Actually this has been one of those ideas that grows worse on you the more you think about it, but well, one has to do it some time and that time for me was last Friday.

The thing is, while one of couse has to have an opinion on beer flavours, the inevitable discussions that this causes does make one wonder to what extent the taste of beer is influenced by commercials and the general image you have of beer brands. So, the idea was to buy some different brands of beers and test them without knowing which was which. This actually serves a two-fold goal: first, you can find out which beers are the best, and secondly, you can test your beer knowledge by checking what beers you recognize.

The test setup

So, we bought two bottles of a whopping sixteen different kinds of beer: Amstel, Bavaria, Bitburger, Brouwmeester (the cheep Aldi brand), Dab (a German brand being served in bottles in de Gloppe, a cafe I go to often), Dommelsch, Grolsch, Heineken, Hertog Jan, Hoegaarden, Hollandia, Jupiler, Leeuw (served in Vera, Palm, Pitt (another cheep beer brand) and Warsteiner.

Now, we did the test with 5 persons (me, Mattijs, Vera, Sander B and Sander P), and each of us had a testing form on which we graded each beer for 5 different categories: aroma, looks (”uiterlijk”), taste (”smaak”) and taste afterwards (”nasmaak”), and a general judgement. Apart from that, we also tried to have a guess at which beer was which. The fill-in form, actually the standard Excel invoice template modified a little by me, can be found here (XLS) or here (PDF). The beer was being served by our lovely serving team Brenda, Elvira, Friso and Inonge (many thanks!) and we received about half a small glass of each beer to test and grade.

Problems

Of course, as could be expected with a lot of beer at hand, and in generally with badly prepared events such as this, some problems were to be expected. With the beer being served in our kitchen and there not being a door between the kitchen and the living room, of course one does tend to see some things some time, which is the reason why everybody guessed the fifth beer, Palm, right. After that, we moved to our closed veranda, which solved that problem.

Another problem is the inevitable cross-influencing one has on this kind of occasion. Sixteen beers is a lot, and no-one of us had actually tasted all beers previously, so when one person distinctly seems to recognize a beer, other people tend to go along with it. So, by tactically copying other people’s guesses one can cetrainly succeed in scoring some points. An example that springs to mind is the Hertog Jan that was giessed right by four out of five, and Vera guessing the Dab right while never actually having had the doubtful pleasure of tasting it.

A final observation is that sixteen beers is just about the number where the whole beer testing thing starts getting a bit boring and people tend to get a bit distracted. It’s just that when you go to the supermarket, there are just so many different kinds of beer that simply deserve to be tested so a selection of essential beers just doesn’t get much smaller.

Test results

Anyway, let’s jump-kick on to the results. As a true mathematician, I of course entered the results in a spreadsheed (graphs, tables and PDF exporting all done with the excellent OpenOffice.org Calc) and throwed some statistics on it to turn the numbers into answers. The resulting sheet can be found here (OOo calc format) or here (PDF). I’ll just go on to draw the most interesting conclusions from our little investigation.

First about our game at seeing who would guess most beers right. The results were a bit disappointing here, though I guess it was to be expected. Truth is, without practice beers are just very hard to distinguish.

I actually had two initial goals: guessing Hertog Jan and Dab. I have had only one bottle of Dab in my whole life, a few months ago, but it struck me having a particularly bad taste (bit of a sulfur-metallic combination), so I guessed I could recognize it, especially since Dab beer has become the archetype of bad beer for me. Hertog Jan, OTOH, is generally considered by us to be one of the best beers around, and a bit of an expensive one at that too, so I thought I ought to recognize it. I am quite proud to say I succeeded in both goals! At the negative side, I didn’t guess Amstel right, which is the beer we usually drink at home, so that was a bit disappointing.

Another interesting point was the Heineken Beer. It seems to be a hip thing nowadays to think of Heineken as having a very bad taste, and both Sander P and Vera suffered quite a bit from this before the test. While they not only didn’t recognize Heineken right, they also actually enjoyed drinking it: Vera gave it an average 6, and Sander P even gave it an above-average 6.5, ranking it at sixth place out of sixteen. Which I found quite funny. Seems that the Heineken myth is now proved to be not true.

Overall, the Challenge Cup (a beer glass with with the two lowest-ranked beers) went to Sander B, who guessed five beers right (including the forementioned Palm). I went second with four, and Sander P only succeeded in getting one beer right other than the Palm giveaway. Prooves the point that many beers are not easily distinguishable to me.

Let’s get on to the final judgement. Here’s the data in a neat table:

As you can see, kudoos go out to Bavaria for making the obviously winning beer, being the best in the aroma, taste, after taste and general categories, and being pretty cheap at that. It finishing above Hertog Jan (and way above, too) was quite a surprise for us. Closer examination learns that Vera was highly responsible for this though, giving Hertog Jan only 5.5 out of 10 whereas the rest of us gave it 7/10 or more. Still, even without that Bavaria would have won though.

Another interesting observation is the high, fourth, rank for Pitt, the second-cheapest beer in the test. Seems Pitt, mistaken for the high-end Jupiler and Leeuw twice, and costing only 70 cents per liter, is definitely going to replace Brouwmeester (10th place) as our low-end beer. That is, if I can’t help it, since I actually placed it on 14th place with a 4.5 grade…

Another thing is the low placement of Leeuw, the beer served in Vera, our favorite club in Groningen. Probably due to having had more than enough change to taste it due to Pooier 666 fame, Sander B and I gave it a 6 and 7, respectively, but the other grades were just plain bad, causing it to reach a disappointing 14th place.

About the last place, I can only say Dab deserved nothing better. I don’t know why this relatively expensive beer is being made at all, or why Sander P cared to give it a 5.5, but the average 2.8 for aftertaste pretty much tells the whole story of this awful beer.

Another interesting thing to look at is the standard deviation of the final judgements of the beers, which is a measurement of the extent to which we agreed about a particular brand of beer. Included in the full report, this number again shows good results for Bavaria, which had the lowest standard deviation value of only 0.49, which means that nobody was really negative about it. The highest standard deviation, interestingly, is for Hoegaarden, a Belgian white beer that ended up at 5th place: it had a standard deviation of 1.75, which is a pretty high figure. And indeed, opinions on this beer ranged from 4 (me) to 8 (Vera). Otherwise, it does seem that for some reason the more commonly known beers do have smaller standard deviation values, whereas people tend to differ more strongly in opinion on the more lower-priced beers.

Finally, as a price-minded Dutchman, I of course had to take a look at the relation between the price of the beers and their ratings. Calculating the correlation coefficient raised the surprising result that there seems to be no relation at all betwee the two factors, with the coefficient ending up as -0.06, actually indicating a very small negative correlation: the cheeper the beer, the beter, indeed!

To conclude things, the graph that just about says it all can be found here: it plots the quality of the beers against their prices (also showing the correlation line). From this graph, we can conclude that Bavaria is pretty obviously the best beer choice available, offering excellent quality at a good price. As a low-ender, Pitt should apparently be the beer of choice.

Overall, it has proved to be both an interesting and much enjoyable experience. It would be nice though to do the same test again and see whether one would actually get the same results. I can pretty well imagine that next time, after some more practicing, we will succeed in getting Hertog Jan on first place…

While OpenOffice.org has been my favorite office suite for some time now (being free and open-source and available for Windows and Linux and stuff), I never found myself in a great need to do macro programming, even though OpenOffice.org has its own Basic-derived (and thus Visual Basic for Applications-like) programming language. So, being a slightly experienced VBA programmer (I did some checks and balancing macro’s for my mothers shop once), I thought I’d give StarBasic, the OOo macro language, a try today.

My task (for the Great Exiciting Beer Test, more on that to follow), was the following: I wanted to have a macro function that can be used in OpenOffice.org Calcs formula’s to calculate the rank of a given score in a list of scores. For example, given the list of numbers [5,4,5,1,2], I’d want a function to return the value 3 for 4, it being the third in rank after the two fives. Now, one can almost feel there would have to be a core function, or a combination of core functions, of OpenOffice.org to do something like that, but I couldn’t find one, so I wrote my own.

Well, I’ll spare you the details of all the trouble I found before getting to the right solution, but here it goes. First, in the “Extra->Macros” menu, choose the “Macros…” item. Now, select the folw in which you want to create the macro (the default selection, “Module1″, is the one to choose if you want the macro to be available for all spreadsheets — otherwise create a new module in the file you’re working in). Hit the “Edit” button and a new window will pop up in which we’re going to write the source code. Now’s the time to actually write some code.

Now, I’m going to assume you already know some Basic. Now, the function we’re going to create (I called it Ranglijst, which is “Ranklist” in English) is going to expect two parameters: the range in which to look, and the value to return the rank for. For example, if the cells A1 to A12 would contain some values and we’d want to know the rank of the value in cell A4, we’d enter something like =Ranglijst(A1:A12;A4) in OpenOffice.org Calc. In the macro call, the first value will be a two-dimensional array (in the function declaration, we call this an Object since I’m not sure what else to call it), and the second one will be a numberic value (the double data type). This means our function declaration will look like:

Function Ranglijst(X as Object, Y as Double)

Allright, so what we’re going to do is go through all the cells in X, and if one has a larger value than Y, we’ll increase the rank we’re going to return. This is pretty trivial to implement, except maybe for one thing: we need to know the dimensions (i.e. size) of the X array. Now, for any competent Basic programmer that will probably not be a problem, which is why it’s not in any FAQ or the Sun documentation or something (grrr… I spent quite some time finding this out), but this is done by calling the Ubound function. Finally, to set the return value, we set the Ranglijst variable to the value we want to return (which is quite different from other programming language’s return statements, too, but I’m glad I still remembered that). This leaves us with the following piece of code:

Function Ranglijst(X as Object, Y as Double)
  Dim maxX as Integer, maxY as Integer, ret as Integer

  ret = 1
  maxX = Ubound(X, 1)
  maxY = Ubound(X, 2)

  for tx = 1 to maxX
    for ty = 1 to maxY
      if (X(tx, ty) > Y) then
        ret = ret + 1
      end if
    next
  next

  Ranglijst = ret
End Function

Now, this is what this eventually looks like in the OpenOffice.org macro editor. When all is done, save the macro file (you can hit the “Compile” button (first button next to the location text field) first to check whether everything’s working).

Finally, we can use our function in our Calc spreadsheet as shown here, and, joy joy, our macro works.

Well, maybe typing this was an utter waste of time, but I found myself to have a bit of trouble getting it to work, so I thought I’d share this with you. More help, including API docs and code snippets, can be found at the OpenOffice.org developer page.

I downloaded a two-VCD version of Shrek 2 the other day. Which is fine for viewing on a PC, but if you want it on a DVD player with (Dutch) subtitles, you need to do some extra work. I think I got it to work (didn’t try on my standalone DVD player though), and here’s how:

First, transform the VCD files into a DVD VOB. I described transcoding in the DVD authoring section earlier, but in this case I was lazy so I decided I’d skip on transcoding to see whether my DVD player would accept VCD data on a DVD. This can be done by at first mpgdemux‘ing (from mjpegtools) the stream, and then mplex‘ing (id.) it again. See DVD authoring for details.

First combining the two files into one and then doing this is not a good idea, I found: mplex would stop halfway through, apparently signalling the end of the first file.

This might raise a problem if you have subtitles for the complete movie that haven’t been split. I figured there had to be a tool for this kind of thing, and guessed it would be written in Perl, and a quick google came up with this great script. Usage is pretty self-explanatory. It doesn’t enable shifting with millisecond precision, but could be trivially hacked to do so (I didn’t test my hack since I didn’t need it, but it should work).

Next you can insert the subtitles in the DVD stream by using spumux (from dvdauthor). A typical config should look like this:

<subpictures>
   <stream>
      <textsub filename="shrekcd2.srt" characterset="ISO8859-1"
         fontsize="14.0" font="arial.ttf" horizontal-alignment="center"
         vertical-alignment="bottom" left-margin="60" right-margin="60"
         top-margin="20" bottom-margin="30" subtitle-fps="25"
         movie-fps="25" movie-width="352" movie-height="288"
      />
   </stream>
</subpictures>

Note the movie-width and movie-height parameters that were adjusted. Also, this font size seems about right (maybe a bit largish) for VCD, but for other resolutions it probably needs adjusting. Now, run the .vob file through spumux:

spumux shrek2.conf < shrek2.vob > shrek2-mux.vob

Finally, this file can be authored by DVDauthor. I didn’t bother making fancy menu’s this time, so I just did:

dvdauthor -o dvd/ shrek1.vob
dvdauthor -o dvd/ shrek2.vob
dvdauthor -o dvd/ -T

The result should burn just fine.

Update: 24-06-2004: The resulting DVD seems to have some problems after all. Though Xine and Media Player Classic on the PC seem to play the thing just fine, my standalone player, the one-and-only Lenco, shows the subtitle video stream in the upper left corner of the image, like it has blown up the movie stream to a more standard DVD size (704×576), but left the subtitles alone. Which causes the subtitles to not only be displayed at about the center of the screen, but also so small that they’re pretty much illegible (especially on our pretty old TV which doesn’t have a very clear image anyway).

Either way, I suspect the solution would be to specify something like movie-width="704" movie-height="576" in the configuration stream. Another thing: it seems that to specify the language of the subtitle, you’ll need to use a DVDauthor configuration file: it cannot be done directly from spumux. Which I’m not going to do, but it should not be a problem.

Update (2): 25-06-2004: I tried it today with movie-width="704" movie-height="576", and though this does satisfy the Lenco (that is, it displays the subtitles at the right place), it rightfully upsets both Xine and Media Player Classic, that refuse to display the subtitles now. Well, it is a Lenco, of course… Probably our Swiss friends (”the originally Swiss brand name of Lenco has been synonymous with quality, reliability and innovation”) are wrong here and this is a firmware bug. What I did here was making the subtitle stream twice as large as the video stream, so I can imagine video players showing only the upper left quart of the subtitle image.

Anyways, I guess one just needs to check what solution works in every separate case. Ah well, glad it works now.

Listened to my recently purchased copy of Hail to the Thief, and I found out the official version was different from the one I had been listening to all the time! Maked sense, because I had downloaded it before the actual release, so I probably had an old leaked version. OTOH, the same version also seems to have been in a Radiohead collection I downloaded, so I was quite surprised… No revolutionary changes, most are just very subtle, but noticable nonetheless. Can’t say which version is better though.

Anyway, another reason to want to rip the album, which the copyright protection tries to prevent me from. Which is ridiculous from today onward since I own the CD now… Luckily, the CDDA Paranoia package for Unix seems to be ripping through the thing quite well (and it’s more than halfway through now…). For some odd reason, my usual ripping package, GRIP, which uses paranoia, seems to choke at the CD, but then again this might be something with the hardware because I didn’t try other CD’s, but doing paranoia from the command line works just fine:

cdparanoia -sB
cdparanoia -vsQ

The first command gives a listing, the second one does the actual ripping. So, if you own the CD and you’d like to have the MP3’s, I’d be glad to share them with you… what a waste of time copyright protection is.

Stumbled across two cheapishly priced Radiohead CD’s today at the Mediamarkt: Hail to the Thief for 8 euro, and Com Lag for 18 euro (though that is admittedly not that cheap I was going to buy it anyway and the other shop had it for 20 euro). So I bought them, at once doubling my Radiohead collection (previously only OK Computer and Meeting People is Easy). Going to buy Kid A soon probably, too.

One of the main reasons I wanted Com Lag (previously featured in Currently Listening To) since I saw it is that the cover art is really brilliant. More info here. The fact that it was originally a Japan release (the front says: “This compilation is for Japan“) also causes this: the booklet has Japanese song texts and a Japanese back paper, which is quite a funny effect. I’m going to have a bit of a closer look when I’m in the train back home.

Also stumbled around what is called “the holy grail of Radiohead bootlegs”: Towering above the Rest, a 10-CD set filled with live tracks and B-sides. Torrent available, obviously. Anyway, I’m really wanting to have this thing now. Also, there’s some FTP site which has the first two CD’s, from which I downloaded the full version of Life in a Glass House, which has a terrific jazzy intro, so I’m really looking forward to listening to the rest of the 10 CD’s.

A bit of a less joyful experience was the fact that my copy of Hail to the Thief was copy-protected! Quite a disappointment, especially given the fact that you’d expect Radiohead of all bands to be against this kind of thing. Anyway, the consequence is that the CD doesn’t play under Linux (tried it with kscd only though), and tries to install some trashy player under Windows. This kind of thing really forces you to download the album. For me, being able to listen to the thing on my MP3 player is just an absolute requirement for the CD to be of much use. I mean, what are these no-dicks at EMI thinking? I’m sorry, but the whole pointlessness of it all just makes me a bit angry. So now I would have to waste my time downloading an album off the internet when I just legally purchased it? I mean, it’s not like this copy protection thingy is going to stop people copying it, now is it? Rather the reverse, I’d say: sort of makes you want to make tons of copies of the CD and throw them at random people in a shopping mall…

That felt good.

In unrelated news, I cleaned my keyboard a week ago or so, but in the process I broke my left CTRL key, which is quite irritating, and my ‘m’ key is working a lot worse now as well, I really have to hit it in the middle to make anything come out of it. So if some m’s are missing, you know why…

The money actually got to my bank account last friday! After the charge of 12 Euro for international fees from my bank, €600 was left, which is nice enough. Finally the end of this chapter I guess.

In the ever-lasting “Currently listening to…” series, today we feature Com Lag, the latest EP by Radiohead.

Now, Radiohead has been my favorite band for quite some time, but I was beginning to long for some new material, so I was pretty happy when I found out about this special edition EP, released in Japan and Australia following the recent Radiohead tour there. Amazon has it from import for $33, but incidentally I also stumbled around it for €20 at a local record store, so I am considering buying the thing.

Basically, half of Com Lag is different versions of songs from the Hail to the Thief album, and the other half is new songs. From the recycled songs department, the album kicks off with a live recording of 2+2=5, but the live version doesn’t add much to the original. The Scatterbrain and Myxomatosis remixes (they couldn’t help calling the latter Remyxomatosis) are much of an improvement on the other hand. Precisely those two songs were my least favorite songs of HttT, and apparently I was not the only one if they picked exactly those two. The I Will recording sounds more cheerful than the original, which is nice.

Over to the new material, we see basically two types of songs on this album: electronic songs reminiscent of Kid A, and some simple balads. Nothing rocky on it, making the choice of adding 2+2=5 all the more strange since it barely fits with the rest. Overall, the new tracks are pretty nice. No real negative exceptions, but on the positive side I tend to really like Paperback writer and I am a wicked child.

The ultimate test for a CD is of course listening to it until you get bored, and over the last (long) weekend I have indeed listened to Com Lag for almost 10 times now. Though this is enough to make you want to listen to something else, the fact that no song gets irritating yet is a good sign. Overall, neat CD, though I don’t think it’s better than any of the last four Radiohead studio albums.

With the elections for the European parliament coming up tomorrow, I did the Stemwijzer once again to finally make up my mind, and today I succeeded in getting the advise I wanted, with GroenLinks, the Dutch environmentalist party on the first place, and PvdA and SP, the other Dutch left-wing parties, on the second and fourth place. This is a relief to me, since I was planning to vote GroenLinks anyway, so this rationalizes the decision.

Now that Pinkpop 2004 is over again, it’s time to grab them audiostreams! Large parts of the festival are available for listening at the really great 3VOOR12 website. Unfortunately, the files are RealAudio files so you can only listen to them, and not download them.

Luckily, there is a solution; if you run Linux, that is (bet there’s something for Windows as well, but the Linux solution works fine for me): you can use the Vsound program, which acts as an intermediary between RealPlayer and the soundcard, capturing everything that goes in between. This way, while listening to a concert, you can record it without loss of quality at the same time. This can be achieved with a command like this:

vsound -f franz.wav -d -t realplay franz.sram

In this case, franz.sram is the downloaded RealAudio stream description file, and franz.wav is the output WAV file. The -d option enables output to file and soundcard at the same time; the -t option is nessecary for programs that do streaming, such as Realplayer.

One more thing: if, while playing the stream, the playback stops for a bit for buffering, which does tend to happen occasionally for me, this will result in a very loud buzz in the resulting audio stream. Luckily, these buzzes are very easily recognizable and removable with Audacity, but after a bit of editing, the result is usually pretty nice.

Sadly and shamefully, Muse have decided their concert cannot be streamed over the internet (along with Lenny Kravitz, but I think that’s hardly something to be sad about), but that still leaves enough nice concerts to listen to. 3VOOR12 has an amazing audio archive with recordings from many older festivals as well, so that’s definitely something to check out.