17 October, 2011

Whoa! Slow Down There, Glenn

On the Glenn Beck Program today, Glenn and his guest by phone were getting all bunched up about a school in Texas which was teaching the Mexican national anthem and pledge.  Not knowing anything more, I'll have to assume the latter is similar to our (the United States') "I pledge allegiance to the flag of the United States of Ameria..."  Initially, I was about as indignant as Glenn and his guest.  From what I've heard, there's been quite enough, thank you, of indoctrination of people into a mindset that America is bad, Americans are aggressors, and such-like.  But then I heard more.

The guest was the father of the girl who was "forced" to recite/sing these.  He related that this was for a Spanish class.  It was then I had more of a reaction like the initial reaction of King Arthur's Knights of the Round Table in "Monty Python and the Holy Grail" when they first encounter the rabbit ("you got me all worked up over nothing!").  I too was schooled in Spanish, albeit in Cheektowaga, NY, not Texas.  This is very likely very mild, and no cause for concern at all.

When learning a language, while not strictly required, it is often quite helpful to learn some of the cultural aspects as well.  It's just standard practice, best I can tell.  While I never had to do anything like memorize any Spanish-speaking country's anthem, it's a fine way to enlarge vocabulary and learn a part of the culture.  The complaint presented was really weak too.

The caller complained about Spanish being derived from Spain, so why should they learn the Mexican anthem?  Why not the Spanish anthem?  That's so ridiculous that it barely requires explanation.  And this is a subject with which I have a slight amount of expertise.  I was taught there are two major variants of Spanish: the European (or Castillian in English, Castellano in Spanish) and the Latin American (LA).  Just like English, there are variations on that, but likewise I'd say there's American English and British English.  We were made aware of the differences in Spanish (chiefly pronunciation), but we concentrated on the LA variant.  You can imagine why.  As residents of America, we're far more likely to encounter people speaking the LA variant.

I personally don't think Glenn is much predisposed to sensationalism, but I think that applies here.  Sorry, but I think obstinance on the part of this Texas family is why this girl got a 13 for this exercise (presumably percent).  Just because you're asked to memorize and repeat some particular passage doesn't mean you are actually agreeing with such a pledge.  It is merely a memorization exercise, combined with knowing when someone hears the pledge or anthem, one is able to identify it and its significance.  It won't be merely a semirandom collection of words to the listener/student.

If this had been during the course of some other, non-Spanish-language, class, I could understand, and agree with, the outrage.  But really...this is much ado about nothing.  But I would agree with part of the referenced article: I don't agree that the US stole parts from Mexico, and would prefer students not be told that.  In particular, it's called "The Gadsden Purchase," not "The Gadsden Annexation."  I will have to admit history is my poorest subject, so I'm not sure of the circumstances of the Texas annexation.

English is a difficult enough language to interpret correctly when its rules are followed, let alone when the speaker or writer chooses not to follow those rules.

"Jeopardy!" replies and randomcaps really suck!

Please join me on Google+!

12 October, 2011

I'm Conflicted About Sibilant Voice Talent

When you hear somone speak the words of an ad (radio or or TV), or give you the options for an IVR, or perform for an animated program, the people who do that are called "voice talent" or "voice actors."  I myself have recorded some voice talent, and have been a minor one, having produced a few PBX recordings for Sitel.  (I also recorded some of the autoattendant announcements, such as at +1(716)871-6400, if it still happens to be there.)  But what I don't quite understand is, why do media producers think sibilant people are good?

If you don't know what sibilance is, it's kind of difficult to explain concisely.  All I can tell you is, when sibilant people pronounce words with the letter "s" in them, you'll definitely know it.  Sometimes, but not always, President Obama speaks with sibiliance.  If a dog is lying down and trying to get to sleep for a nap within a kilometer, a sibiliant person should make them raise their head and point their ears in the sibilant person's general direction.  (Well, maybe not, but I was trying to think of an analogy to illustrate higher than normal high frequency content.)

The problem is, for some people, like myself and a friend of mine named "PJ," hearing a sibilant person talk evokes a reaction like most people have to the scratching of fingernails on a chalkboard.  This is so much so that we have to, as politely as possible, leave the vicinity of sibilant speakers.  One such example I can recall vividly is the birthday party for a mutual friend of ours.  His wife just nails each "s" so hard, we had to sort of try our best to steer clear of her.  Thankfully such high audio frequencies have similar characteristics to high radio frequencies: they're mostly "line-of-sight," so ducking around a corner (such as into a different room) alleviates that feeling of anxiety...so that's all we generally had to do.

Now, I realize that PJ and I may be in the minority, maybe even in the vast minority, but still...don't you think ad producers in particular would want to irritate as few people as possible?  After all, aren't you trying to establish the widest possible audience for the product you're pitching?  I daresay, choosing sibilant voice talent is one way to cut down on your audience size.

The saddest thing about it is, I know there's virtually nothing these people can do about it; that's the vocal tract their genetics have made.  The best we audio engineers can do is add a sibilance controller to the processing chain1.  But this only helps so much.  It's more-or-less only a Band-Aid on the problem.

So, in short, I don't know.  I'm very conflicted.  Sibilant voice talents are usually quite superb in every other respect.  In one sense, I'd rather not see them be fired or not selected for a given voiceover job (because it's something they are rather than something they choose to be), but on the other hand, I want to cringe less.  But just like I am innately unsuited for many jobs, such as being a Buffalo Sabre or a member of a construction crew (either of which requires quite a lot of strength which I don't have, and probably never will), so too these people really could do rather better in some other endeavor...please.  And I do say/write that with a certain amount of regret, please believe me.

1 A sibilance controller is a circuit or (DSP) software which applies an automatic gain control to high frequencies.  It's as if a person were constantly monitoring the audio, and cranking down the treble whenever sibilance is heard (and cranking it back up to normal when there isn't so much sibilance).

English is a difficult enough language to interpret correctly when its rules are followed, let alone when the speaker or writer chooses not to follow those rules.

"Jeopardy!" replies and randomcaps really suck!

Please join me on Google+!

01 October, 2011

A Little Disillusioned by Technology

I know it's pretty wrong to judge almost anything on very few data points, but....It's been a rainy half week here in western New York State, some days to do indoor rather than outdoor things.  So I thought it was the perfect time to make some tweaks to my stereo system.

For a few months now I've been listening to my sound system, primarily an Onkyo TX-8211 receiver/amp, A Behringer DSP8024 equalizer, and some Pioneer speakers, with only a very "basic" euqlization (or "EQ") curve entered into it. And to explain that, I guess I'd have to flash back to a few months ago for a little while.

I have had basically this system since early 2002, when my previous receiver/amp gave out (it appears its anti-impedance-mismatch circuitry1 is "unhappy," including possibly a broken relay that controls switching on the speakers...end result, no relay click, and no sound).  I don't know what parts in particular I have trouble with, whether it's the amps I choose, or the speakers, or just plain buying everything "on the cheap," but I usually end up with something sounding "hollow..." really overly rich in the midrange I guess.  It's something I've nicknamed "the Kraco effect" after the less-than-stellar car speakers brand which were really lacking in good lows and highs.  So I've more-or-less always used some sort of equalization to counteract that, and have what subjectively sounds like more even sound.

The Behringer manual states there is a battery (actually a cell) inside the equalizer (or "EQ") which keeps the settings, response curves, etc. when the power is off, and it should last several years.  Well, "several" was about "nine" in this case, which is pretty darned good.  And finally, earlier this year, a "WARNING: battery low" message showed on the display on powerup/boot.2  Behringer's recommendation is to ship it off to them to replace this so that nothing is lost.  After all, it has 100 EQ curve memories (31 bands each) and dozens of settings, so yes, if I would have done the smarter thing (and either figure out how to back them up over its MIDI port, or at least note down my settings), I wouldn't have caused myself all the vexation.  I wasn't about to spend the money to ship this thing to them and wait for it to come back.  And I don't have a MIDI interface on my 'puter either, if it's even possible to read those sorts of things out of the device.

So, I took it out of the stereo stack, and for the first time ever, I opened up the case some months ago to see what was what inside this wonderful machine.  It turns out it has a CR2032 cell, like the vast majority of computers.  Some time ago, perhaps a year or so, I had gotten a whole card of CR2032 cells at a discount store (Ollie's Bargain Outlet), so I had plenty to spare.  I had it all ready to go, on my living room floor.  And I swapped that thing just about as fast as I could, and I figured the CMOS charge would hold in the RAM through a second or three.  I figured it would probably be unwise to try and swap it with the unit powered up (mainly because I might slip up and short out something with the metal case of the new or old cells for example), so I had it unplugged.  As far as the grief it subsequently caused me, I think maybe I should have attempted this powered up.  When I did power it back on, I saw a message "battery empty: memory cleared."  (whimper...)  Oh, well...since I don't even know someone with a TARDIS so that I could go back in time and do it all differently, I figured there's nothing I can really do except move forward.

So, I buttoned the case back up, put the EQ back in the stereo stack, and I entered in a very basic "anti-Kraco-effect" curve.  But it still sounded a little hollow.  What would really be required was to use its real-time analyzer (RTA) and signal generator to at least start with a flat response for the equipment.  This is done by connecting a microphone (mic) to the EQ's mic input, turning on the signal generator, and having the RTA figure out the response of the amp, speakers, room acoustics, etc.  But that's complicated.  One has to set up a mic who's response characteristics are known (like this one for the Audio Technica Pro2ax that I have):

entering the opposite of that curve into the EQ (a compensation curve), routing the EQ through the system so as much of the system is tested as possible, and finally telling the EQ to take a snapshot at some time and plopping the inverse of that response curve (again, a compensation curve) into one of its memories.

The level captured by the mic is simply too low, even with the EQ's mic gain all the way up, so consequently the amp basically has to be cranked way up in order to get the compensation curve centered somewhere close to zero.  When I was in south Buffalo, I did this, no problem.  This time around though, not so much joy.  The room the system is in now is considerably smaller than where it was before the "big move," and probably concentrated the sound so much that it was uncomfortable to perform the testing.  Thankfully I have some muffs I wear when using power tools and such.

So yesterday, I finally had enough of listening to this Kraco effect, and took some time to enter the inverse of that curve above into one of the memories.  Since it was so dreary outside, this seemed like an ideal day to finally get this done.  And I put on the muffs, cranked it up for a while, and was watching the display.  Muffs of course don't totally block out all sound, so I could hear the white and pink noise during the test.

But suddenly, the noise sounded very wrong, like the noise generator suddenly went wrong somehow, so I just stopped the test.  And I thought when I started it up again it sounded OK.  I finished the test, and I had a white noise "anti-curve,"  and a pink noise "anti-curve."  The pink one seemed to have less of the Kraco effect, so I began with that.

In about a half hour, I had something I thought sounded a lot better than "straight through" (unequalized), almost totally cancelling the Kraco effect.  I listened to WGRF-FM for a couple of songs, and it seemed OK.  Later that night though, I started watching a recording from my MythTV, and as usual, to get really good stereo separation, I put on some headphones (and I left the speakers on to give some subsonic feeling).

"Hmm...this is kinda 'bright'," I thought.  Well..sure enough, I took a closer look at the EQ curve, and there is a steep rise at 2KHz, with a relatively flat (but amplified) response through 16K, then almost up to maximum for the remaining two bands.  I remembered through experience that anything 16K or above has to be almost completely attenuated, or else it seems to start interfering with the stereo subcarrier of my wireless headphones (keeps toggling between stereo and mono).  So those were sharply lowered.  My ears are so aged these days that they can't tell much of a difference anyway at frequencies that high.  I can still just barely hear the flyback transformer of an NTSC monitor, whereas a decade ago I could tell if one was on or off just by listening for just a sec.

It then dawned on me...I must have blown both my left and right tweeters. I must have turned up the gain so much, and put so much power in the high end through the speakers that they just "imploded." I confirmed this today by connecting one of the speakers that I used previously3, and sure enough, it had an unbalanced, tinny sound. Now when I'm listening w/o 'phones, I realize the sound is very lacking in the high end, with very little "detail" to the sound.

Now...here's where the disillusion part begins.

If the tech is working like it should, the compensation curve arrived at through "AUTO-EQ" as Behringer calls it all the higher bands should be "on the roof" (in this case, at +16dB), trying to compensate for the tweeters "not being there?"  Shouldn't the midrange element be able to compensate by putting out more treble?  (Well, maybe not...especially if there is a crossover network in the speaker enclosure.)

Admittedly, all my electronic and acoustic knowledge is self-taught, so I could be missing quite a bit.  And my testing/calibrating methodology might be totally skewed.  I don't know.  But the subjective results say the tech is all screwed up.

I guess for the most part, it doesn't matter that much.  The majority of my listening is talk radio anyway.  And for that, the other EQ I have (a  7 band Sony) sharply cuts off high and low, with most of the energy in midrange where voice is.  And for that input on my mixing board, I put the bass knob "on the floor."  Still...I can see a trip to The Stereo Advantage for some replacement speakers in my future.

1Power amplifiers want to be loaded with the proper impedance (like audio amps with speakers  for example), or they will damage themselves.  It has to do with efficiency of power transfer, the fact that with no load, the power developed on the output of the final transistors would be "reflected back" into the circuit because it "has nowhere to go," thus basically melting itself.  I don't know how to design it, but there is a circuit which was developed which can detect this impedance mismatch/power reflected condition, and shunt the output into a loading resistor instead so that the power "has someplace to go."  In that case, the loading resistor has the same impedance as speakers would, and dissipates the audio power as heat (as any resistive element would do with power applied).

2Yes, this EQ takes about 10 seconds to initialize itself after poweron, so I would say, "boot."

3I don't use these speakers much anymore because there's some sort of rattle/buzz when the volume is up some.  I think either there's a tear I can't quite see, or perhaps the cone is beginning to separate.  I only use them for speech/talk...more elsewhere in this post on that.

English is a difficult enough language to interpret correctly when its rules are followed, let alone when the speaker or writer chooses not to follow those rules.

"Jeopardy!" replies and randomcaps really suck!

Please join me on Google+!