HDD Killers

OK, so if it wasn’t obvious before, it certainly is now, because four days ago I wrote, “I’ll post a few of entries on consecutive days.” Well, sod that, I’m just gonna put it all into one, otherwise more stuff will pile up while I, no doubt, procastinate further.

So yeah, I’ll start with a techy thing. Everyone loves music, but few people know or care about how it’s stored digitally on their computers. Well, I’m going to tell you. First off, the music has to be recorded, to do that it is sampled many thousands of times per second. CD’s have a sample rate of 44100 times per second, with each sample being 16 bits in size. With two audio channels that comes to 1411200 bits / second, which is quite a lot. That’s all fairly straight forward, but where it gets less straight forward is when you try to compress it. There are a few lossless audio formats around, such as FLAC, which compress the audio to about 60 – 70 % of the original file size, which still leaves the files somewhat large.

Then there are lossy audio formats, such as MPEG-1 Audio layer 3 (or MP3 as it’s better known), WMV, OGG, AAC, etc. These use psycho-acoustic models in order to simplify the audio stream so that to most people it sounds similar to the original audio. Quiet noises and high frequencies may be removed from the audio to improve the compression, things you’d probably have to actively listen for to notice their absence, but even just passivly listening, the audio can seem somewhat flatter and less real. 128 Kbps or 192 Kbps MP3 is about the level where most people like their MP3’s to be, discerning listeners prefer MP3’s recorded at higher bit rates such as 256 Kbps or even 320 Kbps, while hardcore audiophiles with the most ultimate audio setups can only make do with losslessly compressed audio.

So it seems that it’s a balancing act, do you want better audio or a smaller file size? When it comes to the internet, smaller files are the be all and end all. No-one wants to wait all day. But wouldn’t it be great if there didn’t have to be a compromise? Current lossless audio codecs offer an average compression ratio of about 1.5 to 1 (compared to uncompressed audio), whereas with a 128 Kbps MP3 file, the compression ratio is 11 to 1. It’s quite a tall order, if not impossible, to compress audio to such a high ratio losslessly, but … perhaps that doesn’t matter. What if it were possible to achieve ratios closer to 2 (700 Kbps) or even 3 (470 Kbps) for lossless audio, would that be enough to encourage the switch over? With internet connections getting faster, hard drive space on the increase and portable music players getting more and more memory, would a bump in lossless compression result in a one size fits all audio file?

Perhaps some time in the future, but if such a bump were to happen now, it would most likely be just the discerning listeners who switched. As you may have been able to tell by now, of late I have been very interested in audio and lossless compression thereof. I would very much like to experiment with C++ to compress some audio files and see how the compression ratios turn out (as soon as I figure out how to work the compiler … and code in C++). I have quite a number of ideas that when combined I think might perhaps be able to achieve compression ratios of 2 or even higher. We shall see.

Now, leaving the techno babble behind, on Wednesday I got thirsty, so I did what anyone would do, I googled for Mountain Dew. I don’t know anywhere in the UK that sells it, so I figured maybe a website would. Turns out I could bypass the brick and mortar completely, there are businesses that actually sell drinks online, ohmigosh! So I ordered a few cans from Cyber Candy, and on Friday they er … they arrived. And were stacked. And photographed. Here’s a medium sized shot (1.5 MP), and here’s a full size shot (10 MP). The shot was taken with a Canon EOS 400D, (you hear that LCF? 400D ftw!). The price wasn’t too bad either, £45.98 including delivery, which works out at a tad over 63.86 pence per can. I’M DRINKING MY MONEY AWAY!!!

Finally, Mensa got in touch and far earlier than expected, only 6 days after I took the test. So much for the 2 – 4 weeks we were told. The requirement for acceptance in Mensa is to be in the top 2% of the population in terms of intelligence. Two separate IQ tests were administered and coming in the top 2% on either one (or both) is a pass. The pass mark IQ’s were 132 for the culture fair test and 148 on the Cattell III B test.

As expected, I didn’t pwn the pants off Mensa, and fell short on both tests. My scores were 124 on the culture fair, (top 7%, 8 IQ points short of the top 2%), and 136 on the Cattell III B (top 6%, 12 IQ points short of the top 2%). So I didn’t do badly at all, in fact I’m rather chuffed with my scores. Byroz didn’t do badly either, but I’ll let him brag about his scores, should he choose to. To anyone wondering why the pass mark IQ’s are so different on the two tests, Mensa has this to say.

As different IQ tests were developed, each was given its own scoring system. Therefore, an IQ of 150 is a meaningless claim unless you know the actual test which was used. In order to compare one IQ test against another, the scores are converted to ‘percentiles’, i.e. where a person’s score falls in comparison to the rest of the population by percentage. Mensa offers membership to anyone whose IQ score places them within the top two per cent of the population, no matter which approved test was used.

So there we go, lossless audio compression, Mountain Dew and Mensa, all in one post. It’s like a variety show in text form. I should start charging for this…