Fighting with computers

Computers are not always friendly.

Wednesday, January 24, 2007

Digital camera madness


I recently tried to learn a bit more about photography (I can recomend you this book). I've owned several film-based cameras over the years (both compact and reflex) and since 1996 when I bought a Casio QV-10 (not that I can recommend that battery eater) all of them have been digital units (Canon S10 and A520 and lately Panasonic DMC-FZ7, all of which I can recomend).

I was thinking to finally jump to the Digital Reflex bandwagon but I was waiting to get a better idea of that segment. But it seems the waters are not that clear. Although it seems that Nikon has settled with the so-called APS-C sensor size, Canon has mixed feelings and now it has three different sizes from full-size 35 mm sensor (like the one on EOS 5D) to APS-C size (like the one on 20D, 30D or 350D).

Why sensor size is important? Well, one of the reasons professional photographers have kept the reflex market going was compatibility. You cannot be compatible if you change the sensor size. And, it does not matter you can still use your old lenses (in case you have any) from your old film reflex camera. The results you get are just not the same.

Here is when the idea of multiplier come into play. Think on a camera focusing a person's face fully. All that image is projected by the lenses on the film or sensor surface. Now think on reducing the area of the sensor. What happens? Now the image only show the nose and maybe the mouth of that face.

What it looks like? It seems now this new sensor is enlarging the image. So it looks like you are getting a longer lens. It looks like you are multipling the lens focal length by a (larger than one) multiplier. So, reducing the sensor translates into an increase of the focal distance (if you keep the same lens).

But why is this important? Well, reflex camera users have created their lenses collection by carefully selecting the different focal lengths to achieve the different types of shots they (or their paying customers) have come to love (i.e. portraits).

The bottom line: No matter the new digital reflex cameras promise you can use your old (and expensive) lenses, the end result will not be the same unless you are getting an expensive camera body with a full-size 35mm sensor (and I am afraid the cheaper one is Canon 5D at around $2500).

I am not saying that going reflex is a bad idea, but don't do if for the wrong reason. At the moment it seems that either Nikon D80, Sony A100 or Canon 400D are solid performers for under $1000 with access to a good set of lenses (all of it with APS-C sensor size, or a 1.6 multiplier).

An interesting alternative is the so-called Four-thrids spec, backed mostly by Olympus that offers a set of reflex cameras with slightly smaller [squared] sensors and a multiplier of 2. These cameras and their lenses can be smaller and lighter but lenses can still be exchanged easily.

Another amazing discovery has been that the multiplier has an effect not only of the focal distance but on the depth of field too. To make a long story short let's say that the multiplier also increases the depth of field by the same amount, which is mostly a welcome effect. However, it also makes more difficult to achieve the, sometimes desirable, blur of the background. This is even more noticeable with non-reflex cameras equipped with smaller sensors where shots have an amazing wide depth of field at almost all focal distances.

I think for the moment I will keep on reading before buying the wrong camera (though my guts are telling me get the Canon 400D). At any rate, please note I have not mentioned at all the word "megapixel", isn't it curious?

Friday, January 12, 2007

File Uploads

Every now and then we all face a common problem poorly solved by today's applications: You want to send a big file to another person (or you want them to download it). For files several megabytes long (or larger) email is not the right tool (as email servers will reject send attempts larger than a few megabytes).

IM applications like MSN Messenger or Jabber might solve the problem, but only if you and your friend have that software installed and have a user account. But there is still another catch: This transfer happens between peers, so both you and your friend need to be online during the whole transfer.

To fix this, some applications like Pando can help, as they provide a multiplatform solution that does not require you both to be online while the transfer takes place. However, you both still need to download and install Pando client.

Upload centers allow you to upload files to be shared with the rest of the world. No user registration is required for some of them but some limits on the file size are enforced. At any rate, maybe you only want to share your file with a few people and these upload centers offer poor or no access control at all.

The type of solution I am thinking about involves a web server and some glue code on it. The idea is quite simple:
  1. You use your browser to upload the file to a server. (The fancy version may include an applet or flash app to offer a progress bar and maybe drag and drop support).
  2. You obtain a random and difficult to write and difficult to remember URL, similar to the ones Google Docs uses for publishing spreadsheets. (This ensures a casual user will not discover any document URL by chance).
  3. You send an email to your friend containing that weird URL.
  4. Your friend clicks on that URL to retrieve the file (again using just a regular browser).
  5. Files are kept for a limitted time on the server (to insure not a lot of storage is required and to lower the probability a curious attacker could get a file not intended for him by just guessing right a URL).
No new software to install on either the sender or the receiver is required. No user account needs to be created. And of course, no clear business model behind to make a commercial success out of it (you still can publish some adds on the system web pages).

A somehow similar system, but with user accounts is called File-O-matic and it is being used by Nova Scotia canadian authorities.

So I have some work to do over the weekend ...

Update: A bit late but here you have a sample (up to 2MB)

Sunday, January 07, 2007

Happy New Wii-ear


It turns out that Santa was quite computer- oriented this year, so we've got a Nintendo's Wii (it seems many people had trouble getting one this season) and a Creative Zen Vision:M.

I knew nothing about the Wii but it's silly name and how much my children wanted to get one. It turns out it has integrated WiFi and even an Opera browser (as the sample figure shows a zoomed in page). Who wants to have another browser? But ... wait, this one can even deal with flash contents ... never wanted to get those videos from YouTube on your TV set? Now you can! (provided you have a wifi network nearby).

Regarding Creative Zen Vision:M I already mentioned on a previous post: It seems to be a better deal if you plan on watching videos on it as the screen is brighter than iPod's. The white unit looks cool and quality made. The weird thing is that the four backlighted buttons require you to actually press them but when you want to click on the central scroll area no presure is required (nor a click will be felt). Scrolling through the music or video files is carried out quite nicely with the vertical scroll pad.

It is a pity that neither Creative nor Apple recognized that customers might want to attach Creative players to Apple computers. I want to thank Richard Low for his superb job on a software solution to this problem. And to me a much easier solution than the iTunes software (yes I do have an iPod too but I am not buying any DRM'ed music).

Another year, another (larger) hard drive for my kid's computer. Once again I want to rearrange partitions. This time I used the HDClone 3.1 without noticing the free version does not seem to offer an auto-expand feauture, so I ended up (10 hours later) with the same partitioning the old drive had but on a larger one. I fixed the problem with a LiveCD of GParted (as Knoppix 5.0 QTPart did not work this time on one NTFS partition). Fortunately GParted nicely expanded the size of the C drive from 60 to 150 GB in a minute or so. I really recomend you this tool for your partition table plumbing work.

Monday, January 01, 2007

HDR photographs


HDR stands for High Dynamic Range and although it is a concept more than ten years old it is new for me.

Several processes are trying to mimic the way we see things: moving pictures, photographic prints, computer displays, etc. The underlying idea is that an accurate representation is achieved when "it looks like the real thing".

Unfortunately, neither prints nor displays can exhibit the dynamic range required for a picture to look the same as the original. While our eyes have a dynamic range of 10000 light levels, a print can only represent around 100 different light levels. On the other hand, the whole range our eyes can see is around a million different levels, just not all at a time (i.e. you need your vision to adapt when you go from a dark place to a sunny one).

While photographic cameras (either chemical or digital ones) can do a great job capturing a scene in one shot, some compromises are done down the line. The end result is that some saturation and some range compression happens in every shot.

If you want to keep good detail of a dark area you select the proper exposure values but then the areas with the more light will appear burned. If you do the opposite then the darker area detail is lost but the burning has been avoided.

If one exposure value is not enough, then use more!! This is the idea behind HDR. You can achieve a higher dynamic range by taking several shots of a single scene using a varying exposure (usually changing the shutter speed). Later, you can combine these several shots into a single image with much higher dynamic range.

Sounds good, but there is one caveat: you cannot either print this photo or show it on a display. Why? Because these procesess lack of the required dynamic range.

What it can be done is to create a new low dynamic range photo from the HDR one that highlights the improvements obtained by all this. This is done by a process called tone mapping which uses different math functions to map the values of HDR pixels into a system with much lower dynamic. There is quite room for creativity here, so don't be surprised that some photos obtained by process look actually quite synthetic.

There are several programs that can help you with this process. I have tested Photomatix and ArtizenHDR. Photoshop 9 CS2 also includes support for HDR photography. And no, The GIMP cannot help you with this process as it is designed for low dynamic range photos only (8 bits per color per pixel).

Have a look at HDR photos on Flickr.

If your camera allows RAW format, then you can obtain images with higher dynamic range already (usually 12 bits per color per pixel) from just one shot. No wonder everybody is so excited about this feature on digital cameras (too bad I've just bought a DMC-FZ7).