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.
You can find my earlier ramblings here:
Ramblings-2002-07
Ramblings-2002-08
It is lucky that I took leave on Monday to recover. I had to rush back to Johor Bahru early in the morning on Sunday to attend a second full-day event. Early means I reached home at 2:40 am, and I had to wake up by 7 am the next morning. Thankfully it wasn't as early as 5:30 am, as on the first day.
On Monday, I wasn't as tired as I was depressed. It make me kind of anti-social, didn't want to go out (as if that's different from normal), and want to lie down all the time to waste time away.
Since this was not the first time I got this kind of mood, I tried to analyse it. First, usually after I have been in a crowd, and then left alone, I will have some withdraw-ant symptom. Second, weddings do affect me because more and more of my friends are getting married.. and I am still alone. But I usually try not to imagine too much. Third, I drank "some" beer/liqour on both nights. The first night was still okay. I drank from a mug, part of the "The Best Men Challenge", due to the accusations that the best men didn't help the bridegroom, and now the best men wanted to avenge him against the table -- mostly gals -- that made him drunk. I estimated it to be about 1-1/2 cups. The effect was mild, but I can clearly feel it.
The second night was worse. After the bridegroom safely made it through the night and into his suite, it's our turn. So we toasted a few times. Since I wasn't driving, I ended up drinking a fair amount -- maybe even more than the bridegroom. It's very repulsive, I tell you, to drink so fast. I usually take my own sweet time to drink. Luckily they mixed with coke. In the end both bottles of coke ran out, so we stopped drinking.
The effect on me was amazing. I have never got so drunk before. As I confessed that night, I drink from time to time because, there are some things I deny even to myself -- I needed the drink to have the courage to tell myself that they are the truth.
Once I used to wonder what I'll be like when I get drunk. Would I lose my tempers, lose sense, cry, vomit or what? Well, since my ex-company's D&D night last year, I wondered no more. I would be normal. This in itself is a revelation, for it tells me something about myself. Am I really who and what I make myself to be?
Anyway, the full effect came about only after I left the suite. I suffered the physical effects: I could sense from the side of my eyes, but I could not focus on the middle, unless I concentrate very hard, and then not very long. I could walk straight, but I could not make decisions. My response and reaction time were slowed. I definitely cannot drive. Fortunately, I was not driving. Given the condition I was in, I kept asking my friend if he is in a condition to drive. He is, he assured me. I was not so certain, but he seemed steady enough.
And, after he dropped me off, my father woke up. He claimed he heard us talking very loudly outside. That's another symptom: talking loudly. I always wondered if my father knew that I drink, and what his reaction would be. Well, he seemed very forgiving that day.
The surprising part for me was that, the next day, I had no hangover, other than that moodiness. Maybe this is because it was dialuted with Coke. Previously I drank "on the rocks" -- actually, my preferred way of drinking; slowly, of course.
I have achieved a long time target: to find my drinking limit. Having found it, I doubt I will drink much any more. I never drink on my own, and I don't have any drinking buddies -- not that I'm looking for them. I only drink with my friends, and they seldom drink, so I seldom drink too.
Drinking is addictive, I tell you. The sirens of the liqour sound out to me every day...
Should you accept toasting during your table-to-table photo shoot or not? Well, sometimes it can get real scary. Downing one full cup of undiluted wine.. I doubt I can even stand half-a-cup. And this is only the beginning. When the strange mixtures come.. wow, good luck.
If you suddenly got drunk, and the best men didn't help, whose fault is it? Should the best man come to his rescue before he got drunk? Or after?
Last weekend was a very auspicious time for the Chinese to get married. According to one of my friends, 22th Sep is the most favourable day in this Chinese Lunar Year. No wonder I got two wedding invitations last weekend. Luckily, both are on different days, because both are my close friends.
This time, I have very great pressure, because both friends didn't want to hire pro photographers. I have not worked under such conditions, nor would I normally agree. I always advise my friends to engage one. I have the tools, some of the skills, but I cannot handle the responsibility.
Yes, it is a great responsibility, because you have to deliver. Due to this, most pros use conservative settings and pretty standard composition. I prefer to be more experimental and go for candid shots.
One thing most people don't realise is how expensive developing can be. Pros have an advantage here because they can get bulk discount. 8 rolls cost S$160, including film. You need to buy an album, decorate it, layout the photos and so on. Very labour intensive.
Then, there is the issue of ownership. I like to retain the negatives. This is why I don't want to be paid. So that I won't have any obligations. Many people asked me why. Let me just say that it is my work, so I'll like to retain it. I always assure my subjects that I will give them -- and only them -- unrestricted access to reprints, if this is an issue with them. Negatives are very delicate. I'll like to minimise damage to them due to reprints, storage and handling.
Ultimately, it is the results that count. I haven't reach the level where I can confidently say I can deliver. No one can. But you must bluff and you yourself must believe it. I can't. I cannot count the number of times I missed good moments because, "I wasn't ready"; "I didn't have the right lens"; "I don't feel like taking it"; "Let's wait for a better expression" etc. And then sometimes my equipment fails me. Out of focus, camera refuse to lock focus, setting the wrong exposure etc. And most of these are due to my style -- and I simply refuse to change my habits. (Yes, I am one stubborn pig.)
What's up with me, I wonder? Tired. Drained. Stressed. And just a moment ago I spilled coke over my notebook. The notebook was on and it continued to work. Thank goodness.
I just took some photos using my Minolta camera again, to do a final check on its readyness before the back-to-back photography workout this weekend. It has been more than one month since I used it.
When I saw the results, I was immediately blown away by its snappiness and how crisp it looks. The last few rolls I took using the Nikon camera do not have this quality.
Does this mean I do not know how to use the Nikon camera/lenses? That I haven't tapped its sweet spot? Or that I have expected too much of these old lenses? Or that I am just plain biased?
I had a bad dream yesterday morning. Not exactly a nightmare, because it happened in the few minutes just before I woke up. This means my brain is already awake, although my body is still sleeping.
There is a HP celebration in a concert hall. It is some kind of karaoke. There are several mics, all active. One person was singing. A large group of HP staff are there, including the head of the division. I was walking outside. Somehow, I was very stressed and went in, took a mic from the shocked person, tapped it a few times to disrupt the singing, threw the mic on the floor and stepped on it. I escaped in the resultant mayhem.
There is a reason that made me do it in the dream. I can't remember it now. I hid in a secluded place to ponder on the consequence of my action. I'm done for, I thought. I expect to see a letter when I return to my desk. I thought of apologising to the head, but will he accept it? The matter is made worse by the fact that the person I took the mic from was a guest from overseas.
Several colleagues passed by the area. News travel fast. They asked me what my plans are. As it were, I only have three choices: quit, wait for the letter, or apologise.
I woke up with a fast beating heart. What the hell was the dream about? I asked myself. Is it due to past grievance? Is it from my sub-conscious? Does it foretell that I might get in trouble? I realised the reason in the dream is the key, but I cannot remember it now!
Somehow, I managed to drop my notebook again. This time, I was gripping it on my right hand and opening my bag to put it in when I lost my grip. Down it went. I immediately used my leg to break the fall. To a certain extent I was successful -- it fell on my toes. Part of the notebook hit the ground hard. When I lost my grip, I instinctly knew, but was unable to obtain a good grip. I let go of a scream as I struggled in vain to grip it. A split millisecond later, I knew it was a goner and moved my leg out to break the fall.
It was painful. It is still painful, after one day. Perhaps this is why the hard disk survive. Luckily, the notebook fell on the LCD hinge, furtherest away from the hard disk. Right after rubbing my toes to ease the pain, I worry about the hard disk. How much data would I lose if it died? Thankfully I just did a major backup. Less than one week, I thought. I shook the notebook lightly to see if the hard disk is a goner. No loose parts. Good. (This is actually a bad thing to do, but I am known to rub salt onto my own wounds...)
Then I inspected the hinge. It gave way due to the impact. Since I have removed all screws from the cover, the LCD screen has separated slightly from the base. What luck, I thought. If it were screwed on, would it have withstood the shock? I snapped both ends back. Since the damage didn't seem severe, I didn't bother to try it -- I need to catch my bus to work. So, off to work.
In office, I opened the notebook. Marks on the screen. Turned out the keyboard became loose and shifted outwards slightly, so they pressed onto the screen. The keyboard was already loose because I have removed the restraints. Luckily, there is no permanent damage.
I was very lucky this time. I always fear one day I will damage the hard disk. The drop this time was from my knee level, and I had managed to grab the notebook before the other end hit the floor.
If you were to choose between money, intelligence and luck, which will you choose? I will pick luck every time, because the other two can be picked up.
Last Thursday we had someone from the Ministry of Law coming down and give a presentation. He is in charge of this Imagine Labs initivative, which he explained to be technological companies reaching out to primary school kids to instill in them the possibilities of the future. Apparently they already had a pilot run with Maxtor, and the students were to design the "hard disks of the future", with mentors and trips to manufactoring plants thrown in. Apparently he is looking for mentors.
Well, I am a bit cynical as to who would give up their valuable weekends to go coach these primary school students? Secondary school, I can understand. What will they learn? How much will they pick up? Also, he should asked for only Singaporeans to attend, since this is for Singapore. I bet the room will be half empty, if he said that. :-)
Also, the one thing that I would like to impress on everyone, whenever I have the chance, is the need to do tests and experiments. You need to determine what you want to test, how to test it, properly setup the experiment, conduct it and evaluate the data. Too few people do objective tests these days -- or know how to.
I won't pretend that all my tests are perfect, or even objective. But I do many tests to help me find a good tradeoff between time, effort and cost. Evaluating exposure, lenses, scanning, preparing an image for display, DVD quality, and most recently, printing. If there is one message I want to pass along, it is, don't assume and don't believe specs at face value. Always test and see for yourself.
After completing my first assignment last week, my project manager decided that since I was so good with doing pen alignment, I will be doing it for another device!
The idea is this. The new printer uses a different chipset and imaging pipeline. However, the hardware is the same as the one I am currently working on. The pen alignment must follow the hardware, so the existing one must be replaced. Since I understood the code fully, I am the best man to work on it.
Actually, I don't mind. When I was working with this module, I had many ideas how to improve it. However, I wisely made the bare minimum modifications -- no sense making more changes than neceessary. With the need to port the code to the new platform, it is a good opportunity to do a massive restructuring of the code. All the functionality will remain the same, of course, only the structure will be changed. The structure refers to the physical function placement, physical dependence and so on.
Personally, I have three goals for it. One, to use the same code for both platforms. I consider it a weakness that during porting, code is often duplicated. The new code is almost never merged back. And when it needed to be, additional effort is needed.
Two, to restructure the code to meet my own standard. My standard is very simple. Expose only interfaces to the outside world. Physical dependence and coupling must be kept at a minimum. Physical organisation must follow the logical organisation. Code should not be duplicated. Code should be kept simple and easy to read.
Three, to improve the tracing facility. Currently the printer will crash if the trace buffer is full, instead of gracefully continuing. Data is sent to the PC in ASCII. Using binary structures would be much more efficient. Also, the tracing is somewhat handicapped with its delayed invocation. It does not allow the same buffer to be reused.
Let's see what I can do.
Someone wrote, on a mailing list, that where photos are concerned, he doesn't talk about lenses just like artists don't talk about their brushes. I disagree. Another poster, in another place and time, posted that to get the best result, exposure accuracy is the most important, followed by tripod, then film processing, and lastly, camera and lenses. Again, I disagree.
Lenses are not the same as brushes. Cameras, maybe. Cameras do not affect the result if you control all the variables. A friend told me he just snap and hope the camera will take a good picture. While I am at a loss at how to encourage him, he will learn, one day.
Lenses is one of the three factors that determine the result. The other two are film and exposure. (Light is implicit, of course) As such, in a discussion about an image, I don't think these can be left out. At the least, we can see implicitly the photographer's skill. Did he did everything "wrong", yet got the shot? Or did he use state-of-the-art equipment but produce merely acceptable photos? What made him choose the film and exposure? Luck? Reliance on the meter? Experience?
Lens determine the outlook the photographer use to look at the world. Why a wide angle? Why not a telephoto and stand further back? Or vice-versa? Film can impart a certain mood. Some film are good for moody pictures; some realistic; and some just makes the world damn colorful. Exposure can be used to control DOF, as well as mood. For the latter, slides work best. Even a 1/2 f-stop bracket on either side can show very different mood.
When I look at a photo, I inevitably will ask for these three values. Usually the photographer will only know the film used. He will most likely use a zoom lens and only know "roughly" which focal length is used. Also, he may be using P or A. For A, he may remember the aperture, but he most likely will not remember the automatically selected shutter speed.
Not that this is bad. An estimate of the focal length is good enough. It is not important that you use 32mm, but a wide-angle focal length. However, in my experience, I have been misled by telephoto shots that look wide angle. As a result, it is a useful trick to know the situations where the usual focal lengths (as we know them) change their status. It is all a matter of subject distance, background distance, subject size and so on. When using a zoom, an estimate to the nearest 24-28-35-50-70 focal length will be good, but currently I can only tell 24, 28-35, 50-70 apart.
For exposure, it is critical to at least remember the aperture. This allows you to reproduce the approximate DOF look. This is what I usually record down in the field unless I am also trying to see if my metering is off -- which it often is. In this case, I write down the shutter speed too. I may or may not write down the lens used, depending whether I think I will be fooled or not.
To iterate, I believe the three factors, the lens, the film and the exposure, must never be ignored. Sure, if you use two same focal length lenses, one Nikkor 24/2.8 lens and the other Sigma 24/2.8 lens, there may not be much apparent difference in the photo. This is the usual reason why people discount lenses. But, they are still not the same! You need to do a controlled test to see the difference, only then can you reliably identify it in normal photos. It's better to know the limitations and weakness of your equipment upfront and work around it, than assume it to be as-good-as-the-real-one, and finally after years of using it, found that all your photos have been compromised in some way. Like digital artifacts, once you realise its presence, you will see it all the time. It gets on your nerves!
Today is my sister's birthday. It is also the 4th anniversary of the day I started working. Wow, it has been four years already. That's pretty quick. What have I accomplished so far? What are my short term and medium term goals?
Yesterday, I encountered a severe design flaw my team's development model. First, the background.
The source code is stored in a version control software. To make your changes, you need to "create task". This creates a new branch. You branch off a label, so if other people later checks into the main branch, you don't see it. I will return to this later.
The branching is done on a need-to basis. You can checkout the files, and later when you check them in, it is stored in your own private branch.
When you are done, you merge your files back to the main branch. To prevent concurrent updates, the script "prepare to findmerge" will attempt to checkout a file that functions as a mutex. If it fails, you cannot proceed. Suppose it succeeds, you happily merge the files back. Hopefully, the files have not been modified by other people since you branched off, or you need to do manual merging. Say, for a file readme.txt, if it has not been modified, your copy will be the latest copy. If it has been modified, you need to do a manual merge. This is extremely troublesome and time consuming. As far as possible, I always try to branch off the latest label, make my changes quickly and merge back.
After merging, you will be brought back to the main branch (via the filter rules). You are supposed to do one build, download the firmware and do a "smoke" test. After it succeeded, you do a "complete task". Basically this adds a new label (one that you supplied) to every file in the whole project and sends an email to every project member to notify them the files that have changed and a description of the changes (that you have to supply).
After this, you have to run additional scripts to generate helper files and HTML files. Finally, you need to upload these files to an NT server where the official releases are kept. Phew! What a lot of steps. For me, I have semi-automated the last few steps. I intend to automate it even further. The scripts prompt quite a number of times for the label. I'm sure it can be done away with.
Yesterday, I was unable to findmerge because the lock was taken. Argh! I split my work to two parts. Earlier in the morning, I had already merged it back to the main branch. I then continued with the other half. By the early afternoon, I was done. However, the lock was taken. I then wait. When the notification email came, I immediately tried to take the lock. Failed again! Argh, someone beat me to it.
Anyway, I looked at the list of changed files and saw several that I have to merge to my copy. This is what I hate the most. I need to retest them. There is a weakness in the Sirius architecture that some files need to be modified very often.
So, what to do but wait?
I dislike the idea of branching off a label. I feel we should see the latest main branch at all times. This allows us to get the latest checked-in files without any effort. As it is now, we have three options: ignore, merge from the main branch to get the latest changes, or create a new task based on the latest label and merge from our old task. All three options take too much effort.
Also, the entire process, especially the merging part, is too cumbersome and takes way too long to execute. This lets a person hog the main branch for too long. I'm sure the time can be shorten.
I have a 2820 dpi film scanner. However, I seldom use its full potential, due to the fact that it produces a 4032 x 2688 RRGGBB image at 2820 dpi in 16-bit mode to give a 65 MB file. Lossless LZW compression brings it down to 55-60 MB.
I am only considering purely resolution. I am not considering multi-pass, which increases scanning time significantly. I definitely want 16-bit, so my files will be all twice as large. Scanning in 8-bit now guarantees that you will need to rescan them in the future. So I do the right thing now.
Since scanning at full-res takes too much space, how about half-res? It gives a 2016 x 1344 image, or 15.5 MB file. The file size is very acceptable. The trade off? The image is just big enough to print at 8 x 12" (at 167 dpi). Quarter-res image, 1008 x 672 (a mere 3.88 MB), is slightly soft and loses some fine details. It is still good for web display, of course.
It is at times like these I wish scanner optics are more flexible. I wish a scanner allows us to choose any dpi we want, not just divisors of the full resolution. Now, I don't mean interpolated resolution, but real resolution by shifting the lens. If a continuous resolution isn't possible, a small set of resolution is fine too.
The Nikon LS-4000 (and LS-8000) and the Polaroid ScanSprint 4000 are 4000 dpi scanners. A half-res resolution of 2000 dpi is very attracive. Minolta Scan-Multi Pro is 4800 dpi. At its full resolution, it can produce a 185 MB image! A half-res image will be almost 92 MB at 2400 dpi. I can only fit 6 images on a single CDR!
The Minolta Scan-Multi Pro is already as good as film scanners can be -- in my opinion. 4800 dpi is a good comppromise between scanning details and file size. Do you really want 6000 dpi (290 MB)? However, the Scan-Multi Pro can be improved by offering true variable scanning resolution between 3200 and 4800 dpi. Currently, it supports true 3200 dpi for medium format. It can be used for 35mm too. It can support 3600, 4000 and 4400 dpi, for example.
How do you cool your notebook. In the tropics, notebooks can get very hot, especially if you use them in non-air con places. And that's why I am most of the time. Therefore, keeping a notebook cool is always on my mind.
First, a story. My sister bought a notebook for herself too. Less than 3 months later, she called me and told me the computer died -- it refused to even power on. Now, this is very serious. A notebook has only one power button. If it doesn't work, it is usually a hardware problem. Aftering asking a few routine questions, I asked her what was she doing when it died. She said she was playing CDs when the notebook suddenly switched off. Huh? That sounded strange. Suddenly, a thought came to me and I asked her where did she put it. In the bag, she said!
Groan, this is the worst thing anyone can do. Notebooks cool from bottom -- actually, I found it out the hot way ttoo, so if you block it, it will get hoter and hoter..
My notebook, the Acer TravelMate 611TXV -- three generations old -- does not have even thermal distribution. This is to be expected. The front right, where the hard disk is, feels just warm, to my surprise. Hard disk technology must have really progressed, or there must be a reason why 2.5" HD cost so much, or it could be spinning at 4500 rpm (not even 5400 rpm). I don't know. I seldom do disk intensive stuff, so the hard disk rarely gets warm.
The right back (towards the middle) is where the CPU and several chips, all with heat sinks, reside. I suppose these chips are the chipset and video controller. This area is always the first to get hot. There is a fan beneath, and an outtake vent at the back, but they don't seem to help much. On the left back is the DVD drive. It is cool all the time, until you play any disc for long period of time. Then it becomes really hot. I always use a software to reduce the speed of the drive. The drive supports 5x, 16x, 20x and 24x. When I am not in a hurry, I choose 5x. It is both cool and quiet. Unfortunately, it works only for CDs. DVDs run at certain speed. I don't know whether it is 1x or 8x.
If I do intensive stuff for a period of time, like watching DVDs or playing games, the whole back gets hot. I avoid doing these unless I have active cooling.
Due to the fan's placement, I always try to elevate the notebook in non-air con places. In my home, I place the legs of the notebook on 8 filter boxes, 2 on each leg. In my JB home, I use two video tapes to stack up the notebook. They help a little, because the fan can suck in more air quickly. For safety, I use a fan to blow at the notebook. The elevation increases the space below the notebook, so the fan can blow through it more effectively. Using a fan really helps.
Mooncake festival is just around the corner. This year, my mother decided to make the mooncakes herself. It's much cheaper, she declared. Well, I asked and found that it costed RM50 per session. Will take a number of years to recoup it back, I think. :-)
Anyway, I am an advocate of do-it-yourself.
I found a new lens hood, the HS-10, for my 85/2 lens. However, it is matte black instead of the original glossy black. The latter complements the lens. I have no idea how the first will look on the lens.
The new lens hood costs S$30. The old one, should I choose to buy it back, is S$15. Which should I buy? I am comtemplating buying both. I like the glossy finish because it goes together with the lens. As a classic lens, I think it deserved at least this fate. However, it may not be that easy to use on-the-road, hence I am looking at the new lens hood. It is definitely easier to use with the clip in place.
I doubt anyone will buy off 3 HS-10 within one week, so I am taking my own sweet time to decide. I can tell you, the longer I take to decide, the higher the chance of me buying both! <grin>
A revolution is underway. Since the mid-eighties, zoom lenses have begun to replace prime ones. Lens makers have proven skeptics (I'm one of them) wrong time and time again. These days, the highend zoom lenses are large, heavy, have a constant f/2.8 aperture, and a performance that will rival prime lenses, especially stopped down. Other zoom lenses are variable aperture, but smaller and lighter. They are more compromised: wide open is often too soft, have too much distortion, and they get softer at the long end. Actually, there is no reason why they must be so, but achieving good performance wide open might cost too much. Besides, photographers are used to stopping down for better performance.
You won't find many constant aperture consumer zoom lenses these days, even if they are f/4. Minolta used to have a 35-70/4, 70-210/4, but they are discontinued. Minolta still makes a 24-50/4 lens. Canon makes a 70-210/4L lens, but this is not a consumer zoom lens. Perhaps lens makers feel that f/4 lenses will lure potential buyers away from the much more expensive f/2.8 lens. But today, third party f/2.8 zoom lenses are available for about half the price. I always crinch when I hear someone opting for such a lens. It is my opinion that the f/2.8 zoom lenses represent the lens maker's state-of-the-art capability. To skip it is foolish.
Anyway, back to f/4 lenses. As a way of comparison, a very good condition 70-210/4 lens cost S$300 to S$350, second hand. The 80-200/2.8 G lens cost S$2300 new. Will you pay this difference for what seems to be just one f-stop difference? Many people who won't. In any case, I have the 70-210/4 lens and I can tell you it is not so hot at f/4, especially at 210mm.
Canon used USM as early as 1989. Nikon finally followed with its AFS in late 1990s (1998?). Finally, Minolta will release its SSM lenses soon. Canon has also released IS lenses. Nikon called it VR. Minolta has nothing equivalent. All these are old news. Canon used DO in one of their lenses, the 400/4 lens. It is remarkably shorter than the 300/2.8 lens, but wider. And it isn't cheap at all.
But something big is brewing. Canon may be removing AF rings from its lower end lenses. Nikon is taking another step to breaking its backward compatibility by releasing lenses without the aperture ring. And also, zoom lenses may have already outperform prime lenses in several area.
Despite the conventional thought that a prime lens must be better than a zoom lens, due to their simplicity and lack of constraints, prime lenses are full of compromises too. First, most of them are old designs. New zoom lenses use exotic glass to achieve never-before performance. Second, they may be unoptimised. What do I mean? Usually, prime lenses are designed to have very little distortion. Zoom lenses, on the other hand, are designed with distortion in mind -- after all, it is usual to get barrel distortion at the short end and pincushioning at the long end. This frees zoom lens to concentrate more on sharpness. For prime lenses, due to size constraint, and the desire to keep things simple -- most primes are straightforward designs with few glasses -- there may be more light falloff, may have poorer sharpness wide open, and may have poorer edge sharpness.
One thing that always surprise me is that, for f/2.8 zoom lenses, there are as many as 20 (or more!) elements in 18 or so groups. Given so many air-to-glass surfaces, how is the contrast kept high? It is a fact that the more glasses there is, the lower the contrast (again, a law of physics). But these f/2.8 lenses don't show it!
The kit lenses are also undergoing a transformation. The first kit lenses are of course the 50mm lenses (the slower kind: f/1.7, f/1.8 or f/2 -- there are no slower 50mm lenses, other than macro), but you haven't seen these bundled as kit lenses since AF cameras came along. The first generation of kit zoom lenses are the 35-70 lenses. Minolta's first kit lens, the 35-70/4 lens, has a very fine reputation. However, I haven't seen it for myself. I couldn't bring myself to buy it (secondhand at S$115) due to its 49mm filter size and lack of lens hood (secondhand lenses rarely have lens hood). For what it is worth, it has an aspherical element! Of course this must have proven too incompetitive, because later Minolta brought out the more usual 35-70/3.5-4.5 lens. Since then, the battle has always been larger zoom range. Naturally, aperture gets smaller and smaller.
These days, the standard kit zoom lens is the 28-80/3.5-5.6 (or f/4-5.6). Or, was. Canon brought out a new 28-90/4-5.6 lens together with the EOS 300. It is slower than its predecessor, the 28-80/3.5-5.6, but who is going to worry about a half-stop difference? Besides, the people who buy this lens may not even know what the number means. Recently Nikon brought out a 28-100/3.5-5.6G lens. I wonder how Minolta will react.
In the past, one step above the kit lenses, you will find the standard zoom lenses, the 28-105/3.5-4.5. Previously they were only 28-85/3.5-4.5, but like all things, they were made longer. Now, they are made even longer, like 24-105/3.5-4.5, 24-120/3.5-5.6, 28-135/3.5-5.6 etc. The shorter 28-85 lenses are long discontinued. The 28-105 lenses may go that way too, since they are being squeezed on both fronts. Now that the kit lenses have the focal lengths as the standard zooms, who is still going to upgrade to them for just half-a-stop difference? As a comparison, the 28-80/3.5-5.6 lens costs about S$280 (but who buys it separately?), and a standard zoom 28-105/3.5-4.5 lens costs about S$520.
If you are lens-savy, you will realise the flaw in my logic. The difference between these lenses are not merely one-stop (for f/2.8 vs f/4), or even half-a-stop (f/4.5 vs f/5.6). There is much less compromise, the optical performance is much better (especially wide open), and the built quality is much better. But does the consumer know this? All he sees is the specs.
The telephotos are the same. It started with 80-200/4.5-5.6 lenses. Then it slowly widen to 70-210/4.5-5.6, then split into three classes: 70-210/4.5-5.6, 75-300/4.5-5.6 and 100-300/4.5-5.6. The first is usually a a companion to the standard kit zoom lens. It is very cheap. The second is a standard telephoto, only slightly better. The third is usually much better due to APO glass element. However, it can be up to twice as expensive as the second one! All these overlapping lenses do nothing to reduce consumer confusion. Who would pay so much more to get the 100-300 lens when you can pay less and get more zoom range? Surely there can't be too much difference, right? Also, the 100-300 lens is priced such that a third party f/2.8 zoom lens becomes an alternative, and is more attractive due to its f/2.8 constant aperture. Such lenses, as we know, must be better.
Currently Nikon has consolidated its lenses. It now only has two lenses in this range: the 70-300/4-5.6G and the 70-300/4-5.6D ED lens. It used to have a 70-240 lens.
Canon has a number of lenses in this range, due to IS. It has two 75-300/4.5-5.6 lenses, one with IS. It has two 100-300 lenses: one f/4.5-5.6 and the other a constant f/5.6. The latter lens is an L lens. However, it doesn't fit into the typical profile of an L lens. Its build quality is second-rate, but obviously optics must be first class. But L or not, f/5.6 is too slow at 100mm, even at 200mm. (Unlike the other lenses, this lens should give good results wide open.) Canon also has a 80-200/4.5-5.6, but obviously it is targeted at the very lowend.
With more digital SLRs on the horizon, watch out for even more new lenses. Already, Sigma has already cashed on the digital frontier with its so-called optimised-for-digital DG lenses. What it really means is that there is severe light falloff and poor edge sharpness wide open. This is because current digital sensors are much less than full-frame -- they are even smaller than the APS sizze. Don't be fooled!
Some of the lenses will replace the old ones totally. Digital sensors require more stringent optical design, due to its high contrast and narrow light acceptance angle. Also watch for never-before odd focal lengths, due to the digital sensor multiplier ratio. Already, Canon has a 22-55mm lens, although that was targeted at its APS SLR. Digital is rewriting the rules of photography, even for lenses. Soon, the lens lineup may bear little resemblence to what we see today!
When you look at a lens, don't just look at the optical performance. The physical construction matters too. But, other than the build quality, do you know what else to look for? Let's go through the items one by one.
Build quality. Plastic vs metal. The age old debate. Older lenses are very heavy due to the use of metal. Newer lenses are far lighter. A closely related issue is the finishing. Lens makers like to differentiate their top-of-the-line lenses with the other lenses. It's quite easy to tell them apart.
What makes an 'L' or 'G' lens is (almost) never stated explicitly by the lens maker, but it is definitely not just the finishing. Minolta states that for a 'G' lens, exotic glasses (aspherical, APO etc) are used, or they have better optical design, like CRC (Close Range Correction) and IF (Internal Focusing).
To me, it means the optical performance at wide apertures is good enough for use. You often have to stop down one or two stops for consumer lenses to reach the sweet spot.
Zoom Ring. Since most people use zoom lenses, presumably for zooming, the zoom ring is very important. Look at the placement, the feel and the rotation.
Focusing Ring. Ever wonder why you don't use manual focus much? The focusing ring could be the cause. Look at the placement, the feel, and the focus throw -- the amount of movement from the nearing focusing distance to infinity. Try a MF lens some time. The ring will be wider, better damped, and the focus throw will be very long. Slower to focus, but it will be more precise.
Canon is eliminating the focusing ring from its low end lenses. Can you imagine this? Well, Canon lenses and cameras are never good for manual focusing anyway..
Due to Canon's USM feature, some of their lenses offer FTM (Full-Time Manual). You can manually focus whenever you like. Pretty nice. Minolta's 'D' lenses have a unique feature too. The focusing ring do not rotate in AF mode -- automatically -- on all its cameras.
Aperture Ring. Canon and Minolta lenses do not have aperture rings. Only Nikon (and other brands who preserve backward compatibility with their MF lenses) does. As such, this is irrelevant to most of us. However, Nikon is doing away with this on its new lenses. The G series do not have the aperture ring. Right now they are all low end lenses, but Nikon is coming out with a 70-200/2.8 AFS VR G. Sign of the times.
Distance scale. This should be on all the lenses, but most low end lenses don't have them. Even some 50/1.8 lens don't have them. Perhaps this is used to differentiate lenses. Kit zoom lenses -- the ones that come with the lowend cameras -- usually do not have this. But presumably they aren't used either, so perhaps this is why Canon decided to leave it out. However, a distance scale is invaluable on a prime lens, together with the following feature.
DOF scale. A DOF scale can be very useful when you want stop down as much as possible for maximum DOF. Rather than carrying a DOF table, why not let the lens tell you? Unfortunately, most zooms don't have DOF scales. This is because DOF varies with focal length, so most zooms have scales on the shorter end only. Only push-pull zoom lenses can have a beautiful DOF scale.
DOF scales are useful for shorter lenses only. Above 50mm, you will quickly notice that even f/32 doesn't give you much DOF. So, a lack of DOF scale on telephotos is no big deal. However, some lowerend prime lenses don't have DOF scales, limiting your full potential. A 50/1.8 lens is often the cheapest prime lens -- a good way to start your lens lineup. It is small, light, fast and sharp. However, the lens maker often cripple this lens in terms of physical construction. Capitalism rules.
Filter Size. As I said before, lenses are getting better. Already, Nikon and Canon have made their f/2.8 zoom lenses 77mm. Minolta still uses 72mm for its 28-70/2.8 and 80-200/2.8 lenses. However, Minolta is coming out with its SSM lenses, and the question to ask is, will they be 72mm or 77mm? Personally, I hope they will be 72mm. You can buy a step-up ring if you want to share with your other 77mm lenses. This means you can't use the lens hood, but Minolta can come out with two lens hoods -- one for 72mm and one for 77mm.. :-)
Rotating front. People who use polarisers a lot always look for front that don't rotate. But even people who don't use polarisers (at all) often look for this, I wonder why? The front may rotate during zooming, or more likely, rotate during focusing. You are more likely to find rotating front on older zoom lenses. All (all?) prime lens have non-rotating fronts. These days, even if the lens does not have a true non-rotating front, a shell covers it up. Can you say, bigger lenses?
As an example where a lens receives unfavourable comments due to its rotating front: the Minolta 80-200/2.8 G lens. However, it was pointed out in the MML (Minolta Mailing List) that this lens is actually very well constructed. First, it has 72mm filter thread, which is a feat, considering that a true 200mm at f/2.8 requires a minimum front diameter of 71.4mm (200mm / 2.8; a law of physics). Now, you must know it is common to cheat at focal lengths. At the wide end, there is usually 1-2mm tolerance. At the tele-end, there can be up to 10-20mm allowance!
So, is this a true 200mm lens? Yes, according to the poster, it is. Compare this to other lenses, they are all 77mm filter size. It is impressive. But, all 200mm prime lenses use 72mm also, so what's the big deal? Remember the cheating? Most 200mm lenses are in fact only 190+ mm, including Minolta's own 200/2.8 lens. Now you know why Nikon only has a 180/2.8, and not a 200/2.8 lens.
IR Mark. In case you use IR film, which has a shifted focusing point, this will let you know how much to adjust. Sadly, most lenses do have this mark anymore. And, most cameras will fog IR films due to the use of IR sensor for more accurate film positioning. The Minolta Dynax 7, for example, uses one to do fanciful mid-roll rewind -- you can forward and reverse to any shot you want, as many times as you like.
If you look at lenses long enough, you will realise that all lenses of the same generation have the same cosmetic design. So far, Minolta has changed its cosmetic design three times. Lenses are discontinued and new ones are introduced all the time. But those that are carried forward, the optical design often remains the same. If you are only after optical performance, an earlier lens is often cheaper.
For a lens lover, there is nothing like a well-stocked secondhand camera shop that lets you see a wide variety of lenses. In a normal camera shop, you may see lots of boxes of lenses, but usually lenses are not put on display. Cameras are. They are what sells. In a secondhand shop, however, cameras and lenses are treated equally.
Now, you can see the lens on the web, or you can even request for the lens catalogue, where they will show you all the current lenses. There is even a table of detailed specification. However, there is a difference when you see the lens in person.
Two lenses of the same focal length and maximum aperture can be very different. Take for example, Canon makes a 50mm f/1.0 lens. You read that right, a f/1 lens. It's incredibly huge: takes 72mm filters and weighs 985 grams. Contrast this to Canon's own 50mm f/1.4 lens: uses 58mm filter and weighs just 290 grams. And there is only one f-stop difference!
Canon has renounced their MF lens lineup (since 1987?), and thus has introduced all sorts of exotic lenses, not to mention using advanced technologies like USM, IS and DO. Some of the exotic lenses are really exotic. After all, it has to catch up with three decades of MF lenses. Some are just plain ultra-fast or ultra-wideangles. Today, Canon has the widest AF lens lineup. It has the fastest, the widest, and is the most flexible.
This is also my main grip with it. As a speed fanatic, I am attracted to Canon's ultra fast lenses. No, I am not interested in the 50/1 and 85/1.2 -- another of their speed demon. Note thaat their AF is anything but fast, and quality is same or worse off than their slower counterparts. The lenses that draw me to Canon are, the 24/1.4L, the 35/1.4L and the 135/2L. To a certain extent, the 200/1.8L too. But I will never be able to afford this lens. All these lenses are oversized. In other words, they are very big compared to the same lens, but one (or two) stop slower.
It is acceptable for the telephotos, since everyone expects them to be large. But it is simply unacceptable for short prime lenses. But first, a short word on the long lenses. You should see for yourself sometime how compact are the 85/2, 100/2 and 135/2.8 lenses. In fact, all three have about the same size. Especially for the 135/2.8 lens, no one will ever suspect that such a small lens is able to pull such long distance. All have filter size less than 60mm.
Fast telephotos have to be large. There is no loophole in the law of physics. The 85/1.4 lens is large and heavy. Everyone who makes one uses a 72mm filter. The fastest 100mm lens is, if I am correct, the Nikkor 100mm f/1.8 AIS. It uses a 62mm filter. All 135/2 lenses have a 72mm filter. So does the slow 200/2.8 lens.
Seeing the lenses and their relative sizes may persaude you to compromise with the smaller lens, with virtually no penalty or loss in performance. And comparing between lens makers let you see what is possible. Did someone take a shortcut somewhere? Is it possible to be faster? Smaller? Cheaper? Better? And what are the compromises each make, in terms of optical quality and physical construction?
And this is where Canon falls short. The Canon 24/1.4L and 35/1.4L lenses are huge; they take 72mm filters. Since Minolta has a 35/1.4G lens, I cannot help but wonder why Minolta requires only a 55mm filter size. Now, Canon uses 'L' to denote its highend lenses. You pay top dollar, but you are assured of top performance in every situation. Minolta uses 'G' to denote its highend lenses. (It is unfortunate that Nikon decided to use 'G' to denote its aperture-ring-less lenses. Currently they are all low-end lenses. I foresee confusion.)
Sure, all lens makers try to standardise on filter sizes. Minolta uses mainly 55mm, followed by 72mm, with a few odd-ball 49mm (lower end lenses), 62mm (mainly 3x zooms) and 77mm (ultra-wide zooms) sizes. Very well planned. Nikon has an MF lens lineup from 20mm to 200mm using 52mm filter size. Very impressive. These days, however, they have a good mix of 52mm, 62mm, 72mm, 77mm and some using 67mm. Canon uses mainly 58mm, with some 72mm and 77mm, and a few oddball 52mm and 67mm sizes.
I am sure with Canon can fit the 35/1.4 lens with a 58mm filter size. As for the 24/1.4 lens, Minolta only has a 24/2.8 lens, so I have no idea how big it should be. However, recently I read up on the MF Nikkor lenses, and there is a Nikkor 24/2 AIS lens. And it takes just a 52mm filter! Surely the jump from 24/2 to 24/1.4 isn't that much?
Finally, I saw a Leica 50mm f/1 lens in a secondhand shop. I never noticed before, because I never gave MF lenses much thought. It is priced impressively (~S$2750), but it is no bigger than a 50/1.4 lens! Finally, I can confirm my long time suspicion that Canon lenses are way too big.
There is nothing like oversized prime lenses that put you off them. The f/2.8 zooms are big and heavy, but everyone knows this is the price you have to pay for high optical performance. However, a prime lens can be very compact and light. But, do you know that?
There is a trend towards bigger glasses. Third party lenses are typically larger. The simple reason is it is easier to design and achieve a better optical performance. It will be more expensive to manufacture. This trend is not necessarily good. Many f/2.8 zooms are using 77mm filters, presumably because the ultra-wide 17-35/2.8 lens requires it (even Minolta's 17-35/3.5 uses 77mm). Although it standardises on one filter size, the cost of 77mm filter is astronomical. A 77mm slim UV filter can cost S$130, compared to a 55mm version at S$40.
It is a common fallacy that anyone can do programming. Sure, if programming means writing code. But what about the quality and how does it measure up to metrics like correctness, robustness, scalability, testability and maintainbility. I haven't mentioned elegance..
(to be continued)
I must confess, I lack much domain knowledge in my new field of work. To compound it, I have no EE background -- much of my work deals with EE stuff.
Today, I downloaded my new set of code to my device, powered it up.. and immediately crashed. Panic. I reassured myself that I have no touched the startup code; that the new code will not be called by the startup; have compiled the right files, etc.
Finally, I suspected it was hardware related (the error code indicated a h/w related error) and emailed my team lead. Thankfully he was free and came over to take a look. He is well-versed in the mechanism and the operations and immediately knew how to bypass it. Quickly, he found the cause of the problem: clogged pens. And it may have short circuited the board. And so it was solved, in a matter of 5 minutes. And why did the pens clogged? The pen caps provided the answer: they were full of ink. Immediately, my team lead asked me, "did you uncap the pens before you did your operations?" Even before he asked, I already knew I had forgotten to uncap the pens -- it just skipped my mind. So, mystery ssolved, he went back to his cubicle and I spent the next half an hour to soak up the ink.
This simple episode brings me down from my lofty tower. All along, looking through the code, I saw many areas that can be improved upon. My conviction that EE-trained Engineers should not do programming grew stronger by the day. But today, I can also say, software programmers should not do hardware stuff. But this simply isn't true -- I know I can do it, I was just careless. Therefore, I must be less critical of EE people. Programming isn't their primary training. Therefore, it make sense that they are weaker in it. As a software person, I resolve to show them the light!
If one day you happen to come across a Crick LP0 printer with ink-filled pen caps, you can be almost certain that it is the one described here. :-)
Imagine you are in a queue. Someone blatantly cuts into the front. What will you do? Or someone asks you to help him buy a ticket. Would you help him?
In the second case, no matter the sob story, sorry, it's no go for me. Nope, not even if you are a pretty gal. How about the first case? Sad to say, so far I haven't spoken out. Sure, I can give many reasons (read: excuses) why I do not do it. But in the end, it comes down to that the silent majority suffers. When cheating prevails, like at the bus 170 queue at the Malaysian Checkpoint, everyone gets into the act. After all, why let another person take advantage? There is nothing to lose. Sometimes, you see two or three buses going and the queue just moved a tiny weeny bit. Bottled anger starts to surface. You must give these people points how they attempt to beat the queue. They are very creative. But it is still wrong.
One thing I learnt is that people cannot be depended to do the right thing, by themselves. There are all sorts of people out there, and you are bound to see people who are less scrupled than you are. However, whenever there is a precedence, whether more people will join in the act depends on the majority. If they voice out, people will be deterred. If they keep quiet, more people will be tempted to take the shortcut.
Just yesterday, I was at the Malaysia checkpoint again, this time in a car. Of course there was a queue. The road is wide enough for three cars, but there is only one queue. The reason is because the road converge ahead to just one car length. Since that is the bottleneck, there remains only one queue for the wider road.
I'm sure you know what happened next. A car sped past to the front of the queue. This is not the first time this happens. Previously, many cars honked, effectively detering other would-be tryers. This time, not a whimper from any car. Since the car cut in just a few cars before me, I really felt like going out and tell him, "Excuse me, there is a queue." As they say, thought is the first step towards realisation.
It is a fact of life that there will be poor people. Actually, why? I don't understand. Even, or especially, I should say, in Singapore, I see the equivalent of beggers. People selling sweets, tissue papers or toys at walkways, people hawking their worthless goods at food courts and coffee shops, and are rewarded by being ignored by city dwellers too busy chasing their dollars. Yes, worthless is a strong word, but that's what they are really selling. Now, I don't consider selling food a waste of time. They may be unlicensed, but as long as they are hygenic and ethical, I have no problems with them. After all, they are selling useful stuff.
Can their lot of life be improved? Are they willing to work for it? I cannot believe that there is no job they are unable to take up. But, should we wait for people to help themselves? Should we not provide the motivation and encouragement for them too?
Over the past few weeks, I scourged the camera shops in search of my next lens. A few lenses are on top of my mind, but I always wondered if I would buy the the lens on top of my list, the Nikkor 85/2 AIS lens, on the spot. Today, I have the chance to find out the answer.
Browsing through one shop, I saw the lens. Could it be the same poor condition 85/2 lens I skipped previously? May be, may not be. This lens has a lens hood! What a rare find! I immediately went in to enquire more about the lens. Good condition. Quite dusty, with a few big spots. Aperture ring is smooth -- too smooth. Once bitten, always check twice. Focusing ring is way too stiff. Filter thread, mount and rear glass looks good. Aperture blades have some sort of mark.
I have some difficulty with the lens hood. I seemed to have some difficulty mounting it, or storing it in the reverse direction. Finally, the owner said the lens hood has a broken clip. Argh. Well, I thought I can get it new, so I told him I don't want it. He gave me a S$15 discount. Jolly good of him. In the end I got the lens for S$320.
Then when I went over to Cathay Photo to enquire, the salesperson immediately knew it was for the 85/2 lens. HS-10? No stock. Estimated price S$38. Estimated time needed: 2 to 3 months, if they have stock. Hmm, not too promising.
I have three choices. First, I can buy back the spoilt lens hood -- I can imagine the shop owner grinning. Second, I can wait and see if the lens hood is available new. Third, I can forget about it and get a compatible hood. The last option is the most practical and will save me a lot of grief. However, I am a perfectionist and when I know I can do better, I will. The 85/2 lens is a classic and to give its dedicated lens hood a miss is unforgivable, the collector in me says.
Well, the story doesn't end there. At home, I found the big dust specks are not dust specks, but fungus! Argh! What to do? Looks like a trip to the Nikon Service Center is imminent..
I always emphasize on efficiency. It is no exception when it comes to dial-up usage. I always dial in, load all my pages, disconnect, read them, prepare all links for the next batch, then log on again. Why do I have to do this? Because of pay-per-minute charging. I believe in maximising the bandwidth usage when I am online. Of course, I am not paranoid to the full extent, but this is my guiding principle.