This is my cyberhome!

November 12, 2014

Canon EOS-M – Poor Man’s RX1

Filed under: Photography — Manish Kumar @ 12:07 am
Tags: , ,

The best way to sum up Canon EOS-M (with 22mm lens) is this – It’s a poor man’s RX1. Of course it will not match RX1’s combo of full-frame sensor with a Zeiss prime but it comes awfully close. And all for a paltry sum of $300 which is what it was selling for as of November 2014.

I have read lots of reviews and examined hundreds of image samples from all kinds of cameras, including the aforementioned RX1. And I can say without any hesitation that this tiny 22mm f/2 lens has ridiculous amounts of resolving power. It’s up there with the best of them. In fact, it’s being limited by the relatively “low-res” sensor. Here is an example.

This is the complete image:

And here is a 100% crop:
Canon 22mm f/2 - 100% Crop

You can actually read those letters and read them comfortably! Notice how each letter is still distinct even at this tiny size. This would have turned into a blurry mess on a lesser lens. That is some serious resolving power. If this does not make your hair stand up, you are either dead or a medium-format user 🙂

I can only imagine what this lens could do on a higher-res sensor. Here is hoping to a 36MP M3.


October 15, 2010

How to fix gapless analysis problems in iTunes?

Filed under: Apple,Audio,Software — Manish Kumar @ 3:56 pm

The Problem:
If you use iTunes, on windows or mac, it is inevitable that you’d come across this dialog where iTunes says that it’s Determining Gapless Playback Information. It doesn’t matter whether you want your albums to play gapless or not. iTunes just goes ahead and does it. There is no way to turn it off. It would have been pretty harmless, except for one little thing – it often gets stuck and the only way out is to reboot your computer (and lose any unsaved data, not to mention the frustration caused). Though there is no way to turn this process off, there is a way to make it go smooth. The solution is to repair your corrupt MP3 files.

The Reason:
There are plenty of reasons for MP3 files being corrupt but the most common one is bad encoders. All the modern encoders are ok but those present during Napster era (late 90s and early 2000s) were really bad. So if you have any songs encoded during that period, chances are that they are not as per MP3 encoding standards. These songs would play just fine but otherwise cause a lot of problems in terms of tagging and gapless analysis etc. If you do a scan on your collection, you’d be surprised by the number of songs with potential problems. These kind of files are the reason iTunes gets stuck.

The Fix:
The fix is pretty simple, fortunately. All you need is foobar and a bit of time. Don’t worry. The songs will NOT be re-encoded or subject to anything else which might reduce their audio quality. These operations are completely harmless. You can take a backup in case you’d like to be really careful.

Ok, here we go:

  1. Download and install foobar. Make sure that you choose all the utilities and extras during the install.
  2. Launch foobar and go to ‘File -> Preferences -> Shell Integration’. In the right-hand side pane, enable the option ‘Folder context menus’. In the same pane, type ‘*.mp3’ for ‘Restrict incoming files to’. This will make sure that only MP3 files come into foobar as this fix is not applicable for AAC or WAV files etc.
  3. Go to your songs folder in Windows explorer, right-click and choose ‘Enqueue in foobar2000’. Or you and drag your songs folder to foobar.
  4. In foobar, Select all files, right-click and choose ‘Utilities -> Rebuild MP3 Stream’. This will recreate the MP3 file by removing all the non-standard data and other garbage. It will also fix the tags by rewriting them in a standard-compliant way. It took about 3 hours to fix 20k files on my aging PC. A nice side benefit of this fix is that a few more songs will now show up in your iTunes library because of standards-compliant tags.
  5. Once step 4 is over, select all files again, right-click and choose ‘Utilities -> Fix VBR MP3 Header’. This will add the missing VBR header data to the Variable Bit Rate files. The absence of this header causes problems in seeking and also in calculating the duration of song which impacts gapless analysis. This process also took about 3 hours on my somewhat old PC.
  6. This step is optional but I did it anyway. Clean up your iTunes library and add all the songs again. The only thing I care about is song ratings which I store in BPM field so I don’t lose them even if I clean up the iTunes library.
  7. Let iTunes go through ‘Determining Gapless Playback Information’. Typically it won’t get stuck now but there are chances that it still won’t like a few of the songs. The only fix for these songs is to remove them from iTunes library. In my case, I had about 1000 songs in a particular folder which were giving problems. These were the songs which I had converted from FLAC to MP3 a long time back, prehaps using a crappy non-standard encoder. I just removed these songs from iTunes, re-encoded them from FLAC using Lame and everything went fine.

To validate your MP3 files, you can use a free program called MP3val. It can even fix some of the common problems.

February 25, 2009

How to make Nikon Capture NX2 run faster.

Filed under: Photography,Software — Manish Kumar @ 10:14 am
Tags: , ,

To say that Nikon Capture NX (or Capture Nx2) is slow would be an understatement. The reason for its slowness is that it ships with older versions of certain Windows files which it uses. So all we need to do to make it run faster is to install the latest and greatest versions of these files. It really is that simple and it makes a significant difference. But why Nikon doesn’t document it or doesn’t do it by default is beyond me.

Here are the steps (Please read the whole post before proceeding with the install):

1. Download and install Microsoft .NET Framework 3.5 with Service Pack 1 from this page on Microsoft’s site. Capture NX2 installer specifies that it needs .NET framework 2.0 but if you install .NET framework 3.5, you do not need .NET Framework 2.0 or any other earlier version.

Note: If you do not want to install .NET Framework 3.5 for any reason and want to use version 2.0 itself, download the latest version of 2.0 with SP1 from this page.

Install updates, if any, through Windows Update (in Internet Explorer). If you already have .NET framework installed (2.0 or later), use Windows Update to get the latest version. Reboot if prompted.

2. Download and install Microsoft Visual C++ 2008 SP1 Redistributable Package from this page on Microsoft’s site. It’s a very small download (only 4 MB). If you install this, you do not need Microsoft Visual C++ 2005 Redistributable Package or any other earlier version. Capture NX2 ships with a very old version of this file (from 2005 or earlier) which is the main reason it runs slow.

3. Reboot your computer.

4. Install Nikon Capture NX (or NX2). If you have already installed it, you do not have to install it again. But if you are doing a fresh install, it’s better if you install it in the end.

5. Enjoy editing. (more…)

December 15, 2007

Making your DSLR show Linear Histogram

Filed under: Photography,Software — Manish Kumar @ 4:39 pm

In addition to having greater flexibility during post-processing, one of the main reasons for shooting raw is to capture greater dynamic range. But if you are relying on the histogram on your camera’s LCD to judge exposure, you may not be capturing the full dynamic range that your camera is capable of.

The Problem:
The histogram shown on the LCD of a digital camera is based on the gamma-corrected jpeg image and not on the linear raw data. This is true even if you are shooting raw. It is not possible technically to create this histogram from raw data because raw data is unusable until it is processed and converted into an image. Since the raw files typically have an exposure latitude of one f-stop and the blown out highlights can be recovered easily in the raw converter, we don’t want the in-camera histogram to treat this as over-exposure. Otherwise we’d be doing negative exposure compensation to get a pleasing in-camera histogram and thus losing valuable shadow detail. In other words, instead of following ‘Expose to the Right (for jpeg)’ (ETTR) principle, we want to follow ‘Expose to the Right (for raw)’. We want to push the shadows as far to the right as possible without really blowing out the highlights. And we want to do this because the way a digital sensor works, shadows are given very little importance when compared to highlights.

Human Vision vs. Sensor Vision:
Imagine you are sitting in a room where there is one light bulb of 100w burning. You can see things around you and the room appears to be at a certain level of brightness. Now, if a second 100w bulb is turned on, the room appears brighter but it does not appear twice as bright. You’d need lot more bulbs to make the room appear twice as bright. The same is true of human ears when it comes to listening ability. Doubling the sound pressure does not make the music sound twice as loud. This relationship between the actual light intensity and perceived light intensity is said to be logarithmic in nature. In other words, an increase in physical stimuli does not produce an equal increase in its perception. This relationship is described in mathematical terms by Weber–Fechner law or Stevens’ power law.

On the other hand, digital sensors (CCD or CMOS) in our digital cameras are linear. Doubling the ambient light will make things appear twice as bright to the digital sensor. This is an important point as the raw files in our digital cameras store linear data.

Linear Data:
The digital sensor in our cameras being linear, the raw files produced by them are also linear. The brightest f-stop in a linear file is twice as bright as the next f-stop and thus requires twice as many levels. In a 12-bit raw file, there are a total of 4096 brightness levels. That means that the brightest f-stop has 2048 levels in it, the next f-stop has 1024 levels and so on. The deep shadows in this linear file will have only 16 levels! Imagine the banding that will occur if you try to open up these shadows. Thus the brightest f-stop has far more levels in it than our eyes can perceive while the lower f-stops have far fewer levels than required. This non-uniform distribution of luminosity data can be fixed by applying gamma correction on the linear data.

A cathode-ray tube (CRT), which is used in the computer monitors, is inherently non-linear. The intensity of light reproduced at the screen of a CRT monitor is a non-linear function of its voltage input. This relationship between the input voltage and the output intensity is described by a parameter called gamma. To compensate for this non-linearity, the input signal to a CRT is encoded in such a way so as to make the light produced by the monitor perceptually uniform. This is called gamma correction.

As it turns out, the non-linearity of our eyes is very nearly the inverse of non-linearity of CRT. Thus gamma-correcting a raw image (converting to jpeg) not only takes care of the CRT non-linearity, it is also the best perceptual encoding for visual data. It makes the best use of limited bits (8 bits or 255 levels) available to achieve the most appealing visual reproduction. MP3 encoding does the similar thing for audio data. Note that it’s only the input to the CRT (jpeg on the hard disk) which is gamma-corrected. Once an image is displayed on the CRT screen, it’s as good as seeing the image subject in real life. It is this gamma-corrected image from which the LCD histogram is derived.

LCD Histogram:
To create an in-camera histogram out of raw data, four things have to be done. First, the raw data has to be de-mosaiced, and second, a color space has to be imposed on it. The third step is to apply gamma correction. If a histogram were to be plotted at this stage, it would be bunched up to the left because even though the brightest f-stop occupies half the total number of levels (2048 in a 12-bit raw file) and thus the complete right half of the histogram, there are very few pixels in it. This is true for typical images which comprise mostly of mid-tones and do not contain much shadow or highlight data. This is the histogram (from linear data) that we are after.

The fourth step is to apply all other settings like sharpening, contrast, saturation etc. and create a jpeg image. This jpeg image is then used to plot the in-camera histogram. This histogram looks very different from the linear histogram because gamma correction spreads the luminosity data uniformly (perceptually) across the tonal range. As long as the linear histogram does not show highlight clipping, we are ok even though the gamma-corrected version might indicate that clipping is happening.

So all we need to do to get a linear histogram is get rid of gamma correction and we can do that through custom curves (on Nikon cameras at least).

Custom Curves:
Most of the Nikon DSLRs have the option to upload a custom tone curve to be applied to the jpeg images and hence to the in-camera histogram. The custom tone curve (or the default tone curves for that matter) is comprised of two parts – an implicit gamma correction and the actual tone manipulation. All the tone curves work on gamma corrected images by default. What we are looking for is a tone curve, which:

  1. Does not apply gamma correction so that our data stays linear, and
  2. Does not apply any tone manipulation so that the linear data stays pristine.

Enter ToneUp Studio.

Applying a Custom Curve:
ToneUp Studio is a raw converter written by Todd Gibbs, and indie software developer. For all of $14 dollars, it not only provides custom curve uploading, but also tethered shooting (only for Nikon cameras). And you have the option to use Nikon SDK so that you see your images exactly as Nikon intended. Here is what is to be done to make your camera show raw histogram using ToneUp Studio (This feature is available only in latest beta as of now).

I have tested this process on D80 and it should work fine on other Nikon cameras too. Please do check the camera compatibility list on the author’s website.

  1. Put your camera in PTP (Picture Transfer Protocol) mode instead of USB mass storage mode and turn it off.
  2. Connect your camera to your computer with the USB cable and turn it on.
  3. Launch ToneUp Studio. You have to connect your camera before you launch this software otherwise it won’t recognize the camera.
  4. Go to Edit-> Preferences and enable the option to “Disable Gamma curve when uploading curves”.
  5. Go to File->New Curve. A new window with a straight line curve will be displayed. Choose File->Upload Curve. A progress bar will be displayed while the custom curve gets uploaded to the camera.
  6. Turn off your camera and disconnect.
  7. In your camera menu, choose Optimize Image->Custom Tone curve.
  8. Shoot.

That’s it. The images on your camera’s LCD will appear too dark but that’s ok. When you open your images in a raw converter, this custom curve will get ignored and you’ll see a normal (gamma corrected) image. If it looks over-exposed, just use the highlight recovery slider and bring back the detail. If you are using Capture NX or any other Nikon SDK based software, just choose a different tone curve.


  1. ToneUp Studio
  2. Expose to the Right
  3. Exposing for RAW
  4. Weber-Fechner Law
  5. Steven’s Power Law
  6. Gamma Correction
  7. Gamma FAQ

November 21, 2006

A review of Westone UM1 IEMs.

Filed under: Audio — Manish Kumar @ 1:12 pm

I have been using Philips HP805 cans for over a year now and HP800 before that. The only thing I didn’t like about those phones was the lack of sound isolation. I had also heard lot of good things about IEMs in general (curse you head-fi) so I thought I would get a pair. These being my first IEMs, I didn’t want to spend too much. After spending countless hours pouring over reviews and sound quality analysis, Westone UM1s seemed to have the best balance of price and sound quality. I got them from earphonesolutions for $109.

Packaging and Box Contents:
They come packed in a very nice high quality plastic box. The included carry case could have been better though. It is made of strong clothlike material and is quite sturdy but gets squished when pressed hard. I like it but wish it were made of solid plastic or metal though.
It comes with 5 pairs of foam tips, a wax cleaning tool, a carry pouch, and some printed instructions. The phones are not packed in the pouch but have their own recess in the plastic tray. Impressive packaging overall.

First impressions:
These are one tiny pair of IEMs. I was surprised on seeing them for the first time. Did they send me the wrong phones? They looked much bigger in the photos! Did I just pay $100 for this tiny piece of plastic? Turned out many other people had the same impressions, esp. first time IEM buyers like me.
And the cable is very thin too. It’s braided type and almost like a thread. I started having doubts about its sound quality, coming from thick cables of philips HP805 headphones. Then I realized why it was so. These phones need very little power to operate. The soldering lead on my philips headphones looks bigger than the entire UM1 driver.

Build Quality:
They are so small and everything is so tightly packed in them that it acts like one solid piece of plastic. You can drop them, toss them around, and give them a beating and they would still work fine. Care needs to be taken for the tip though where foam is inserted. That is a small plastic pipe protruding from the main unit and could break under moderate force.

One of the reason that I went for these phones was the comfort. And I was not disappointed. I can literally forget that I am wearing them. There is no pain or fatigue even after hours of usage. And they don’t protrude out of your ear at all. If I wear them and look in the mirror (from front), I can’t see them! You can even sleep (on your side) while wearing them and not feel a thing.

The UM1s have a very thin braided type cable which is extremely resistant to microphonics. They keep rubbing against my cloths and my office ID tag but never pick up any extraneous sound. Several people have this problem with some Shure and Etymotic models but UM1s are just fine.

My Setup:
iPod 5G (Apple lossless) -> Turbo dock -> Turbo mini-to-mini -> GoVibe 5 -> UM1

Sound Quality:
As with all IEMs, sound quality depends highly on the correct insertion and getting a proper seal. If the seal is not proper, these would sound muddy and won’t have any bass. So make sure that you insert them correctly before giving them a listen.
I typically use Chesky demo disc for making sound quality impressions ’cause I have listened to it countless times as part of headphone testing and know exactly what to expect. In fact, they tell you before each song what you should be listening for; a transcription of which can be found here. This would be my guide for judging UM1s. Just a note – The chesky demo disc is actually a SACD so the Redbook CD layer won’t be as detailed but it is still a very good reference.

  • High Resolution – Rebecca Pidgeon – Spanish Harlem
    Rebeca’s voice does sound real and you really feel that you can touch her. There are many times when her voice overpowers the shaker so you can’t really isolate its sound. The metal balls in the shaker move in a different pattern each time it is shaken and that’s what causes them to sound different. At the few places where shaker sound is prominent, each shaker does sound different from the one before it but this fact is not immediately apparent. You have to listen to it 3-4 times and even then the difference is not crystal clear. Though not extremely detailed, I would rate the UM1 resolution to be much better than average.
  • Depth – Sara K. – If I Could Sing Your Blues
    The trumpet that opens this song does sound distant and from slightly right to the stage. Sara’s voice sounds natural with just the right amount of reverberation. I could occasionaly hear the guitar pick but not too many times. The guitar is also full and warm but I could not judge the size of the studio. Not sure who is at fault here. Excellent performace by UM1 here.
  • Midrange Purity – Livingston Taylor – Grandma’s Hands
    There are 3 backup singers in this song and I could figure that out with UM1. But the finger snaps never had that flesh and bone quality to them. They sounded artificial to me. They sounded the same way on my philips HP805 too. Please let me know if your phones can reproduce it. Anyway, it is too much to expect this kind of timbre from such tiny and single-driver phones. Livingston’s voice does seem to have “chest” to it and sounds very real and human.
  • Naturalness – Ana Caram – Correnteza
    Excellent performance here by UM1. Ana’s voice sounds highly real, without any harshness or electronic gear. Mid-range is anyway the strongest point of these phones. The mids are neither too forward nor too recessed. They have just the right amount of laid-back quality that I enjoy.
  • Presence – McCoy Tyner & Joe Henderson – Ask Me Now
    No matter how much I listened to this recording, I could never hear the sound of the reflections off the rear wall. This just confirms that these phones are not too resolving. The mechanical sound of the keys was very clear though. The trumpet, at times, did not have the right timbre but it was still very much enjoyable.
  • Focus – Vivaldi, Flute Concerto in D
    I could easily pick out each instrument apart from the others. All the flute notes were distinct and quite clear, without any hint of blurriness.
  • Transients – Solisti New York – Stravinsky, The Royal March
    UM1 had no problem handling this piece. I think this quality is because of its tiny size and not inspite of it. The IEM driver displacement is extremely small when compared with the displacement of a full-size speaker diaphragm. So it has no problem starting and stopping with extreme precision. I do think the drum sound was not as tight as it could be but until there is a high-end system to compare against, I can not say decisively.

SQ impressions by frequency range:

  • Bass
    Track for bass tightness – “Hotel California” from “Hell Freezes Over”
    Track for bass extension – “Hero” (Jet Li) soundtrack by Tan Dun
    Very good, tight, and punchy bass. It doesn’t go as low as other high-end phones but it’s enough for normal music. Getting this kind of bass depends highly on having a good seal. If I turn up the volume slightly, I can get that chest thumping bass in Hotel California, right after the guitar intro. The bass is not too over-powering and not too diffuse, just the right kind.
  • Mids
    Strongest point of these phones. Absolutely no complaints here.
  • Highs
    This is the weakest point of these phones. There is a severe treble roll-off which makes them lose the details in the highs. I typically use “Latin” or “Treble booster” EQ on my iPod which helps a bit but still you feel something is lacking. This might be due to these being single-driver phones. On the plus side, this makes them non-fatiguing and you can listen to them for extended periods of time.
  • Soundstage
    Very good. I could easily make out the location of each instrument on several songs. On Sara K’s “If I could sing your blues”, the trumpet does sound distant and coming from the right side of the stage. On “Westmister Choir – Britten, Festival Te Deum”, you can really feel as if you are standing in a large cathedral. Impressive performance.

These are currently among the best IEMs you can buy for $100. After reading about expensive UEs, Shures, and Etymotics, I was not expecting much but I was completely surprised by their performance. Highly recommended especially for the first time IEM buyers.


  1. UM1 first impressions
  2. Initial impressions on the UM1’s
  3. Westone UM1 VS Etymotic ER-6 Review
  4. UE SuperFi 5 Pro vs. Westone UM1
  5. UM1s not living up to expectations
  6. Why aren’t there that many UM1 Reviews?
  7. This review on head-fi

July 4, 2006

How to install Mac OS x86 on a PC

Filed under: Apple — Manish Kumar @ 8:34 am

I have finally managed to install and boot into (these are two separate things in this context as will soon become clear) Mac OS x86 on my PC after struggling for more than two weeks. I used HotISO 10.4.6 install DVD. I don’t know much about legal stuff but it is probably illegal to do this without owning apple hardware. I own an iBook G4 so maybe it is ok. I don’t know. Please don’t quote me on this.

Before the install:
These are the things that you have to take care of before you start the install. 90% of people asking for help in the forums are there because they did not pay attention to this part. So before you start-

  1. Make sure that your DVD drive is set up as primary slave. If it is not, you would still be able to install the OS but you won’t be able to boot into it. It would say ‘waitng for root device’ or ‘b0 error’ and just hang there. Or the PC would keep on rebooting itself. Some of these problems can be fixed if you are able to boot from the DVD and go into terminal but you won’t be able to boot from the DVD either (after the install).
  2. Both your keyboard and mouse should either be PS/2 or both should be USB; you cannot mix and match. If you do, you won’t be able to boot into the OS after the install. It would complain about keyboard being not present and just stop there.
  3. Make a primary partition while you are in Windows. You can use Partition Magic or any other such tool to do non-destructive partitioning. Once created, mark this partition as ‘Active’ otherwise you won’t be able to boot into Mac OS X. You don’t have to format it at this point.
    You might want to make the size of the new partition sufficiently different from the size of Windows partition so that they are easy to tell apart in the Mac OS installer. A minimum size of 6GB is needed though.
  4. You don’t need Linux, fdisk, Acronis OS selector or any other such tool. Whatever you need is right there on the DVD itself. Some people suggest sacrificing a goat but I doubt if that’ll be of any help.

Install process:

  1. Boot off the install DVD and keep clicking until it’s time to choose the volume on which to install Mac OS X. The list would be blank as of now. Go to ‘Utilities –> Disk Utility’. The left side of the panel would now show all the partitions. Select the one you want and click on ‘erase’ tab in the right-hand panel. Choose ‘Journaled’ as the file-system, name the volume, and click on the ‘erase’ button. Try to avoid naming this volume with spaces in the name like ‘Mac OS X’ etc. I named this volume as ‘Tiger’. This kind of name would come in handy if you have to launch the terminal and copy files here and there.
  2. The next important panel is ‘Installation Type’. Choose ‘Custom’ and it would show you the list of packages to be installed. Pay attention to the last entry there which contains the patches and other drivers which make this whole exercise possible. Select only those patches here which match your system hardware. So if you have an intel CPU, do NOT select anything related to AMD. This is important. If you are not sure whether your CPU supports SSE2 or SSE3, download this small standalone utility. Again, select SSE3 only if your processor supports it.
  3. Sit back and relax.
  4. Installer would reboot automatically when it’s done and Mac OS X would start. Go through the simple setup process and enjoy.

If only computers were so simple! Post-installtion help forums on OSx86 project are a good place to start if you are having any problems with your setup. Make sure to go through the Technical FAQ first.


  1. Niall Douglas’s page on installing Mac OS x86
  2. OSx86 project

May 23, 2006

Kernel Streaming with iTunes.

Filed under: Apple,Audio,Software — Manish Kumar @ 5:36 am

iTunes is considered to be the best music organizer and foobar to be the best music player. But what if you could have the interface of iTunes but guts of foobar in one application? Enter Multi-plugin.

Here is what you’ll need:

  1. foobar (0.9.4) and Kernel Streaming plugin for foobar.
  2. iTunes (for Windows, ver 7.0.1)
  3. Multi-plugin (latest version is 2.4.2 as of September 22, 2006)

Install iTunes, foobar, and Multi-plugin. To install Kernel Streaming plugin for foobar, just copy the file “foo_out_ks.dll” to foobar components directory.

Run foobar, go to file-> preferences -> playback -> output and choose Kernel Streaming as the output method.

Run iTunes, go to Edit -> preferences -> Multi-plugin and make the following changes:

  1. Under “Appearance”, choose the default skin.
  2. Under “Other”, choose “foobar2000 passthrough”.

Now whenever you hit play on a song in iTunes, it would be played through foobar. This means that you get access to all the foobar DSP goodness too. Enjoy!


  1. You might get the message “foobar not installed” when you enable “foobar2000 passthrough”. This happens because Multi-plugin looks for foobar information in the wrong place in the registry. It’s pretty easy to fix though. Just transfer the keys from “HKEY_LOCAL_MACHINE\Software\Foobar2000” to “HKEY_CURRENT_USER\Software\Foobar2000”.
    To do that, go to “HKEY_LOCAL_MACHINE\Software\Foobar2000” in the registry and choose File->Export. Then open the resulting *.reg file using notepad and change “HKEY_LOCAL_MACHINE\Software\Foobar2000” to “HKEY_CURRENT_USER\Software\Foobar2000”. Save the file and double-click. Done!
  2. You need to set the volume to maximum in iTunes to avoid any loss of quality because of attenuation.
  3. Even if you don’t notice any improvement in the sound quality, this would at least take care of the songs skipping in iTunes (a big problem in iTunes 7.x.x).
  4. You can not run both foobar and iTunes at the same time. Otherwise you’d get error “KS output error: error opening device”. In fact, you don’t have to run foobar to use Multi-plugin. iTunes would launch it automatically in the background.

Update on September 20, 2006:
Links now point to the latest version of Multi-plugin, 2.4.2.

May 8, 2006

Notepad Vs Word Processor & Left Brain Vs Right Brain.

Filed under: writing — Manish Kumar @ 11:01 am

Christopher, a 15 year old boy in "The Curious Incident of the Dog in the Night-Time", suffers from autism. The left half of his brain is undeveloped so that the right half has to make up for that deficiency. His left brain has literally taken over his right brain and he has lost all the abilities that right brain is associated with. He cannot understand why his father is angry with him or why his actions annoy others. He takes things literally and is not able to understand the nuances or intonations in speech. When asked by a policeman how old he was, he replies, "I am 15 years and 3 months and 2 days".

Our left brain is associated with things that require logic, reasoning, words, grammer, list making etc and other verbal things while right brain deals with imagination, drawing, music, feelings, and other non-verbal things. Being a left-brained kid, Christopher is very good in mathematics and reasoning. He can solve complex puzzles in an instant but cannot understand why hitting a policeman is wrong. And he'd certainly be very poor in writing fiction.

Writing is essentially a right-brain activity constrained by left-brain rules. Some people are naturally gifted writers (Saki) while some acquire this art through hardwork (Stephen King). But can the medium of writing, the means of putting words on paper, make a difference in the quality of the resulting prose? Can you write a higher quality essay by composing it in a word processor than on a notepad? And how is it affected by left/right division of human brain functionality?

Editing on-the-fly:
One of the rules (or rather the only rule) for writing well is to write first and edit later. i.e. don't worry about grammer or spelling or coherency while writing your first draft. Just let your thoughts run wild. Do not pause to re-read; do not judge the quality of your prose; and certainly do not try to rephrase your sentences. In other words, write with your right brain and edit with your left brain. Notepad and typewriter are more conducive to this type of writing than a word processor. Once you comming something to paper using a notepad or a typewriter, it's not easy to go back and change it. So you are forced to move forward and let your right brain take control. But with a word processor, the tendency to edit on the fly is very strong; the words there are not concrete, unlike paper. The left brain takes over and this interrupts the flow of thought and slows you down.

Writing speed:
Human brain thinks at a very fast pace and it is not possible to write down those thoughts at the same speed. Out of five ideas that occur to you, you might be able to wtire down only one or two of them, losing the rest. Some ideas so ephemeral or so unique that they never occur again. So for taking down your thoughts fast, word processor is the tool of choice. It is common to achieve typing speeds of 60 wpm or more with a little practice. Notepad fares worst in this regard. Some typewriters do allow you to type fast but a mechanical typewriter cannot hope to compete with an electronic computer.

Thinking while writing:
Some people think as they write. I.e. they don't have any well formed thoughts before they put the pen to the paper. They just let the plot develop as their pen moves. This type of writing is best done on a word processor because typing with both hands engages both halves of the brain, thus maximizing your creativity. A notepad or a typewriter is better suited for those who like to think first and write later. For them, the act of writing is not so much writing but recording of their thoughts.

Writing as drawing:
People with a very strong right brain tend to think of writing (with a pen) as an expressive act. When they write in cursive, it's like they are drawing, like they are moving brush on a painting. This helps them connect with their inner artist and makes them much more articulative. J. K. Rowling, author of Harry Potter series of books, prefers to write on a lined paper in longhand. She calls laptops 'so eighteenth century'.

A programmer, on the other hand, would do best as a writer when he thinks of writing as coding. Though coding is not strictly a traditional artistic field, it requires creativity nonetheless. This way of thinking adapts natually to writing on a computer.

Writing being stressful:
Some people find the act of writing on a paper to be too stressful. You have to take care of margins, spacing between words, spacing between letters, keeping a straight line etc. This puts so much pressure on the brain that there is no energy left to do the actual thinking. An extreme case of this is called Dysgraphia. So a dysgraphic person would write much better using a word processor where all these things are taken care of for him. Same goes for people who are too lazy to write on paper or are just not used to it.

Adopting a wrong style:
Whatever an author writes has to be submitted to the copy editors at the publishing house. And in today's wired world, this invariably means emailing your manuscript. This forces the writer to compose using a word processor as it would be too tedious to transcribe from a notepad later. In fact, most publishing houses provide MS Word templates which are to be used to compose your manuscripts! If a writer is forced to adopt a writing style that does not suit him, the result is bound to be of inferior quality.

Now I understand why I write better using a word processor. I am a programmer, I think as I write, and I am too lazy to write using a pen. So there you go!

March 7, 2006

Introduction to Handwriting Analysis.

Filed under: Learning — Manish Kumar @ 7:18 am

They say it doesn’t matter what you say. What matters is how you say it. How was your body language? Were your eyes darting from place to place? Were you fidgeting? These visual cues carry more weightage ’cause they come from our unconscious mind. We have no contol over them. They are the ultimate giveaways. Eyes don’t lie as they say.

Our handwriting is one such visual cue. Our hand is just a medium to hold the pen. The actual writing is done by the mind. Or rather, by our subconscious mind. Notice how your pen just glides across the page, forming letters and words. Only thing you are aware of is what you are writing, not how you are writing it. You don’t notice how big your letters are, how much space is there between two words, how much pressure you are putting on the page, or how you are forming the individual letters. Which of three humps is bigger in your m? Are your Os open? Are your Ts crossed?

Handwriting analysis involves looking at all these visual cues and doing a personality assessment. It cannot tell what you had for breakfast and it cannot tell whether a brick would fall on your head. It can just tell what kind of a person you are. Short tempered? Egoistic? Friendly? Generous? Selfish? Satisfied with life? The answers may surprise you. Think you are frank and honest? Take a handwriting test.

The pen doesn’t lie as they say.

February 20, 2006

How to buy an apartment in Bangalore.

Filed under: Opinions — Manish Kumar @ 6:56 am

Even though this post is specific to Bangalore, the general principles should apply everywhere. I have not included everything related to buying an apartment here like location, security etc. Just those things which are not so obvious to new house-hunters.

Size: Do not buy a smaller unit thinking that you can always buy a bigger one later. Banish the thought forever. If you can’t afford a bigger house now, you won’t be able to do so after 5 years either. Property rates are doubling every two years whereas the salary is not. And once you get settled in a house, it is almost impossible to shift.
Looking ahead 10 years, you’d have kids who would need a room of their own. Your parents might come to stay with you. You might acquire a ton of household stuff which you’d have to put somewhere. So buy the biggest house you can afford. Look for one with atleast 3 bedrooms with a total area of about 2000 sqft. It might pinch a little now but will sure come in handy later.

Built-up area: The area of the apartment quoted by the builder is called super built-up area. This is the area for which you are paying the money. This super built-up area includes the area reserved for corridors, playgrounds, gardens and lift etc. The area in which you actually get to live in is called built-up area which is typically 80% of super built-up area. So if you are buying a 2000 sqft apartment, you get only 1600 sqft out of that. Some builders claim to give 85% but that is the limit.

Loan: This is the one thing that causes most anxiety to new homebuyers. There are many things to be taken care of while taking a loan.

  1. Pre-EMI: In a typical payment schedule, the bank releases a part of loan, say 10%, at each stage of construction. By the time you take possession, the bank would have paid the entire loan amount to the builder. If the construction takes 18 months, you have to pay the interest for 18 months on the money the bank has released. As and when the bank releases money, the amount on which you have to pay interest goes up. This interest amount is called Pre-EMI.
    The alternative to Pre-EMI is to ask the bank to release the full loan to the builder in the beginning itself. Then you can start paying full EMIs to the bank instead of paying Pre-EMIs. Even though paying full EMIs sounds bad when you can get away with paying much less Pre-EMIs, it is actually better and will save you as much as 5 lakhs!
    When you pay full money to the builder up-front, you get a discount of 4-5%. This works out to be 1.5-2 lakhs, depending upon the cost of the apartment.
    The Pre-EMIs that you pay do not count towards your loan. I.e. they do not bring your loan down. You are just paying a convenience fee to the bank. Most people go for this option ’cause they can’t pay full EMI and the rent at the same time.
  2. Interest rate: You can go for either fixed or floating rate of interest. Floating rates generally change every quarter but it is up to the bank. Fixed rates are of two types – fixed for a term and fixed for full tenure. The fixed term is typically 3 years after which there is a revision to the rate, depending on the market condition at that time. There are very few banks which offer fixed rate for full tenure. ICICI is one such bank. Some banks like Kotak offer loan with interest rate linked to the Fixed Deposit rate.
    You can change your loan from fixed to floating rate later and vice-versa but banks typically charge 0.5% of outstanding principle amount for this. There is one hidden cost though here. Your interest and principle components for EMI would be calculated again and you might end up paying more.
    As of Feb 2006, the floating rate is 7.75%, three year fixed is 8.25%, and fixed for full tenure is 8.75%.
  3. Pre-closure: Most loans last for about 6-7 years even though they were originally taken for 15-20 years. If you get some extra money and want to close off your loan, banks typically charge you 2% of the remaining loan amount. Some banks do not allow you to do this at all. In ICICI, you don’t have to pay any penalty for this if you leave 12 EMIs.
    Also check if you can pay more than your EMI once in a while. Banks typically allow you to make excess payments once in a quarter or once in 6 months.
  4. Insurance: Banks typically fund up to 85% of the apartment cost. Some banks fund up to 90% if you take loan insurance but 90% is the upper limit. The loan insurance premium is typically 8-10k per year. It covers things like disability, unemployment, death, or loss of property due to fire and theft etc. Unlike life insurance, you won’t get anything back at the end of coverage term. You might be better off taking simple life insurance if you are not concerned about job security etc.
  5. Tax exemption: You get tax benefits on pre-EMI and EMI only in the year in which you are taking possession. See more details here.
  6. Loan disbursement: All the builders have home loan tie-ups with various banks and they also have loan agents who deal with those particular banks. After your loan has been sanctioned, it takes a lot of co-ordination between the builder and the bank to disburse the money. All this becomes much easier if you take the loan through these builder appointed agents.

Cost: Like there is an ex-showroom price for cars, apartments have an ex-builder price (I just made up that term). A typical price of, say, 2000/- per sqft quoted by the builder does not include charges for Water supply, Electricity, Car parking, Service tax, VAT, Registration, and legal expenses etc. You won’t get any wardrobes or kitchen shelves either. Add 30% of the base cost for these things (total cost now = 2600/-). The builder should be able to tell you exactly how much would these things cost. Your loan eligibility is calculated on the sum total of all the above costs.

Premiums: Some builders, or rather all of them, ask for a premium for corner units or upper floors. In other cities, lower level floors cost more than the higher ones but in Bangalore it’s the other way around. This premium rate is typically 20/- to 50/- per sqft per floor. There is no problem in paying it except that it is not shown in any of the documents. Your house would still be registered at a rate of 2000/- only. See if your builder can waive it off or reduce it; most do.

Amenities: All the apartment complexes are advertised to have a club house, swimming pool, gym, garden, and playground etc. It would be a shame to call that pit a swimming pool but the point here is that you are made to pay for these things. If you are going to buy a house in one of these projects, there is nothing you can do to avoid paying for them. But what you can do is to buy a house in a much smaller apartment complex. These complexes typically have only 12-20 units and don’t have any of the above “luxuries”. You can see these kind of complexes everywhere in residential areas like Koramangala. The prices are about same as big complexes but then you get to live in the city and in a much better locality.

Salespeople: Do NOT trust the marketing executives who take you to the site and give you a tour. They would promise anything just to sell you the damn apartment. Always call up the customer care department and verify it with them. Better still, go to their office and have a look at the papers yourself.

When to buy: This one is simple. Buy it as soon as possible. Buy it now. Don’t worry about leaving Bangalore and going back to your hometown. you can always sell it later. And at a profit too. The best time to buy an apartment is during the pre-launch offer. These offers run for about 2 months and the rates are up to 200/- per sqft lower. As soon as the builder gets the plans approved from the govt authorities (and the project is officially launched), the rates shoot up (and keep on shooting up). The number of units on offer during pre-launch is very less though.
Another thing to note is that builders do not release (put up for sale) all the units once they launch the project. They keep some units for selling them later at a higher price.


  1. BDA Sites – A highly useful and relevant discussion group on MSN. Do not trust everything you read there though.
  2. Official website of BDA
  3. Official website of BMRDA
Next Page »

Create a free website or blog at