2004-09-30

Overloading white space in C++

This paper by Bjarne Stroustrup is an April Fool's joke. And you'll only get it if you are a C++ programmer. But it is so well written that if you didn't notice the date you might think it was real.



The best idea in the fake paper is letting you write C++ programs using Unicode, which is a giant character set that encompasses all the characters in the whole world. (Unicode is used internally in Windows XP, amazingly.) It's fun to browse through and be amazed at the characters that have been tracked down and formalized as Unicode.



For example, Dingbats are an official standard. What could be better than having Dingbats as an official international standard?

2004-09-29

Skylines

We had a visitor recently who said that L.A. doesn't have a skyline. It's not as good as Seattle's, but it is sort-of there.



Here is Seattle's, as seen from a ferry approaching the city:





Seattle Skyline






Here is Los Angeles', as seen across a sea of asphalt:





Los Angeles Skyline




2004-09-28

Forbes.com: Microsoft anti-spam effort dies

Forbes.com: Microsoft anti-spam effort dies



I still like my idea: lower the cost of a digital signature for email to a buck. Everybody buys one for any email accounts they care about, and everyone rejects mail that isn't digitally signed.



The end.

2004-09-27

USATODAY.com - Virgin to launch commercial space flights

USATODAY.com - Virgin to launch commercial space flights



Speaking of Virgin ... Richard Branson is launching "Virgin Galactic" which will provide suborbital space flights for $208,000.00.



I met Branson a couple of times when I worked at Virgin Interactive. He's a walking pile of charisma, which is interesting, because he's basically shy.



I have a saying: there are two kinds of CEOs. The kind that can add and subtract, and the kind that multiply.



The kind that can add and subtract figure out how much something costs to make and then charge a bit more, thus making money.



The kind that can multiply are the ones that figure out how much something costs to make, then make a spreadsheet that shows if they sell huge numbers of the thing, they'll make money. These guys usually fall prey to "the big spreadsheet lie" and generally lose money - and lots of it.



Branson is the first kind of CEO - he only goes into businesses where the operating margin makes sense. He loves to find businesses that have high operating margins and then undercut the market by discounting a bit. As a result he's been fabulously successful.



He's also smart about when he unloads a company - once the market gets crowded he usually gets out.



The only problem with buying a company from Branson is that unless you really understand the Virgin culture and work to maintain it, you're likely to lose a lot of good people, as corporate-ness sets it. Virgin is one of those few "anti-corporate" companies that actually pull it off.



2004-09-26

Gamasutra - Features - "Applying Game Design to Virtual Environments" [01.01.98]

Gamasutra - Features - "Applying Game Design to Virtual Environments" [01.01.98]



Another article - this one on Virtual Environments - by that same guy (registration with Gamasutra may be required).

Confusing



No Exit Exit




Click for the bigger picture. (The picture is from some old bathrooms in the the LA Convention Center - in an old portion of the property that was not updated in the big expansion in 1993.)

2004-09-24

Jack to the Future! (and Ratings Rants)

Finally, after much waiting, the final four episodes of Samurai Jack are going to air this Saturday on Cartoon Network.



I was lucky enough to work on a Samurai Jack game for PS2 and Gamecube. The development process had a lot of ups and downs but one of the "ups" was the chance to work directly with some of the talented people at Cartoon Network Studios that actually make the show. It's rare when you work on a licensed property that you actually get to meet with the talent (although that is starting to change, thank goodness).



We met with Genndy a couple of times and took some direction from him. He'd look at a model, make one change, and it was way better! Samurai Jack is a very stylized show and it was great to get inside Genndy's head a little bit and understand how and why that style exists.



Another highlight was when Genndy said, "That animation is pretty good!" I have David Hunt to thank for that.



The nice people at Warner Bros. and Cartoon Network gave me and a couple of other team members a cool maquette, which sits on my desk here at home now.









My personal opinion about making games for kids is that the rating system really gets in the way.



You can't make a low "T" game and have it succeed.



Let me explain: the ratings that matter are "E" (Everyone), "T" (Teen), and "M" (Mature).



The E rating has to be non-objectionable to parents of six year olds. That currently means almost no violence - even cartoon violence. If you have a game built around combat of any kind you get a T rating. This is because of the increased graphic quality of current generation games. The exact same game on PS1 or Nintendo 64 would get an E rating. Gameboy Advance "E" games can be quite violent because the depiction of violence is very low-resolution.



"T" ratings are divided into ranges: "Low T", "Mid T", and "High T". A game that used to be an "E" on the N64 would be a "Low T" on the Gamecube because of the better "more realistic" graphical quality.



Unfortunately, these distinctions between "Low T" and "Mid T" are not made by Moms and Dads when they go to the store to buy Christmas presents. So a game like Samurai Jack or even Super Smash Bros. which are made for kids 8-12 get "T" ratings, which completely misrepresents the situation.



I think there needs to be a "PT" or "Pre-Teen" rating. That tells Mom what she wants to know - this game has some action but it's okay for kids.



If we had "PT" then games like Sponge Bob which are definitely "E" games would have their place and then games like Super Smash Bros. and (of course) Samurai Jack would have their place.



Another problem with the "Low T" rating - print and online game magazines review the game and compare it against other more hardcore ("Mid T") games. "PT" would set their expectations correctly and we would get more accurate reviews.



It's kind of stupid for the ratings board to make these distinctions of "Low T", "Mid T", and "High T" and not make those distinctions obvious to Mom and Dad.



I would think the same distinction needs to be made for "High T" games. They need a rating too. Medal of Honor is an "M" game except that there is (1) no blood and (2) it deals with a historic event. So it is a "High T" game.



BTW, I'm not making this shit up. All of this "Low T" etc. stuff comes from a talk I attended that was presented by the ESRB. Medal of Honor was specifically addressed in the talk.



My undestanding is that the ESRB wants a pretty simple ratings system, which is why there is no "PT" or "HT" rating. I think that thinking was fine when Moms and Dads were less sophisticated about games, but games are mainstream enough and there is enough awareness of these distinctions that the overly simple system the ESRB has now is not doing anyone any favors anymore.



My proposal just exposes distinctions that the ESRB is making internally already.



Do you know how many people set the rating for a game?



Two.



Three people actually look at the game, but if two agree on the rating, then that's it.



I read a science fiction story once where statistics and polling were so advanced that voting for president came down to sampling one person's opinion. The rest of it was interpolated by computers.



The ESRB is almost there.



2004-09-22

Too many words

I went down to Borders in Redmond to buy the latest Lemony Snicket book. I noticed that Neal Stephenson was going to be talking and signing books there later in the evening. I also noticed a lot of nerdy people hanging around. I actually noticed the nerdy people first, and then saw the notice that Neal Stephenson was dropping by, and it all made sense.



I met Neal Stephenson, who is a shy guy, so I'm surprised he is venturing out to talk to strangers, when I worked at Virgin Interactive and he came by to talk to me about turning Snow Crash into a cool CD-ROM game like 7th Guest. When I told him it would cost upwards of a million dollars he lost interest, as I think he wanted to fund it himself.



I like his books, in general. I even liked the Baroque Cycle - or I should say I liked the idea of it. I couldn't actually finish even the first book, although I tried.



There were just too many words.



I like my fiction in bite-sized hunks.



Maybe when it's all in paperback, and I can actually read it without injuring myself (the hardback books are very heavy!), I'll give it another try.

"Loved the Picture"

'Loved the picture' says Guy Johnson. Guy's site is here. He's proud of the fact that there is very little text in his web site - it's all pictures until you drill down into his gambling log.

"Kichiguy" is his nickname. In Japanese it means, "crazy person", as in "crazy, drooling, insane, lunatic, scary kind of guy."

2004-09-21

Collapse



Photo (C) 2004 Stephen Clarke-Willson





This is the Air France terminal at Charles de Gaulle airport outside Paris.



My wife and I went to Paris (without the kids!) for a week to celebrate 20 years together.



A portion of this roof collapsed two months after we were there, killing several people.



_______________________



(Subject change: Posting pictures with "Hello" is too easy. I might end up posting a lot of pictures. Except there is no automatic copyright notice. The "hello" people should add that as a feature.)

Hello



Okay - apparently I can upload pictures with "Hello", a free program that works with Blogger. We'll see how this looks. Posted by Hello



This is from a lightning storm we had here in Sammamish last June. It used to be impossible to get a picture like this without great skill and patience. But in this case, the picture was actually shot on miniDV and then I grabbed an interesting looking frame later on the computer. Again, this works because digital cameras are sensitive to low light.

Plog, Phlog?

I wish I could just paste pictures in here. I guess that would make it a "plog" - a photo log. No, it turns out a plog is something different.



How about a phlog? Yup - that's it.



Anyway, I have to type in actual HTML in order to put a thumbnail in here.



I have taken thousands of pictures with my handy digital cameras and some of them, perhaps by accident, are good.



My friend Guy Johnson, who loves taking pictures, says a good photo strikes an emotional chord in you. You don't need to stare at it - you take a look and hopefully POW!



My wife thought this picture of my daughter at an Ice Cream Parlor in Prescott, Arizona does that.



So click below, and hopefully, POW!





Ice Cream Parlor




"Knife, this is Variable"

"Knife, this is Variable." That's a line from a Tom Clancy book. It has nothing to do with this post.



If you are curious about audio compression, it is worth compressing some of your favorite music with a variable compression rate. Take your favorite encoder and set it to "VBR 100", which means "keep this music as close to the original as possible." It's an interesting experiment, because, for instance, the highest constant bit rate (CBR) that Windows Media allows is 192 kbps, but if you use VBR 100, the encoder will decide to encode music at up to 320 kbps!



You're more likely to have an mp3 encoder and you can do the same with that - set it to variable rate encoding at 100% and then play the file in WinAmp. You can watch the bit rate go up and down with the music.



It's tempting to encode everything with VBR100 and get the best sounding music you can get. The reason I don't is that VBR is a hack (a brilliant hack) that was invented by Real and doesn't work in all players. My minidisc software (SonicStage - what crap) in particular can't stand mp3 files that are encoded at more than 256 kbps and a VBR file is likely to jump up over that.



Also, every mp3 player I have seen can't seek accurately in the file, because the variable sized packets of encoded information make it impossible for it to calculate the correct location in a file. That gets annoying.



I've found that 160 kbps mp3 files rarely have artifacts that you would notice in everyday use. 160 used to be my default encoding rate but I've switched to encoding at 192 kbps for mp3 as my default because I feel it gives me some headroom in case I want to edit a file once or twice, which will cause recompression. For professional work, I encode at 192 kbps wma, which sounds terrific.

2004-09-20

Bach to the Future!

CPU Bach is a program that Sid Meier wrote for the 3DO that composes music in the style of JS Bach.

If you download this (24 megabytes), you can get 101 minutes of very pleasant very-Bach-like music to listen to. It's really good. Listening to it reminds me of listening to an NPR Sunday morning show.

I have an old 3DO machine hooked up to my stereo with the CPU Bach disk in it. The program never gained much popularity because it was only on the 3DO. Meier put it on the 3DO because at the time it had superior sound output compared to the PC (this is 10 years ago?).

Anyway, turning on the 3DO machine results in it automatically composing and playing Bach-like music. So I captured 101 minutes of it which I put into this huge 240 megabyte mp3 file. If I posted that, and four people downloaded it, then my bandwidth limit of a gigabyte a day would be used up ... I needed more compression ... I listened closely to the music and decided that there wasn't that much going on frequency-wise, and that I could probably compress it down to a really small wma file that would sound good. So I did, and that's what's posted. The resulting file is only 24 megs and it sounds quite pleasant.

Bach to the future!

2004-09-18

The Apprentice - NOT

At first I thought, Trump has really screwed the pooch on this one. He fired "the best guy". What kind of idiot would fire the best guy?



But Bradford wasn't the best guy. Trump's mistake was calling him the best guy and then firing him. But I guess he was trying to make a point.



It took me an hour to come around to Trump's point-of-view, which is that, when you have a really great advantage in business, don't piss it away, no matter how much you think "you don't need it." A successful business is always firing on as many cylinders as possible.



So Bradford was fired for over confidence. I know a guy who was over confident - he risked millions of dollars on a market play that didn't work out. He ignored one of Trump's main rules: minimize the downside and because of that the downside came and bit him in the ass.



My wife has a saying she likes: "never challenge worse," as when people say, "things can't get any worse!" Yes, they can!



Ask Bradford.

2004-09-17

CBS Fonts

There has been a lot of noise about faked records of George Bush's service record in the Air National Guard.



This particular site (or try here if that doesn't work, or better yet, go here for a brief summary of the typography issues) goes into gory detail about how the fonts used are from Microsoft Word and a laser printer and not from a typewriter or whatever they had available in 1972. To be honest, I didn't read the whole thing - I got the point pretty fast.



But I did skim it all, and if you are interested in typography it is an interesting read.



Back in college (1985, 1986), Donald Knuth was writing his five book series Computers and Typesetting - it was a big deal, as it gave us poor slobs writing our Ph.D.s complete reproducible control over our papers. This was good! The old way of typing it on a typewriter and getting it exactly right was hard! A laser printer! A good thing! Incredibly detailed typesetting program! A good thing! (Martha Stewart would like typsetting with Tex.)



The cool thing was that Knuth would pay you if you found a typo in his drafts, which he published on the Internet. I found one, quite late in the process, and received a check for $2.56. I was so poor in college that I cashed it. Moron. I wish I had that check now - what a cool keepsake! (I did save an early draft of The TeX Book.)



My wife, God bless her, bought me the five book C&S series as my graduation present from the Ph.D. program. Since my "old family" was completely anti-intellectual (I am the only one to finish college out of four kids), this touched me greatly.



I wonder who CBS hired to verify the documents. Heads will roll over this one. A couple of more days of denial and someone is a goner.



(An alternate view that proposes that it was an old Selectric that made the documents is here.)





2004-09-16

Voltage

One of the companies that Ann Winblad has invested in is called Voltage. You can visit their site at (what else) http://www.voltage.com .



What they have to offer is digitally signed email without the overhead of the PKI - public key infrastructure.



I had been thinking that it is easy enough to get rid of all Spam - Verisign lowers it's cost for a digital signature to a buck, and everyone buys one. Presto, no more forged email. If the message isn't signed, it is easily rejected by Outlook or any email client.



What a simple fix!



But Voltage has gone one better - they derive a key from your email address and nothing else!



I have no idea how they are going to make money on this - it does require a server on their part to generate a key derived from your email address - so maybe they will charge for that.



Either way - get the price down to a buck a year, and no more Spam!

Ann Winblad

Speaking of Gates, my wife and I attended a dinner hosted by the MIT Enterprise Forum of the Northwest. The speaker was Ann Winblad. She is most famous as one of Gates' old girlfriends. She's actually an accomplished Venture Capitalist working out of the San Jose area.



Her feeling is that we are finally digging out of the VC "free money" extravaganza that caused the big meltdown in 1999 - 2001.



She also said a big sea change is underway, where all kinds of investments that the VC community would have avoided in the past are okay now - consumer products, vertically integrated businesses, and the big win right now is service oriented businesses. Her view is that the infrastructure technology for gluing services together is sufficiently advanced that it now makes sense to build a company focused on a particular market segment - for instance, accounting for hog farmers - rather than a big monolithic product like Oracle or SAP - and to rent the software to the hog farmer rather than selling it.



The funnist moment, which most people didn't notice, was when she said, "Venture Capitalists have opinions. Everyone has an opinion. Opinions are .... are .... opinions .... everyone has an opinion, right?" You could tell she really was about to say, "Opinions are like assholes - everyone has one," but decided midstream to reconsider her use of the word "asshole" with the MIT crowd.



I enjoyed the evening and it was fun having my wife along. The guy sitting next to me was responsible for Xbox hardware quality, so we chatted some about the game business. That was fun too. My wife is in charge of our investments and she chatted with the lady next to her about investing.



And the food was okay too.

2004-09-15

Gates

I met Bill Gates when I worked at Virgin Interactive because I was the producer of "The 7th Guest."



Graeme Devine, who programmed Guest, had been working with Microsoft on getting it to run under Windows so Gates could show it at a Multimedia conference.



Thanks to Graeme, I was able to tag along to the rehearsal for the presentation, which was the night before the actual event. I think this took place at Moscone Center in San Francisco.



Gates dropped by and talked with Graeme and pretty much ignored me. But I was listening. The thing that struck me was how relaxed he was. He was only worth about $7B then - I thought of him as The Seven Billion Dollar Man.



After chatting with Graeme he did a dry-run of his presentation. He was the most charismatic speaker I had ever seen. He was laughing, making jokes about CD-I, and generally having a great time. His staff was working really hard to make sure everything went great and while you could see they were working hard and focused they didn't seem unhappy or like people that were being forced into unpleasant labor. They liked working for him.



It was an incredible experience for me, because Gates was completely unlike his public persona, which is a nerdy, goofy, nervous-nelly kind of guy.



How could this be?



Graeme and I left - but before we got into our car, I realized I need to go back and and go to the bathroom. On my way to the bathroom, Gates and an associate (maybe Glaser - I can't remember) were coming out of the auditorium. With every step, Gates slowly transformed into his nerdy self. It was amazing.



I realized from that that when Gates was in his element - the dry-run for his talk - surrounded by his staff, where he was in complete control - he was the most charismatic person on earth. I could see why people worked so hard and were so eager to please him and generally worked their asses off.



But out of his element, he was a very nervous dude, which is they guy we see on TV.



I read in article in Slate once by a reporter that had dinner with Gates and he commented on how relaxed and interesting Gates was. The reporter had to go to great lengths to point out that even though Microsoft owned Slate, the reporter was not an apologist for Gates or Microsoft. He had to defend his comments that Gates was an interesting dinner companion. That is how at-odds Gates' public and private personas are.



So - a big question answered - how could that nerdy dude we see on TV build a huge empire? Answer - he's different in private.

2004-09-14

Slashdot | Rob Glaser Responds, Talks Up Real Networks

Slashdot Rob Glaser Responds, Talks Up Real Networks



The article is an interview with Rob Glaser, who I met once when he worked at Microsoft on the "Tandy VIS", which was just like Xbox, except a 286 processor, one megabyte of RAM, and one megabyte of ROM. It ran a cutdown version of Windows. I should write more about that some day. I still own one and it might still work.



Rob talks about why online songs cost what they do, amongst other interesting things.



When I met Rob, which was around 1992, he was in charge of Multimedia stuff at Microsoft. I was at the MS campus for a VIS conference. Gates was supposed to come by and say hi and tell us about Microsoft's commitment to multimedia (remember, this is 1992!). He didn't show up so we saw a little video tape instead where he told us about Microsoft's committment to multimedia.



In the tape, Gates kept smiling at odd times.



I was sitting at a lunch table with Glaser and I asked him about it - what's with the weird timing with the smiles?



He told me that Gates is very uncomfortable with PR and that his PR people had tried to train him to smile more, and so every so often he would remember to smile, regardless of what he was talking about.



Later, when I really met Gates, I found him the be the most charismatic person I had ever met.



But that's another story.

2004-09-13

Culling

I just read a startling article in IEEE Software (Sept/Oct 2004) called "Why Culling Software Colleagues is Popular", by Peter Middleton, Ho Woo Lee, and Shahruk A. Irani.



The article's main points are:

  • The staff likes it when you get rid of marginal performers; and
  • Managers don't like to do it.

A third interesting point was that

  • Erratic performers annoy their colleagues.

The article explains how Jack Welch as CEO of GE implemented a bell curve for giving out raises and forcing managers to let people go. I think Microsoft does this too. Anyway, Welch found it was hard to get his managers to implement the system, and of course, it got harder as time went on, because the curve moved up as the people at the lower end of the curve were let go.

The astounding thing was that the remaining staff loved it. The reason? Those who remained perceived that those who were let go were holding them back.

One of the main ways the "lesser-performers" annoyed their colleagues was by erratic performance. I've had some people work for me that were pretty erratic, but I always tolerated it. This, in spite of people coming into my office to complain about the erratic people on a fairly regular basis. I never had "erratic" as one of my hot buttons for letting someone go. For me to let someone go, their performance had to be obviously bad, or their skills just plain didn't fit what we wanted to do (those latter kinds of firings are really hard).

Now, after this article, I am seriously rethinking that stance. I'm adding "erratic" to my list of things to be on the lookout for in performance. Now that I think about it, I've never seen an employee evaluation form that asks for an evaluation of how erratic someone's performance is. For instance, you can hit deadlines, which is often mentioned on the form, but still be erratic about intermediate work products, and the like, which can cause other people to miss deadlines.

Food for thought.

2004-09-11

AmphiPod

I was at Target in the electronics section while in Arizona on a trip and I saw this cute thing for carrying your iPod around - the AmphiPod. Once I got back to the hotel I tried it out ... and my iPod didn't fit! What kind of crap was that?

Actually, it almost fits, which turns out to be good enough.

Well, the AmphiPod was never made for carrying an iPod, as you can see if you visit their site.

The model I have is this one which is made for runners wearing loose-fitting clothing.

I guess the buyer for electronics at Target thought as I did - hey, a cool accessory for your iPod!

Well, as I said, it works well enough - it's not exactly the fit you would expect, but it is quite handy if you are wearing loose-fitting clothing and want to carry your iPod or iPod mini around.

Of course, it shouldn't be confused with these other AmphiPods.

2004-09-08

Connectedness

I have some CGI pictures posted here. Suddenly I started getting a lot of hits at that part of my web site.



It turns out some people starting linking them in as background pictures for their web sites.



Here's one. Here's another one. And another one complete with Cave music from Mario Bros.



And here are some people talking about my rendition of Chopsticks (mp3 or MIDI).



I learned all this because I stumbled across a cool web reporting tool that my ISP (pair.com) supports which I had never enabled. Cool beans!



Whaddya know.

2004-09-06

256

The day after I mentioned that you can recompress high-bit-rate files a few times without worrying about it, Buy.com announced that many of their titles are now available at 256 kbps.

2004-09-05

Recompress

If you do a little reading about MP3 and how lossy compression works you'll soon come across a warning to avoid recompressing your files. In general, this is a good idea, since even if you have a super-high-bit rate file, a little is lost just in doing the math to convert between the time domain (the wave file) and the frequency domain (the mp3 file).



But there are exceptions and they are significant. First, recompressing a file once isn't generally a problem with a file that is encoded at 160 kbits or above. Recompressing twice probably isn't a problem either. Recompressing 100 times is bad.



The Sony Minidisc makes use of the fact that recompressing once isn't bad - every time you put an mp3 file onto the Minidisc it is converted into Sony's proprietary Atrac format.



So there you have a product that requires recompression.



Recompressing, of course, is worse at lower bit rates, because more information is thrown away each time. You might think that once the information is thrown away the first time, there would be enough bits to recompress without loss, but unfortunaely due to the way it all works, subtle problems with alignment and the math of it cause further loss. On the other hand, if you have tons of bits, then very little is thrown away each time.



2004-09-04

AllofMP3 - seems legal!

I downloaded my first song for $0.035 and I thought, this has to be illegal, so I'll just write-off the rest of my $5.00 and give up.



As far as I can tell, all of the over 4500 files I have encoded on my computer are legal. I had a few songs from the old Napster but I tracked down legal copies of those, either on CD or from on-line stores.



Interestingly, and as a complete aside, in normal life I would never have listened to Eminem. He was on Saturday Night and did a bad job. But he was so much in the news that I finally went to the old Napster and grabbed a couple of songs from him. I was blown away! I bought all of his albums. Napster worked as an advertising medium. Personally, I think the record companies should give away low-bit-rate versions of all the music in the world. It's the perfect replacement for AM and FM radio.



But anyway... after doing some research, I decided that AllOfMp3.com was legal enough - I mean, how much due-diligence can your average consumer do? I checked around, and it appears to be legal.



So I bought about 40 songs, which cost about $2.50 .



It's actually a pain to download that many songs. I would get an email when the song was encoded, click on it, and then it would s l o w l y download. They have some kind of download manager you can get, but I haven't tried that. I guess that's the next thing on the list.



I'm sure the record companies hate AllOfMp3.com, but in this case, they're going to have to work through all kinds of international copyright laws, so AllOfMp3.com is probably legal for a few more years.



2004-09-03

Allofmp3 review

Allofmp3 review



AllOfMp3.com charges by the megabyte! One or two cents a megabyte. And you can choose the bit rate, so if you want a crappy file, it's cheaper than a high bit rate file. But at these rates, why wouldn't you grab the best quality you can? And there is no DRM.



I was pretty sure this site must be illegal, since I bought a song for less than 4 cents. (They do math down to the 1/1000th of a cent, because they can.)



They have songs you can't get in the US.



The review above claims it is all legal due the vagaries of international copyright.



Now that's amazing.

2004-09-02

Bits

If you have enough bits to store lots of frequency data, then "compressed" music can sound just like the original.



What's interesting is what happens when you run out of bits to store all the frequency data.



That's when the "psychoacoustic" modeling kicks in. The encoder wants to throw away data and so the psychoacoustic model is used to throw away the frequencies you are least likely to hear.



A big optimization is to simply throw away big hunks of information: switching from stereo to mono, or using a trick called joint-stereo, where each frequency is allocated to only one side of the stereo image. Also just throwing away the high end frequencies is good when you're encoding at low bit rates. You've heard it a million times that people can hear up to 20kHz but that's a kid with perfect hearing. Dropping that limit down significantly gets rid of a lot of frequencies.



Even so, at really low bit rates, there still aren't enough bits for all of the frequency data you might want to store, so the frequencies that your ear just most likely won't hear are tossed.



I'm listening to this internet radio station right now on decent computer speakers and it sounds pretty good! It's a 20kbps .wma stream from WKSU (an NPR classical music station). The volume isn't too loud so I don't hear how bad it sounds and I'm busy typing this so I'm not being too critical.



Listening with headphones, as I am now, is more irritating. It just sounds hollow. That's probably because a boatload of frequencies are missing.



The most irritating artifact of low-bit rate encoding, especially in mp3 files, is something I'll call flutter, since I don't know if there is an official digital term for what happens.



Interestingly enough, music is compressed into frames (about 38 per second), just like movies are in frames. Each frame is encoded with separate rules. If you do Variable Bit Rate encoding, then the rules for each frame are really different - including the bit rate.



The flutter, I believe, comes from each frame getting encoded with slightly different rules. In one frame there is a certain frequency and in the next frame it is missing. This would occur for frequencies that were on the border for getting tossed. At a low enough bit rate, there will be lots of these kinds of frequencies. So they "flutter" in and out and it is REALLY irritating.



Someone should fix that in an encoder - put in some history or something so the frequencies on the edge don't flip back and forth. The mp3 format doesn't need to change - just the encoder.



Somone get to work on that!

2004-09-01

Lossless

There are two parts to this whole whoop-tee-do about music compression - the part you hear about, called "psychoacoustic modeling", and the part you don't hear about, called nothing, because you never hear about it.



The first part is somewhat obscure as to what it actually is. I spent a day at an Audio Engineering Seminar at the University of Washington and learned more about analog-to-digital coverters and compression and human hearing than I thought possible. (We have world famous sound guys here in the Pacific Northwest because of Microsoft, Real, and Mackie.)



When I first heard of psychoacoustic modeling, I thought it meant that they had figured out the psychology of listening, and knew, for instance, that if the drums were loud, you weren't concentrating on the trumpet, or some such magic.



Actually, it probably has very little to do with psychology. It's more about how your ear actually works physically. Of course, since we don't cut people open to see how their hearing works while they are alive, the details of how human hearing works are unknown at this time.



Anyway, what psychoacoustic modeling actually means is that when two frequencies are really close to each other, then depending on their relative volumes, you may only hear one of them.



So to compress drums and still have them sound good, you could probably throw away half or more of all the frequencies in the drums.



The more frequencies there are in the music, the more you can throw away without most people hearing it.



There you go. That's the first part - the "famous" part - the psychoacoustic modeling part.



The second part you never hear about is simply this: missing frequencies are expensive to encode the regular way (in .wav or pcm files or as they say, in the time domain), but can be cheaply encoded in the frequency domain (the way .wma or .mp3 or .rax files are stored).



What this means is that if your music has a low enough number of frequencies in it, then an mp3 file can nearly losslessly encode it. It's simply a more efficient encoding. Nothing needs to be thrown away except silence.



Well, I guess it's true that if psychoacoustic modeling throws away the frequencies you can't hear, then the ones that aren't there would count in that, but that's not generally what they are talking about.



Silence is golden. In this case, silent frequencies are golden.



The ramification of this is that you can compress music with mp3 or wma to a high degree and literally not lose any of the real music. Some math errors creep in, but they are minor.



When Microsoft says wma at 192 kbps is "transparent CD quality", and Sony's new Hi-MD Minidisc player can "transparently encode" CDs at 256 kbps, they mean it!



That compression is nearly lossless.



Nobody will ever tell you that, except me. I think they (whoever they are) don't want to draw attention to the fact that at high bit rates, "lossy" compression isn't.

iPod vs. The Cassette

Pretty Funny.



Cassettes are still semi-viable. Cassettes sound okay in your car if you have a good car stereo. There's so much road noise in your car a lot of things sound pretty good.



What's really nice, if you still want to listen to cassettes, is to build a playlist with Winamp that is as close to 45 minutes long as possible so you know it will fit nicely on one side of a 90 minute tape. (Don't use 120 minute tapes - too fragile.) It was always a pain in the past to figure out if the next song would fit on the tape!



Now, with computers, it's easy!