WEBVTT
Kind: captions
Language: en
Style:
::cue(c.cyan) { color: cyan; }
::cue(c.yellow) { color: yellow; }
##

1
00:00:00.200 --> 00:00:02.020
This video is sponsored by
CuriosityStream

2
00:00:02.020 --> 00:00:04.540
and my own Nebula original series.

3
00:00:04.540 --> 00:00:05.860
More on that later.

4
00:00:05.860 --> 00:00:09.340
In 2019, a load of music videos
that had been uploaded to YouTube

5
00:00:09.340 --> 00:00:12.100
in the early days of the platform
were remastered.

6
00:00:12.100 --> 00:00:14.380
YouTube let the music labels keep
the same video IDs,

7
00:00:14.380 --> 00:00:15.800
along with all the views and comments,

8
00:00:15.800 --> 00:00:18.360
so suddenly,
all those old videos were in HD!

9
00:00:18.360 --> 00:00:19.480
And they look...

10
00:00:19.480 --> 00:00:21.160
well, it depends on the video.

11
00:00:21.160 --> 00:00:24.240
Smash Mouth's <i>All Star</i> looks awful.

12
00:00:24.240 --> 00:00:26.100
Now, there are some brilliant
remasters out there.

13
00:00:26.100 --> 00:00:30.160
Wham!'s <i>Last Christmas</i> is now in 4K,
and it looks spectacular.

14
00:00:30.160 --> 00:00:32.780
And that's because it was shot on film,
full cinema quality,

15
00:00:32.780 --> 00:00:36.060
the sort of thing you can literally project
on the side of a building

16
00:00:36.060 --> 00:00:37.540
and it will look great.

17
00:00:37.540 --> 00:00:41.260
Same with Freddie Mercury's <i>Living On My Own</i>,
the remaster looks stunning.

18
00:00:41.260 --> 00:00:42.800
In fact, I think that looks better,

19
00:00:42.800 --> 00:00:45.380
because they haven't converted it
to widescreen:

20
00:00:45.380 --> 00:00:49.260
they've kept it at the "aspect ratio"
of old '80s television.

21
00:00:49.260 --> 00:00:52.260
Because music videos
weren't meant for the cinema.

22
00:00:52.260 --> 00:00:55.780
They were meant for television.
It's in the name: music video.

23
00:00:55.780 --> 00:00:57.620
So one of the first steps in editing

24
00:00:57.620 --> 00:01:00.360
will have been putting the film
through a telecine,

25
00:01:00.360 --> 00:01:03.420
which is a very fancy way of
pointing a TV camera at a film projector

26
00:01:03.420 --> 00:01:05.220
and saving the results to videotape.

27
00:01:05.220 --> 00:01:09.020
It is way more precise than your average
movie-theatre pirate with a camcorder,

28
00:01:09.020 --> 00:01:13.320
but the result is that all that film quality
gets crushed down

29
00:01:13.320 --> 00:01:16.960
onto a standard definition TV picture
stored on magnetic tape.

30
00:01:16.960 --> 00:01:21.520
480 horizontal lines in the US,
576 in Europe. That's it.

31
00:01:21.520 --> 00:01:23.860
Any more quality than that is gone.

32
00:01:23.860 --> 00:01:26.300
Editing on videotape is
much cheaper and easier, though,

33
00:01:26.300 --> 00:01:29.660
and if you're putting the final version
out on television, well...

34
00:01:29.660 --> 00:01:32.120
why would you bother with
editing actual physical film,

35
00:01:32.120 --> 00:01:33.840
which is what you had to do
back then?

36
00:01:33.840 --> 00:01:38.680
A few videos might have gone to that expense,
if it was part of a big concert film, say.

37
00:01:38.680 --> 00:01:42.940
So the <i>Rolling Stones Rock and Roll Circus </i>
was played at the New York Film Festival,

38
00:01:42.940 --> 00:01:46.140
and you can see the cinema camera
with its big, bulky rolls of film

39
00:01:46.140 --> 00:01:48.200
in the background of
some of the shots.

40
00:01:48.200 --> 00:01:53.860
But the final version of
most music videos were, well, videos.

41
00:01:53.860 --> 00:01:55.540
So for the remaster of <i>Last Christmas,</i>

42
00:01:55.540 --> 00:01:59.100
a team of engineers had to track down
the original reels of film,

43
00:01:59.100 --> 00:02:01.920
and rescan them using
modern technology at 4K,

44
00:02:01.920 --> 00:02:04.560
and then, they painstakingly took
those 4K files

45
00:02:04.560 --> 00:02:07.620
and recreated every edit
in the original video,

46
00:02:07.620 --> 00:02:10.100
frame for frame,
cut for cut.

47
00:02:10.100 --> 00:02:13.620
Actually, they were only able to find
seven out of the eight original reels of film,

48
00:02:13.620 --> 00:02:15.580
so there are a few shots in there

49
00:02:15.580 --> 00:02:19.280
where the only source material was
the TV-quality music video.

50
00:02:19.280 --> 00:02:20.500
And you can tell,

51
00:02:20.500 --> 00:02:23.940
'cause this is what an old-school, standard-definition,
analogue signal looks like

52
00:02:23.940 --> 00:02:28.240
when you use world-class technicians
to upscale it to 4K.

53
00:02:28.240 --> 00:02:30.780
It's not too bad? I guess?

54
00:02:30.780 --> 00:02:32.500
<i>Living On My Own</i>
also has a couple of shots

55
00:02:32.500 --> 00:02:35.980
where the old '80s digital effects
have to be recreated with modern tech,

56
00:02:35.980 --> 00:02:38.960
and... they're not a perfect match.

57
00:02:38.960 --> 00:02:41.460
And there are also a couple of shots
that have been replaced,

58
00:02:41.460 --> 00:02:44.780
or the timing slightly changed,
no idea why.

59
00:02:44.780 --> 00:02:47.020
But again, it's not a bad job.

60
00:02:47.020 --> 00:02:49.480
Now, there is a cheaper
remastering option

61
00:02:49.480 --> 00:02:52.340
that doesn't involve recreating
the whole edit from scratch:

62
00:02:52.340 --> 00:02:54.160
you can take the final TV edit

63
00:02:54.160 --> 00:02:58.140
from the best, cleanest, professional-grade
videotape you can find,

64
00:02:58.140 --> 00:03:00.040
and upscale that to HD.

65
00:03:00.040 --> 00:03:01.800
A lot of remastered videos
have done this.

66
00:03:01.800 --> 00:03:04.060
It won't be as sharp as
the original film,

67
00:03:04.060 --> 00:03:06.380
but it will look better than
people are used to.

68
00:03:06.380 --> 00:03:10.700
Which is what they've done with
the remastered version of <i>Bohemian Rhapsody.</i>

69
00:03:11.880 --> 00:03:12.520
And look,

70
00:03:12.520 --> 00:03:16.720
I can see why record companies would want
an HD version of <i>Bohemian Rhapsody.</i>

71
00:03:16.720 --> 00:03:20.280
It's iconic.
It has over a billion views on YouTube.

72
00:03:20.280 --> 00:03:22.740
But this was not shot on film.

73
00:03:22.740 --> 00:03:25.640
It was recorded on TV cameras,
edited on videotape—

74
00:03:25.640 --> 00:03:27.100
you can tell by
the look of the lighting—

75
00:03:27.100 --> 00:03:30.960
and by the fact that these effects required
live feedback to the camera.

76
00:03:30.960 --> 00:03:34.980
you could only get this by pointing
a television camera directly at a monitor,

77
00:03:34.980 --> 00:03:38.180
which means there was never
any high-quality film version,

78
00:03:38.180 --> 00:03:41.300
this was only ever in standard definition.

79
00:03:41.300 --> 00:03:45.700
And remember, old-school analogue television
in Europe runs at 50 frames a second.

80
00:03:45.700 --> 00:03:47.920
60 in the US,
but 50 in Europe.

81
00:03:47.920 --> 00:03:51.760
Half of the horizontal lines get skipped
each time, which is called interlacing.

82
00:03:51.760 --> 00:03:54.720
And it means you get less flickering
on old equipment.

83
00:03:54.720 --> 00:04:00.840
So if you want the authentic, original experience,
as seen back in the '70s,

84
00:04:00.840 --> 00:04:04.000
then <i>Bohemian Rhapsody</i> should be played
at standard definition,

85
00:04:04.000 --> 00:04:05.760
at 50 interlaced frames a second.

86
00:04:05.760 --> 00:04:07.780
That's what was recorded.

87
00:04:07.780 --> 00:04:11.280
Instead, what you see on YouTube is
a computer's best effort

88
00:04:11.280 --> 00:04:16.560
to make a high-definition, 25 frame a second,
film-like picture from that limited data,

89
00:04:16.560 --> 00:04:19.960
filling in the gaps,
turning jagged edges into smooth lines.

90
00:04:19.960 --> 00:04:23.760
That's called 'deinterlacing'
and it's not done badly here.

91
00:04:23.760 --> 00:04:28.120
'Remastering' doesn't mean 'restoration'.
It means 'recreation'.

92
00:04:28.120 --> 00:04:32.380
You'll see marketing blurbs saying
remasters are "the way it was meant to be seen".

93
00:04:32.380 --> 00:04:35.000
Which is a great tagline, but...

94
00:04:35.000 --> 00:04:36.060
no, it wasn't.

95
00:04:36.060 --> 00:04:39.900
Remastering is, to an extent,
changing the historical record.

96
00:04:39.900 --> 00:04:42.340
There are fan remasters of
Bohemian Rhapsody out there,

97
00:04:42.340 --> 00:04:47.880
upscaled from VHS, 50 frames a second,
that are much more true to the original.

98
00:04:47.880 --> 00:04:51.960
But at least Bohemian Rhapsody got
a decent HD remaster.

99
00:04:51.960 --> 00:04:54.200
It's not like what they've done
to Smash Mouth.

100
00:04:54.200 --> 00:04:56.420
Look at his face in this scene.

101
00:04:56.420 --> 00:04:59.800
That's not a clear high-definition picture,
that's some kind of demon child!

102
00:04:59.800 --> 00:05:03.140
It's scaled up wrongly,
it's over-sharpened.

103
00:05:03.140 --> 00:05:07.080
Look at the diagonal lines here,
they're jagged and stepped.

104
00:05:07.080 --> 00:05:11.420
I tracked down a standard-definition version
from years ago that someone had reuploaded,

105
00:05:11.420 --> 00:05:15.020
and... yeah, I can see
why a computer would do that.

106
00:05:15.020 --> 00:05:18.460
And, oh yeah, it's badly deinterlaced
in some parts too,

107
00:05:18.460 --> 00:05:23.040
because here, you can see the cheerleader is
in two places at once in the same frame.

108
00:05:23.040 --> 00:05:25.760
It's not bleed-over
from the next frame either.

109
00:05:25.760 --> 00:05:27.820
She's in two different places there as well.

110
00:05:27.820 --> 00:05:30.280
Same for this scene with the dancer.

111
00:05:30.280 --> 00:05:33.200
Why is there a dancer in just one scene
in this music video?

112
00:05:34.200 --> 00:05:38.200
But those deinterlacing artifacts,
I don't understand how they're even possible.

113
00:05:38.200 --> 00:05:41.420
Did they shoot some of this
on a TV camera at 60 frames a second

114
00:05:41.420 --> 00:05:43.320
and then grade it to look like film?

115
00:05:43.320 --> 00:05:44.620
That doesn't make any sense,

116
00:05:44.620 --> 00:05:47.220
but that's the only way
that those artifacts could happen.

117
00:05:47.220 --> 00:05:50.760
But other parts look fine, and look like
they're on film at 30 frames a second!

118
00:05:50.760 --> 00:05:54.780
And the parts that they pulled from
the movie that it was a tie-in to — <i>Mystery Men</i> —

119
00:05:54.780 --> 00:05:56.040
those run at 24 frames a second,

120
00:05:56.040 --> 00:05:59.500
and you can see in this transition, the right side
is badly deinterlaced with a ghost image

121
00:05:59.500 --> 00:06:02.440
and the left side has every fifth frame doubled
so the frame rates match.

122
00:06:02.440 --> 00:06:03.440
I— I—

123
00:06:03.440 --> 00:06:07.000
I genuinely cannot understand
what has happened here.

124
00:06:07.000 --> 00:06:09.520
It is a complete and utter mess.

125
00:06:09.520 --> 00:06:13.100
Presumably some underpaid and overworked
VFX tech was given

126
00:06:13.100 --> 00:06:16.840
a low-quality, already-broken
copy of it and told:

127
00:06:16.840 --> 00:06:19.940
"It's just Smash Mouth, we don't care.
Just make it work."

128
00:06:19.940 --> 00:06:22.600
So they threw the video into a
basic resizer to make it HD,

129
00:06:22.600 --> 00:06:24.920
massively dialed up the sharpness,

130
00:06:24.920 --> 00:06:28.280
and tried to squeeze the aspect ratio
into... whatever this is.

131
00:06:28.280 --> 00:06:31.860
The HD remaster looks worse
than the one it replaced.

132
00:06:31.860 --> 00:06:35.260
It's sharper, sure,
but it's not better.

133
00:06:35.260 --> 00:06:37.580
And if you are going to change
the historical record,

134
00:06:37.580 --> 00:06:41.340
it does seem a terrible option
to make it worse.

135
00:06:41.340 --> 00:06:44.620
I've got an original series over on Nebula.
Here's the trailer.

136
00:06:44.620 --> 00:06:47.560
<i>I invited five people to play some games.</i>

137
00:06:47.560 --> 00:06:49.580
<c.yellow>I trust no one.
<c.cyan>None of us are trustworthy.

138
00:06:49.580 --> 00:06:52.600
<i>...in an environment designed to
slowly break their team apart.</i>

139
00:06:52.600 --> 00:06:54.140
<c.yellow>This is real money!

140
00:06:54.140 --> 00:06:56.660
<i>But all they knew is
they'd be sat around a table</i>

141
00:06:56.660 --> 00:06:59.880
<i>trying to win real cash: $10,000.</i>

142
00:06:59.880 --> 00:07:02.020
<c.yellow>The vibe's changed after that theft.

143
00:07:02.020 --> 00:07:05.440
<i>This is a show about trust,
about loyalty,</i>

144
00:07:05.440 --> 00:07:06.380
<i>and about Money.</i>

145
00:07:06.380 --> 00:07:09.540
<c.yellow><i>Tom wants the chaos.</i>
<i>(all laugh)</i>

146
00:07:09.540 --> 00:07:12.480
Nebula is a home for new,
in-depth and experimental content

147
00:07:12.480 --> 00:07:14.940
and collaborations from
education video creators

148
00:07:14.940 --> 00:07:16.520
that you may well have heard of.

149
00:07:16.520 --> 00:07:19.440
and it's bundled for free
at this link with CuriosityStream,

150
00:07:19.440 --> 00:07:22.460
a subscription streaming service
with thousands of

151
00:07:22.460 --> 00:07:25.820
big-budget, professional documentaries
and nonfiction titles.

152
00:07:25.820 --> 00:07:29.820
CuriosityStream is $2.99/month
or only $19.99/year.

153
00:07:29.820 --> 00:07:33.540
You can get a 30-day free trial,
including access to Nebula and my series,

154
00:07:33.540 --> 00:07:37.060
by going to curiositystream.com/tomscott.

155
00:07:38.060 --> 00:07:41.600
[Caption+ by JS*
https://caption.plus]