WEBVTT
Kind: captions
Language: en

1
00:00:00.080 --> 00:00:02.560
Last week, I put up a video that was filmed
here

2
00:00:02.560 --> 00:00:04.959
at the Cambridge Centre for Computing History.

3
00:00:04.959 --> 00:00:07.933
And folks seemed to like the video, but a
few comments made it clear

4
00:00:07.940 --> 00:00:09.420
that there was a slight problem.

5
00:00:09.420 --> 00:00:12.720
And it was a problem that made me
really angry with myself.

6
00:00:12.720 --> 00:00:16.180
There was a very subtle high-pitched noise
in the background of the video,

7
00:00:16.180 --> 00:00:18.500
and I hadn't noticed it before uploading.

8
00:00:18.500 --> 00:00:19.820
So I thought I should explain why:

9
00:00:19.820 --> 00:00:22.350
first, why the noise was there at all,

10
00:00:22.350 --> 00:00:25.380
and second, why I didn't notice it.

11
00:00:25.380 --> 00:00:30.800
This museum is filled with old computers with
cathode ray tube screens: CRTs.

12
00:00:30.800 --> 00:00:32.689
Some of these screens are
specialised computer monitors,

13
00:00:32.689 --> 00:00:36.180
but others, like these over here, are basically
just old-school televisions.

14
00:00:36.180 --> 00:00:38.890
At the back of each monitor
is an electron gun,

15
00:00:38.890 --> 00:00:40.850
and it fires out a beam of electrons.

16
00:00:40.850 --> 00:00:44.270
Wherever they hit,
the phosphors on the screen light up.

17
00:00:44.270 --> 00:00:45.930
But the electron gun doesn't move at all.

18
00:00:45.930 --> 00:00:47.460
It couldn't possibly move fast enough.

19
00:00:47.460 --> 00:00:49.860
Instead, it just fires in one direction,

20
00:00:49.860 --> 00:00:52.879
and electromagnets bend the beam of electrons.

21
00:00:52.879 --> 00:00:55.980
The more power that's sent to the magnets,
the more the beam bends.

22
00:00:55.980 --> 00:00:59.360
So for every frame of video,
that beam scans the whole screen,

23
00:00:59.360 --> 00:01:01.420
left to right, top to bottom.

24
00:01:01.420 --> 00:01:03.030
I'm going to ignore the vertical movement,

25
00:01:03.030 --> 00:01:05.370
what we're interested in is the horizontal.

26
00:01:05.370 --> 00:01:09.090
Inside there is a clever bit of kit called
a flyback transformer.

27
00:01:09.090 --> 00:01:11.310
It outputs a sawtooth signal to the electromagnet,

28
00:01:11.310 --> 00:01:15.049
which means that it ramps up, and then drops
very quickly back down.

29
00:01:15.049 --> 00:01:18.859
And as it ramps up, the electromagnet bends
the beam across the screen

30
00:01:18.859 --> 00:01:22.409
before it quickly snaps straight back ready
for the next line.

31
00:01:22.409 --> 00:01:28.139
That electron beam scans across 625 lines
of the TV signal, 25 times a second,

32
00:01:28.139 --> 00:01:35.319
so that flyback transformer needs to
ramp up and drop down 15,625 times a second.

33
00:01:35.319 --> 00:01:38.380
And because of something called magnetostriction,

34
00:01:38.380 --> 00:01:43.090
it changes shape a tiny, tiny, tiny amount
every time it does that,

35
00:01:43.090 --> 00:01:46.419
which generates a noise
at that high-pitched frequency.

36
00:01:46.419 --> 00:01:49.649
It's the same frequency that
every old-school TV produces.

37
00:01:49.649 --> 00:01:50.649
At least in Europe.

38
00:01:50.649 --> 00:01:52.060
The numbers in America are a little different,

39
00:01:52.060 --> 00:01:54.259
but they still end up at about the same frequency.

40
00:01:54.259 --> 00:01:56.419
Now, with modern audio editing tools

41
00:01:56.419 --> 00:02:00.680
it is really easy to isolate that frequency
and remove it.

42
00:02:00.680 --> 00:02:04.270
Until now, I've been removing it
from this video, but let me tell you:

43
00:02:04.270 --> 00:02:08.360
it's here, and in this room,
with all these monitors on, it is loud.

44
00:02:10.300 --> 00:02:12.060
Now I knew the sound was recorded originally,

45
00:02:12.069 --> 00:02:15.670
because I could hear it in this room, I knew
my microphone was picking it up.

46
00:02:15.670 --> 00:02:19.700
I just thought the automatic noise reduction
stuff I use would take care of it.

47
00:02:19.700 --> 00:02:23.500
It didn't: but it also turns out these headphones
that I use to check audio

48
00:02:23.500 --> 00:02:25.310
aren't great at high frequencies,

49
00:02:25.310 --> 00:02:29.379
and so with my voice drowning it out I just
didn't notice the sound.

50
00:02:29.379 --> 00:02:32.980
There aren't many quiet bits in that video,
I don't shut up.

51
00:02:32.980 --> 00:02:36.000
And for a while, I was angry with myself,
because it was an unforced error.

52
00:02:36.000 --> 00:02:39.349
I wasn't under time pressure,
I checked through half a dozen times,

53
00:02:39.349 --> 00:02:43.310
I just didn't think to check the sound
with my eyes.

54
00:02:43.310 --> 00:02:47.840
That mistake might have stopped the message
getting as far as it could.

55
00:02:47.840 --> 00:02:50.370
But I think there was another reason
that I was angry.

56
00:02:50.370 --> 00:02:53.580
Spider Robinson once wrote that
"anger is fear in disguise",

57
00:02:53.580 --> 00:02:56.150
and I'm not sure that's always true, but...

58
00:02:56.150 --> 00:02:58.540
it wasn't just my headphones.

59
00:02:58.540 --> 00:03:01.159
As you get older, your range of hearing 
starts to narrow.

60
00:03:01.159 --> 00:03:03.890
It starts to get more difficult
to hear high frequencies.

61
00:03:03.890 --> 00:03:06.890
And I used to pride myself
on how good my hearing was,

62
00:03:06.890 --> 00:03:10.409
I could hear higher frequencies than I was
supposed to for my age.

63
00:03:10.409 --> 00:03:14.859
Those 17,000Hz mosquito alarms designed to
make teenagers go away,

64
00:03:14.859 --> 00:03:16.730
they irritated me through my twenties.

65
00:03:16.730 --> 00:03:21.010
And, sure, I can hear the loud, high-pitched
noise here, 'cos it's really obvious,

66
00:03:21.010 --> 00:03:23.170
but in that video?

67
00:03:23.170 --> 00:03:24.170
Nope.

68
00:03:24.170 --> 00:03:27.159
I have to strain to even notice it, even as
some folks in the comments

69
00:03:27.159 --> 00:03:29.909
said it was loud enough to stop them watching.

70
00:03:29.909 --> 00:03:31.560
I did a hearing test.

71
00:03:31.560 --> 00:03:34.040
I can't hear mosquito alarms any more.

72
00:03:34.040 --> 00:03:36.569
I can't hear anything higher than these monitors.

73
00:03:36.569 --> 00:03:40.000
That video, that mistake I made,

74
00:03:40.000 --> 00:03:44.110
was a sudden and sharp reminder
that I'm getting older.

75
00:03:44.110 --> 00:03:48.439
That some of the things I used to do,
I can't any more.

76
00:03:48.439 --> 00:03:50.209
So protect your hearing,

77
00:03:50.209 --> 00:03:52.769
and next time someone forgets to check for

78
00:03:52.769 --> 00:03:55.629
a high pitched noise in the background and
it annoys you...

79
00:03:55.629 --> 00:03:58.279
well, at least you can still hear it.