Rendered at 18:21:36 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
BobMcBob 21 hours ago [-]
Tech Tangents is one of the best retro channels on youtube but by retro I dont mean glorified nostalgia either. Shelby puts a lot of work into his videos and likes to showcase what awesome engineering went into some of the early tech that was practically magic. Love the channel and glad to see it on HN.
bcraven 21 hours ago [-]
[flagged]
Nux 21 hours ago [-]
[flagged]
geon 6 hours ago [-]
As I understand it, this only works with still images that scroll vertically.
Each revolution is one frame, so if you compare the data paths next to each other, they don't make up an image but the same single line of several consecutive frames.
Scrolling a still image makes the same line on the screen cover a different part of the image each frame, so you can sort of make out what the original image looked like.
The end credits should show up as a single tall image, since the only limit to the height is the radius of the disc.
harrall 1 hours ago [-]
If I’m not mistaken, it also doesn’t work with LaserDiscs… it works only with CEDs.
The title is a bit misleading. LaserDiscs are digital but CEDs are analog.
Very different technologies.
NoMoreNicksLeft 57 minutes ago [-]
I was under the impression that laserdisc encoded the video signal in analog. Some (all?) audio tracks were digital/pcm, however.
itisit 30 minutes ago [-]
LaserDiscs are not quite a digital medium. The video is an FM analog signal (composite). Digital PCM audio was introduced about half a decade into the format's life.
sixothree 4 hours ago [-]
Great points.
And it just so happens in this video that the scrolling was at a speed where everything was preserved and also not stretched. Too slow and it would be elongated vertically on the disc and too fast and it would be compressed and you'd miss information.
throwaway27448 13 hours ago [-]
Is there a version that doesn't require watching a video please? This would be 10x faster and easier as a text blob
nilslice 12 hours ago [-]
yes, you must save it to a laserdisc, and then observe it though a highly magnified digital microscope looking for a specific frame.
Long answer: Just jump around. Once you get to the last 1/3rd of the video, there's a lot of close ups of the part of the laserdisc that recorded the credit sequence. The on-screen (laserdisc) text is clearly visible.
The author moves a flashlight around to show how the angle is important; something that won't come across as well in a blog entry.
3 hours ago [-]
imtringued 11 hours ago [-]
You can ask your human to watch the video and write the text blob for you.
comboy 3 hours ago [-]
Just post it to social media to generate the text description for free (multiple humans in parallel work better)
FrozenSynapse 10 hours ago [-]
[flagged]
bonoboTP 9 hours ago [-]
It's likely based on just the transcript, even if it describes visual things, it likely guesses those things from the transcript text only.
Maybe it's better now, but that was how it did it recently. To be convinced that it "watches" the video, I would need to see evidence of it referring to facts that are strictly only possible to know from the video, but not guessable from the audio.
ilija139 6 hours ago [-]
You can try it with your own recorded video. I record myself doing exercises and Gemini gives me really good feedback on my form.
ptman 10 hours ago [-]
use gemini and ask it to summarize a youtube link
rustyhancock 20 hours ago [-]
Here's a screen capture of the end credits visible on the disc the videos worth it but I do think sometimes you need to start with the money shot
https://ibb.co/v4KK88fF
crazydoggers 19 hours ago [-]
This is from a CED not a laserdisc.
iamtedd 14 hours ago [-]
If you watch the video, he looked at both.
crazydoggers 12 hours ago [-]
Yes, but the end credits mentioned by the parent and the link to the image is the CED. That and the title of this post make it seem like this level of image clarity is from a LaserDisc which its not. I think it's worth being clear.
jmkd 6 hours ago [-]
Watch the video if you believe it's worth being clear. The credits are perfectly visible on both the LaserDisc AND the CED. Approx 22:00-24:00 is the laser disc and 25:00 onwards is the CED. Enjoy.
sixothree 4 hours ago [-]
And to support your point even further, in the video he captures the credits off a disc with constant angular velocity, but he _also_ completely lucks out and captures an image from a disc with constant linear velocity.
I'm so sick of this negative attitude. I get it when it comes to politics or more complex systems or conceptual ideas. But holy hell, we're talking about a "money shot" to get people interested in the subject.
jmkd 5 hours ago [-]
Wow. Downvoted for factual clarification and helpful citation. No idea what drives such people, but must logically assume deception and obfuscation.
irl_zebra 4 hours ago [-]
Ugh, I'm pretty sure it's Big CED deceptively here to muddy the waters and obfuscate superior LaserDisc Technology. Perhaps flag the post so the moderators can see it and make sure we root Big CED out of these forums and out of our lives for good.
jmkd 4 hours ago [-]
I'm making placards now. Meet you on the corner of Pioneer and 78th st.
K0balt 3 hours ago [-]
I’m in.
VorpalWay 20 hours ago [-]
The live stream of this had more interesting things as well, such as looking at the ink on mimeographs compared to inkjet printing. Long and rambly as live streams tend to be, but it is there if anyone cares.
I'm not that familiar with CED but the fact that we can see the images with microscopes is because these are analog discs? And that was because computing power back then was non-existent so they didn't use any kind of compression?
grishka 4 hours ago [-]
This is 1970s tech, so of course there was no compression. It's basically a composite video signal etched into the medium. Every revolution of the disc is one full frame (or two fields). So when you look at it, the adjacent grooves represent the same lines of video in different frames.
Now imagine if you picked a Y coordinate and extracted the corresponding row of pixels from every frame of a video file, and stacked them vertically. If there was scrolling text, or if the camera moved vertically, you would see a meaningful image, the same way a scanner obtains a complete image by moving a single-row sensor across whatever you're scanning. This is the same effect, just arising from the way the video signal is laid out on the disc.
gf000 11 hours ago [-]
The key point is that there has to be a slow vertical panning happening as actual content. If that happens, then the on-disk representation of a color channel can end up physically below/above what happens before/after in the movie, drawing out the "actual content". This is why end credits were the most likely visible artifacts.
One other important aspect is that by changing the angle of lighting, he could basically filter out data at a relevant wavelength.
--
At least that's what I got from the video.
ErroneousBosh 9 hours ago [-]
There wasn't any computing involved. The video was recorded as a sampled analogue signal, with pits of varying length setting the "output voltage".
If you look at a Constant Angular Velocity disc you can actually see "spokes" radiating out from the centre, with two broad ones 180° apart. The narrow spokes are the horizontal sync pulses occuring every 0.576° - the disc rotates at 25 revs per second and each concentric track is one complete frame. The broader spokes are of course the vertical sync pulses and colour burst occurring every 1/50th of a second.
If you're in the US or Japan, these numbers are 30 revs per second, 0.686° and 1/60th of a second, because of the lower resolution video standard, but it doesn't look like Laserdisc was much of a "thing" in those countries.
Here in the UK, in the 1980s all the schools took part in a thing called "The Domesday Project" [1] - the name is a reference to The Domesday Book, a survey of England and Wales carried out in the 11th century by William the Conqueror.
The Domesday Discs were CAV Laserdiscs that were played in a special player with a SCSI interface, attached to a BBC Micro computer. Because each concentric track was a complete frame it was possible to get perfect still frame video by just keeping the head still, so you could look at photos of places all around the UK and read a bit of information about them genlocked over the top.
Uncompressed digital encoding might still result in recognizable structures, but probably not as nicely as here.
throwaway85825 11 hours ago [-]
Only the CD had uncompressed data.
happycube 5 hours ago [-]
CD's have repatterened uncompressed data using scrambling codes to keep the waveform stable (i.e. ~50% on/off)
And a Laserdisc with digital sound literally has that CD audio EFM waveform in the lower frequencies.
brazzy 8 hours ago [-]
CED, not CD. And LDs most definitely also have uncompressed data, resulting in visible text at 22:30 in the video.
actionfromafar 7 hours ago [-]
This is where we get into semantics, but it sort of has compressed color data, in the analog domain. :)
throwaway27448 13 hours ago [-]
This is a distinction without meaning: all digital anything is analogue if you look closely enough
> And that was because computing power back then was non-existent so they didn't use any kind of compression?
Compression is not a medium-level detail. You can easily store compressed data on a laserdisc.
roywiggins 11 hours ago [-]
Your CED or laserdisc player needs to be smart enough to be able to decode whatever you put on it, which- in the era that they were relevant- pretty severely limited what you could do.
system2 13 hours ago [-]
No, with digital, you need encoding. How can you even compare binary with embedded images.
throwaway27448 13 hours ago [-]
> How can you even compare binary with embedded images.
How are the images encoded?
system2 60 minutes ago [-]
If you watch the video, you can see that the images are burnt to the medium and can be seen optically. With 1 and 0, you cannot do this. All you'd see would be random dots everywhere.
brazzy 11 hours ago [-]
In ways that ensure that they are not visually recognizable on the physical medium afterwards, because the visual layout represents a whole lot of redundancy, and the job of compression is to remove redundancy. If the end result has any recognizable patterns, the compression is not doing its job well.
csours 19 hours ago [-]
Not nearly as cool, but I was able to show a colleague the letters in a raster image section of a pdf using xxd by varying the output width
EvanAnderson 18 hours ago [-]
I wrote a simple tool, when I was a kid, that dumped binaries into VGA mode 0x13 and allowed me to vary the width. Mode 0x13 is one byte per pixel so it was just a simple REP MOVSB to put data into the buffer (no worrying about bitplanes). It was so useful in reverse engineering software. Besides raster data, regular data structures often jump out.
Retr0id 17 hours ago [-]
Fun fact, if you load a file with extension ".data" into GIMP there's a UI where you can set the pixel format and adjust the width/height with sliders
userbinator 15 hours ago [-]
Photoshop does the same with the .raw extension.
CamelCaseCondo 13 hours ago [-]
I did the same thing to evaluate random number generators by drawing pixels with the count value. You see a pattern, line or clusters? Bad generator.
sixothree 57 minutes ago [-]
I feel like everyone needs a peek inside a file in this manner at least once in their life. I just like to show people various file headers.
Without you showing them, they might never see this kind of stuff.
smusamashah 20 hours ago [-]
So CAV (constant angular velocity) is an encoding format for laser disks. When something is written with CAV, it is basically analogue data and therefore repeating patterns can be recognized on the disk.
wtallis 20 hours ago [-]
No, CAV has nothing to do with encoding, and both analog and digital formats have used both CAV and CLV and hybrids of the two.
The legible text seen in the microscope images happens because of the combination of LaserDisc recording a raw and uncompressed encoding of the analog video signal, the way that LaserDisc used CAV to store an integer number of frames per track so that the image data for corresponding on-screen locations of subsequent frames would be aligned at the same radial position on the disc, and the credits scrolling vertically at constant speed.
If LaserDisc had used a digital encoding (especially a compressed encoding), the data on disc may still have had discernible patterns but the text would not necessarily have been legible. If it had used CAV but not stored a whole number of frames per track, then temporal and spatial locality on screen would not have corresponded so well to spatial locality on disc. And the vertically-scrolling credits are pretty much the only kind of content that can produce the recognizable and legible images on the disc surface.
I think the fact that the aspect ratio of the text came out approximately right probably is a consequence of the scrolling speed of the credits being chosen to suit the vertical resolution of the video. If the text had appeared squished in the microscope, it would probably have been moving too fast on screen to be clearly legible.
xattt 19 hours ago [-]
Is the image seen via microscope basically a readout of the image on a spectroscope?
Those can have near-legible images, but most of the time they are not.
wtallis 2 minutes ago [-]
For the images seen through the microscope, the horizontal axis is the same as the horizontal axis of the video. The vertical axis of what the microscope sees is the time axis of the video: it's essentially a slit camera view of a particular individual scan line position within the frame.
VerifiedReports 16 hours ago [-]
No; it's strictly a reference to constant RPM or variable RPM.
CAV discs contained one frame per rotation. While this meant you could only fit half an hour on one side of a disc, it did give you perfect slow-motion and freeze-frames.
I worked in a video store and loved LaserDiscs. The Duran Duran video album was CAV, and the Pioneer LD-700 had such a fast transport mechanism and remote control that I could to DJ-style "scratching" with it.
jmkd 20 hours ago [-]
Not simply repeating patterns, readable text from the credits as shown in the video.
accrual 20 hours ago [-]
Here's where the text comes into focus, pretty cool:
The data being written to the disk is the same in CAV or CLV disks, but the player just needs to know how to spin the disk at the right speed so that the laser can read the pits/lands correctly. It is purely a detail about the speed that the disk is spun at so they can cram more data on it with CLV disks.
What CAV LaserDiscs allow for, though, is to make it extremely obvious where scanlines and blanking intervals are in the video signal.
ralferoo 19 hours ago [-]
Actually amazing being able to read the text like that, and on two different types of discs. Great video, was much better than I was expecting it to be from the title!
luzionlighting 9 hours ago [-]
Interesting to see how optical technology has evolved. In lighting design we also deal a lot with how light interacts with surfaces, lenses and reflectors.
Even small changes in optics can drastically change how light spreads or how uniform illumination appears in a space.
amelius 20 hours ago [-]
But the opto mechanical parts of a laserdisc reader are way more interesting than a microscope.
9 hours ago [-]
oofbey 18 hours ago [-]
Fun fact about laser discs. They are analogue not digital. CD’s store digital information with the presence or absence of pits. Fairly ancient but still fundamentally feels like a very old version of a thumb drive.
Laser discs are not digital. They encode the analogue video signal’s value as the length of the pit. It is digitized in the time domain - sampled at some frequency, but the “vertical” signal value is stored entirely analogue. In terms of encoding it’s more similar to a VHS tape than a CD. Kinda crazy.
bregma 8 hours ago [-]
> Laser discs are not digital... It is digitized in the time domain
Laser disks are 100% digital (as you said, they store digits in the time domain).
They don't encode their data using binary like a CD does.
"Binary" and "digital" are two separate and unrelated concepts.
actionfromafar 7 hours ago [-]
Um... I think they store "PCM encoded-ish" but the length of the pits are not discrete on / off like on a CD but various arbitrary lengths, so analog.
The sound was also analog to begin with, then the same encoding as CDs, then after that AC-3 and DTS.
a-dub 17 hours ago [-]
yeah i remember learning this as a kid and being surprised. i originally thought laserdiscs were modern high tech, but then they turned out to actually be from the late 70s/early 80s with the primitive analog video encoding where red book audio cds of the mid to late 80s were actually digital.
VerifiedReports 16 hours ago [-]
BUT... Pioneer put AC-3 (Dolby Digital) surround on LaserDiscs before DVDs came out. So LaserDiscs were the first video medium to offer digital sound at home.
And at that point, most players sold were combo players that could also play CDs.
And there was one more disc format: CD Video. It was a CD-sized digital single that also had a LaserDisc section for the (analog) music video. I have a couple; one is Bon Jovi.
throwaway85825 11 hours ago [-]
Was CD video compressed? I thought it existed at the same time as DVD but cheaper.
yourusername 10 hours ago [-]
That's Video CD. It existed before DVD but survived alongside it (mainly in Asia) as a cheaper alternative.
a-dub 3 hours ago [-]
no, apparently there was both. i was familiar with video cd which was mpeg-1 on a cd-rom (with some weird partitioning scheme). cd video is apparently a very obscure hybrid format with an analog video section and a digital audio section. https://en.wikipedia.org/wiki/CD_Video
mh- 17 hours ago [-]
I just learned this in my 40s and am surprised. Very cool.
qsera 12 hours ago [-]
So how does writes work? Does an analog signal translate into pit-lengths with absolute precision?
throwaway290 12 hours ago [-]
Everything is analog when it gets to real world
IAmGraydon 5 hours ago [-]
Just an aside - I love that he's using The Mind's Eye (https://www.imdb.com/title/tt0167285/), which is a video that really captured my imagination when I was young (along with Beyond the Mind's Eye) and changed the course of my life as it got me into 3D animation at a young age.
SV_BubbleTime 16 hours ago [-]
Very cool but, I was hoping he was going to spin it and align with the camera’s refresh rate.
system2 13 hours ago [-]
That would be cool with a variable motor and a 3d printed mount maybe.
Aardwolf 9 hours ago [-]
Now I wonder if something similar is possible with the magnetic fields on VHS tape
grishka 3 hours ago [-]
I thought about that recently while researching how VCRs work before attempting to fix one. I didn't even think about seeing the actual video signal, I was just curious what the diagonal lines and control pulses on the tape look like. There are many other things as well that would be interesting to look at (all kinds of tapes, all kinds of floppies, hard drive platters, magstripe cards), but unfortunately I don't think there exists a technology capable of visualizing magnetic fields with enough precision.
valleyer 8 hours ago [-]
Thoughts on that: (1) you'd need a way to visualize the magnetic fields, (2) the data is frequency modulated, (3) due to helical scan, the video field lines do not line up evenly one over the next as they did so nicely in the Laserdisc / CED (there'd be a skew).
So I don't want to say it's impossible, but I think it would require a lot more creativity.
JohnnyLarue 17 hours ago [-]
[dead]
throwaway85825 11 hours ago [-]
[flagged]
macshome 16 hours ago [-]
[flagged]
ZeWaka 15 hours ago [-]
A LaserDisc was featured in the video alongside the CED video disk.
jmkd 6 hours ago [-]
Both are featured in the video and both have credits visible through the microscope.
Each revolution is one frame, so if you compare the data paths next to each other, they don't make up an image but the same single line of several consecutive frames.
Scrolling a still image makes the same line on the screen cover a different part of the image each frame, so you can sort of make out what the original image looked like.
The end credits should show up as a single tall image, since the only limit to the height is the radius of the disc.
The title is a bit misleading. LaserDiscs are digital but CEDs are analog.
Very different technologies.
And it just so happens in this video that the scrolling was at a speed where everything was preserved and also not stretched. Too slow and it would be elongated vertically on the disc and too fast and it would be compressed and you'd miss information.
writeup from the author linked in the video description
https://youtu.be/qZuR-772cks?si=rYM4EjvV7VeTEzx8&t=1570
Long answer: Just jump around. Once you get to the last 1/3rd of the video, there's a lot of close ups of the part of the laserdisc that recorded the credit sequence. The on-screen (laserdisc) text is clearly visible.
The author moves a flashlight around to show how the angle is important; something that won't come across as well in a blog entry.
Maybe it's better now, but that was how it did it recently. To be convinced that it "watches" the video, I would need to see evidence of it referring to facts that are strictly only possible to know from the video, but not guessable from the audio.
I'm so sick of this negative attitude. I get it when it comes to politics or more complex systems or conceptual ideas. But holy hell, we're talking about a "money shot" to get people interested in the subject.
https://m.youtube.com/watch?v=zIsCswtkozI (mimeograph around 3:36:00 mark)
Now imagine if you picked a Y coordinate and extracted the corresponding row of pixels from every frame of a video file, and stacked them vertically. If there was scrolling text, or if the camera moved vertically, you would see a meaningful image, the same way a scanner obtains a complete image by moving a single-row sensor across whatever you're scanning. This is the same effect, just arising from the way the video signal is laid out on the disc.
One other important aspect is that by changing the angle of lighting, he could basically filter out data at a relevant wavelength.
--
At least that's what I got from the video.
If you look at a Constant Angular Velocity disc you can actually see "spokes" radiating out from the centre, with two broad ones 180° apart. The narrow spokes are the horizontal sync pulses occuring every 0.576° - the disc rotates at 25 revs per second and each concentric track is one complete frame. The broader spokes are of course the vertical sync pulses and colour burst occurring every 1/50th of a second.
If you're in the US or Japan, these numbers are 30 revs per second, 0.686° and 1/60th of a second, because of the lower resolution video standard, but it doesn't look like Laserdisc was much of a "thing" in those countries.
Here in the UK, in the 1980s all the schools took part in a thing called "The Domesday Project" [1] - the name is a reference to The Domesday Book, a survey of England and Wales carried out in the 11th century by William the Conqueror.
The Domesday Discs were CAV Laserdiscs that were played in a special player with a SCSI interface, attached to a BBC Micro computer. Because each concentric track was a complete frame it was possible to get perfect still frame video by just keeping the head still, so you could look at photos of places all around the UK and read a bit of information about them genlocked over the top.
[1] https://en.wikipedia.org/wiki/BBC_Domesday_Project
And a Laserdisc with digital sound literally has that CD audio EFM waveform in the lower frequencies.
> And that was because computing power back then was non-existent so they didn't use any kind of compression?
Compression is not a medium-level detail. You can easily store compressed data on a laserdisc.
How are the images encoded?
Without you showing them, they might never see this kind of stuff.
The legible text seen in the microscope images happens because of the combination of LaserDisc recording a raw and uncompressed encoding of the analog video signal, the way that LaserDisc used CAV to store an integer number of frames per track so that the image data for corresponding on-screen locations of subsequent frames would be aligned at the same radial position on the disc, and the credits scrolling vertically at constant speed.
If LaserDisc had used a digital encoding (especially a compressed encoding), the data on disc may still have had discernible patterns but the text would not necessarily have been legible. If it had used CAV but not stored a whole number of frames per track, then temporal and spatial locality on screen would not have corresponded so well to spatial locality on disc. And the vertically-scrolling credits are pretty much the only kind of content that can produce the recognizable and legible images on the disc surface.
I think the fact that the aspect ratio of the text came out approximately right probably is a consequence of the scrolling speed of the credits being chosen to suit the vertical resolution of the video. If the text had appeared squished in the microscope, it would probably have been moving too fast on screen to be clearly legible.
Those can have near-legible images, but most of the time they are not.
CAV discs contained one frame per rotation. While this meant you could only fit half an hour on one side of a disc, it did give you perfect slow-motion and freeze-frames.
I worked in a video store and loved LaserDiscs. The Duran Duran video album was CAV, and the Pioneer LD-700 had such a fast transport mechanism and remote control that I could to DJ-style "scratching" with it.
https://youtu.be/qZuR-772cks?t=1540
The data being written to the disk is the same in CAV or CLV disks, but the player just needs to know how to spin the disk at the right speed so that the laser can read the pits/lands correctly. It is purely a detail about the speed that the disk is spun at so they can cram more data on it with CLV disks.
What CAV LaserDiscs allow for, though, is to make it extremely obvious where scanlines and blanking intervals are in the video signal.
Even small changes in optics can drastically change how light spreads or how uniform illumination appears in a space.
Laser discs are not digital. They encode the analogue video signal’s value as the length of the pit. It is digitized in the time domain - sampled at some frequency, but the “vertical” signal value is stored entirely analogue. In terms of encoding it’s more similar to a VHS tape than a CD. Kinda crazy.
Laser disks are 100% digital (as you said, they store digits in the time domain).
They don't encode their data using binary like a CD does.
"Binary" and "digital" are two separate and unrelated concepts.
The sound was also analog to begin with, then the same encoding as CDs, then after that AC-3 and DTS.
And at that point, most players sold were combo players that could also play CDs.
And there was one more disc format: CD Video. It was a CD-sized digital single that also had a LaserDisc section for the (analog) music video. I have a couple; one is Bon Jovi.
So I don't want to say it's impossible, but I think it would require a lot more creativity.