the camera does not remember the flash-off state. I have very few problems with it. Whenever I turn on the camera and I think it will use flash, I just press the flash button a few times.
As already mentioned, the spot AF sensor makes it difficult to shoot blindly, which is my style of shooting. It is difficult for casual users too, as my parents and brother can attest. They frequently forgot to focus, lock, then compose.
Now, everything I said has been negative! But, when things work well
-- the result is incredible. The photos aree sharp. Sometimes they are as
sharp as the ones taken with my SLR lenses, sometimes not. And the colors,
they are saturated; very saturated. I love the colors.
I bought this to replace my lost T5 in Mar 2002. Fed up with the T5's shortcomings, I decided to give this camera a try. I shot 11 rolls so far.
My first impression is, the camera is so small! It fills only about 60% of the T5's pouch. It has a clam-like shell protecting its lens, giving it weather-resistant capability. And it looks slick.
The Mju II has a reasonably low shutter-lag. While it still requires time to focus, it is pretty short. So far I have not lost any shots yet.
The wide AF sensor makes it easy to shoot blind. However, the film advance is very noisy. Everyone will know you have taken a photo. So, if you think it would affect the mood or people's reaction, be stealthy for the first shot and make it count!
The noisy film advance makes it difficult to shoot in places where photography is restricted. Luckily, the ambient noise outside drowns out the camera's noise easily. Also, it is easy to slip the Mju II in and out of your pocket. Still, it is not as quick to use, as the camera does not remember the flash-off mode. You must set it again after you "reset" the camera, or risk everyone's attention.
On the few occasions I used the flash, I get red-eye almost all the time. I seldom use flash, but it is needed sometimes for fill.
The photos are sharp enough, but they are definitely not comparable
to ones taken with my SLR lenses. It may be possible to tell them
apart, but with lots of mistake, of course. As for colors, I am disappointed.
The photos are not colorful at all. They look rather normal with Fuji Superia
400. More realistic, perhaps.
How about the lack of zoom? Isn't it a huge hindrance? No, it is not. You just need a change in mentality. Most people think they absolutely need a zoom. But think about it, do you frequently use the focal lengths in-between? Most of the time the view is either not wide enough or the subject is still too far away. No zoom can solve this.
Most people think to take a close up shot of a something, you need to zoom in. There's another way: stand closer. While not applicable all the time, this is the better way. You may be shy or afraid, but you should also try it. The aperture for telephoto on P&S is too small anyway, at f/9.5 or f/11. Absolutely useless indoors.
You should learn to move around and discover how to the background changes with respect to the subject. It will take more effort, but it will also make you a better photographer.
These days, I always ask myself, "Is that all to me?", "Have I realised my potential?". What is the answer?
My advice to all friends, colleagues and any one who cares to listen: do not be complacent. You may have it good today, but will the carpet be pulled from under you one day? Make sure that day never comes by always keeping one step ahead.
Not Here Yet.. may I point out, is my initials. :-)
Two years ago, I stated that digital cameras have several shortcomings: low resolution, limited memory capacity, high shutter lag, poor battery life and noise in low light. Have things change?
Yes, if you own the state-of-the-art EOS D60 or the even more-state-of-the-art than state-of-the-art EOS 1D, you will have one digital SLR that performs.. almost as well as a film SLR.
I do not intend to give you a review of the EOS D60 here. Go read a detailed review of it somewhere and see what it can do. Okay, now that you are back, we can continue. But first wipe the saliva from your mouth. Fascinating camera, isn't it?
6 Megapixels, for most of us, is better than film, especially those of us using ISO 400 films. One attractive aspect of digital cameras is that they have no grain, which allows you to enlarge an image and still look good.
The shortcomings? Plenty.
The price. For S$3888, you can get the EOS D60. Or you can get the state-of-the-art film camera EOS 1v. The D60 is equivalent to the EOS 30, which costs S$850. As such, its features are closer to it than the EOS 1v. Say, how much would a body based on the 1v cost? Only a cool 10 grand, and with only 4.1 Megapixels. Pay more for less? We'll discuss more about this later.
Did I mention the price is for the camera body only? No lens and no vertical-grip? S$3888 can get you a pretty good camera (EOS 30) and several very good lenses.
No matter how you look at it, it is expensive for an amateur. Professionals are different. They can easily make back the money with a few sales or contracts.
The interface. Okay, you can afford the camera. Now, can you use it? Use it well? As it is now, all digital cameras still do not have an intuitive interface. There are simply too many functions, so there has to be a menu-based navigation system. The problem is when some functions are nested deeply and you need to change them pretty frequently. It will frustrate you.
Why is it like this? Digital cameras do not just take pictures. They allow you to view and process the images on-the-spot. That's why they are more complex. If you don't do it, good. But I doubt there is not one digital photographer who is not over reliant on the LCD and the histogram. How did we ever survive before this, I wonder?
Shutter lag. I'm not actually referring to shutter lag, which is the time from when you press the shutter release button to the time the shutter actually release. This is pretty quick. I'm referring to the cycle time; the time you can take the next shot. This generation of D-SLRs has a buffer that allows you to shoot several images before the buffer is full.
The question is, how many images? At what resolution? The specs give a very good figure, but this is meaningless if you shoot raw. And one thing to look out for is the buffer write-out time. When the buffer is full, it takes a while before you can take the next shot, and the next next shot.
Limited capacity. How many rolls of film do you shoot in one trip? If you shoot very little, digital cameras are good for you. If you shoot a lot, they are not here yet. 128 MB flash disk is now standard. How many images can you fit in it? It depends on the size of your images. A 3072 x 2048 image (as produced by the EOS D60) will need 7.4 MB (at 12-bit). This is if you store it raw. 7.4 MB per image, 3 images per second! Now you realise the bandwidth requirements.
So, let's say we compromise and store in JPEG-Large/Fine (3072 x 2048), which people say has no visible effect. It takes up just 2.5 MB (an estimate given for the D60). If you don't need the quality, you can store in JPEG-Large/Normal (3072 x 2048) which takes up just 1.3 MB.
But what's the problem with JPEG? First, it's a output medium. As such, it is lossy. It discards not-important data in order to reduce the size. You lose the data forever. While this is not important for an output medium, it is important for digital cameras, where acquiring the image is just the first step in the digital workflow.
This is the reason why we have all the Histogram, White-Balancing, Brightness, Contrast, Color Saturation, Sharpness controls -- and one basket full of filters -- in the camera. These have to be done before the JPEG compression to get an optimal result.
JPEG introduces artifacts and removes the fine details. The first is obvious. You'll know it when you see it. The second is not so obvious. JPEG works by discarding high-frequency data. What this means in practice is that the fine details are removed. Look at fine lines and areas of high contrast. They will be smoothened. Strangely, this means a smoother out-of-focus rendition, which can be a good thing.
Another thing about JPEG currently is that it only supports 8-bit per color channel. Currently, digital cameras are capable of 12-bit colors. The extra bits come in useful in processing. You can manipulate the image without introducing banding and other artifacts.
So, raw or JPEG-Large/Fine? Your call.
Back to the storage. For a 128 MB flash, you are looking at 17 raw images or 51 JPEG-Large/Fine images. That's definitely not good enough. What's the biggest capacity available? Either a 1 GB Microdrive or a 512 MB CF (Compact Flash). You opt for the latter because the former is non-solid state, so it is fragile and prone to mechanical failure. At 512 MB, you can store 69 raw images or 204 JPEG-Large/Fine images. Good enough? Consider that 69 images is only two rolls of film. How many 512 MB CFs can you afford?
Battery life. Digital cameras still eat battery for lunch. Not surprising, considering the computation power required. How long can one single charge last? Can it last at least one full day of shooting? If so, how about the next day? Do you have a place to charge it? How heavy is an extra battery?
Noise. The EOS D60 uses CMOS instead of CCD for its sensor, so it has less noise at high ISO sensitivity. The noise I am referring to is not audible noise, but digital noise -- random pixels that don't belong to the image. They show up in low light, because the light signal is small compared to the inherent sensor noise.
Current digital SLRs work well at ISO 400. They work pretty well in low light too. A compact-class digital camera may produce acceptable results up to 2 seconds. The SLR-class cameras can go up to 30 seconds and more.
One "cheat" which many manufacturers do is pixel remapping. This solves the hot pixel, dead pixel and noise problems. The pixels are still there, just that they are hidden from you.
Other factors to consider.
Latitude. Think slide film. Both CCD/CMOS and slide film have limited latitude. 5-6 f-stops. This is fixed forever. You cannot pop in a different film and get 12 f-stops latitude.
Focal length multiplier. Current digital SLRs use existing 35mm mount for lens compatibility. However, due to technical limitations, most do not utilise the full 35 mm frame (36 x 24 mm). The EOS D60, for example, utilise only 22.7 x 15.1 mm. However, the image is treated as if it comes from a full 35 mm frame, so the resultant enlargement is 1.6x.
This means all current lenses are multipled by this ratio. A 28 mm wide angle lens becomes a normal 45 mm lens. A normal 50 mm lens becomes a short-telephoto 80 mm lens. It is good for telephoto users, as a 300/2.8 lens becomes a 480/2.8 lens. A double bonus, because there is no 500/2.8 lens.
But is this magnification real? It is not. A 50 mm lens is still a 50 mm lens. It captures a certain image. Let's say we enlarge it to 4 x 6". Now, go back to the 35 mm frame, mark the center 22.7 x 15.1 mm, crop it and enlarge it to the same 4 x 6". This is what it all is: an in-camera cropping. We can do it already, if we want to.
Once you get pass this hocus-pocus, you will really be annoyed that you can no longer take wide angle photos. Even the ultra-ultra-wide angle 14 mm lens gives you an equivalent of 22 mm lens. The second thing that will annoy you is the fact that the perspective remains the same. So, if you use an ultra-wide 20 mm lens to achieve a wide-angle 32 mm view, you will get all the perspective distortion and optical abberations of the 20 mm lens.
There are digital SLRs that are full-frame, of course. But they are the minority -- and very expensive. Ignoring them, the best is Canon's EOS 1D, with a multiplier of 1.3x. This is still acceptable.
"Megapixels". You should be aware that each CCD or CMOS sensor can sense only one color. Therefore, 6 Megalpixels refers to the total number of single-color pixels. Currently there are two green sensors to every red and blue sensor, and they are laid out according to bayer's pattern. And clever extrapolation is done during raw->tif conversion. It works most of the time, but it has been known to fail.
New technology is coming. New sensors are sensitive to RGB. These will take a while to catch up.
Flash. It is unfortunate that we have taken a step backwards because we currently cannot achieve OTF-TTL flash (Off-The-Film). We need to use pre-flash, which is irritating. You need the latest flash to do this, and you need a new flash later when we can do OTF-TTL flash.
Non-standard raw file. Currently none of the raw file formats are public and standardised. You need to use proprietary software to convert the raw files to TIFF files. There is no guarantee that the format will not change in the future, or that there will still be software readers to read them. Not very encouraging, right? Digital hardware may obsolete fast, but digital data obsoletes even faster.
Dust. I read on the net this is a common problem for SLRs. Dust is attracted to the CCD. And dust enters the camera every time you change lens. So, don't change your lens often, if at all. Definitely doesn't work for people who use single focal length lenses.
Diffraction. I read on the net this is a problem, especially for cameras with small sensors. This is still not so bad for digital SLRs. At least they can still be stopped down to f/16 and below. Contrast this to compact-class digital cameras. They only go to f/8. This problem strives from squeezing so many pixels on a small sensor.
Weight. The EOS D60 is heavy. Very heavy. The body alone weighs 855 grams. In comparison, the Minolta Dynax 7, which is Minolta's state-of-the-art technological flagship, weighs only 575 grams.
Lens compatibility. Yes, the digital camera uses the same lens mount. But can you use just any lens? Most likely you have to buy a "digital" lens. Strange, but true. Otherwise you will get color fringing, light falloff and the sort. I'll appreciate if anyone can tell me more.
Depreciation. Like all digital products, digital cameras depreciate the very minute you buy it. The camera will totally depreciate in two years time, when its successor comes out. What used to be state-of-the-art will become totally unusable compared to its successor. Film cameras can withstand the test of time, because there is nothing much to improve upon.
Depth-of-view. Note that I never mentioned DOF. D-SLRs can have pretty shallow DOFs, although not as good as film SLRs.