Time to wheel out one of my favorite quotes about the signature of a medium:
"Whatever you now find weird, ugly, uncomfortable and nasty about a new medium will surely become its signature. CD distortion, the jitteriness of digital video, the crap sound of 8-bit - all of these will be cherished and emulated as soon as they can be avoided. It’s the sound of failure: so much modern art is the sound of things going out of control, of a medium pushing to its limits and breaking apart. The distorted guitar sound is the sound of something too loud for the medium supposed to carry it. The blues singer with the cracked voice is the sound of an emotional cry too powerful for the throat that releases it. The excitement of grainy film, of bleached-out black and white, is the excitement of witnessing events too momentous for the medium assigned to record them." - Brian Eno
Idle thought: I don't think I've ever seen one of these TV emulator things implement the situation where the vertical oscillator was slightly wrong and you get the picture slowly looping up the screen.
This one does. You can configure the noise injected into the signal and when it gets too much, it loses sync and the picture starts rolling. It's actually a software NTSC modulator/demodulator, not just an effect to simulate it.
I actually posted ntsc-rs as it came up in my research - I'm also looking for something like what you're describing..!
I was also looking into https://codeberg.org/fsphil/hacktv which generates a variety of different analog tv signals (meant to be broadcast using HackRF) - but yes, I want the opposite - an analog-receiver-emulator...? And one that would be "ok" with incorrect signals // fail like an analog TV would... :-)
You're not getting the full experience of analogue telly artifacts until you emulate colour subcarrier phase shift and colour burst detection failure. (-:
I once tried to fully analyze the amazing NTSC emulation used in OpenEmulator. I went down a rabbit hole that involved losing motivation several lessons in to a signal processing class on YouTube, but for those interested, I did at least pull quite a lot of it apart here: https://observablehq.com/@zellyn/apple-ii-ntsc-emulation-ope...
I also ported it to JavaScript (linked from above page)
I educated anyone who asked about the NTSC filter over the years because I wanted to see less-optimized implementations of it, given how much faster hardware is than the mid 2000s (it precalculated the kernels for every color at every phase offset and did the signed RGB math during rendering). There's something satisfying about being able to recreate the peculiarities of old hardware we grew up with, as a way of demystifying it.
I was immediately reminded of the fake VHS line artifacts for Stranger Things - A Bad Lip Reading[0], which I assume are sort of a bit about the fake film grain things during the opening titles in the Stranger Things show.
I do love that this is an area of such active development. But I'm curious to see what the artifact simulation crowd thinks of it. I most often encounter them as shaders for emulators and such, but of course this kind of structure degradation of a pristine video is also in high demand these days for video production. Producers want that 90s-camcorder look but crews can't actually use the clunky 90s-camcorder hardware and formats.
I'm actually surprised there isn't much of a scene for authentic camcorder footage - directors love to bust out real black and white film cameras for stuff?
Film is a fun, interesting, authentic, and useful medium for filmmakers, and there are established workflows for it. A camcorder writing interlaced video to miniDV may have its charms (I still have a great old Panasonic 3CCD one) but as a filmmaking tool it would be really inconvenient. Shooting in an ordinary digital workflow and adding the effect later is a no brainer production-wise.
That said, I would not be surprised to see camcorders, DV or VHS or whatever, rise up as a Polaroid-like alternative to smartphone cameras! Old digital point and shoots are already popular that way.
In 2009, I recorded a video of the after effects of a torrential downpour in Toronto on a Sony HDV camera. I also called up a few news stations to see if I could sell it.
I ended up reaching CFTO (CTV Toronto), and took the footage over to Channel 9 Court. What happened next took me by complete surprise.
The flagship station of a national network had no deck in the building that would play HDV mini DV tapes. I hadn’t brought my camcorder or my MBP either, so I couldn't quickly convert it into a format that they could use.
I ended up going home, and exporting via FCP and burning onto a DVD. It worked, I got to see the inside of a news station and I got $135 for it. The news broadcast later that day showed about 10 seconds of my footage, which by extrapolation, was the highest-ever hourly rate I’ve ever earned: ~$48,600/hour.
The lesson here was that DV and DV-adjacent workflows were difficult in a pro context even when they were mainstream in the consumer market.
I've never been a smartphone user, and have moved from a Flip Camcorder, to various point-and-shoots in video mode (never liked very much), and just in the last 3 years, have discovered that Sony handicams are now pocket-sized, I never considered carrying around one before, but it's actually completely reasonable.
The model (HDRCX405) is wonderful, 30x optical zoom a real value-add over smartphones, but also I just love the ergonomics in general, very easy to pick it up, and start a video within a second.
That said, Sony discontinued the low-end handicam line last year (this model went from $200 new to $800 used), which is really unfortunately, right as I hope this niche might gain momentum.
To have true VHS effect, I think we should train AI for this, examples from digital videos to record on true VHS tape, on multiples VHS devices then digitalize and calculate difference between original and digitalized from VHS.
Then even we could have filter like: VHS Panasonic, VHS Sony...
I might argue that generating and decoding an actual NTSC signal, as the OP project does, would be true in ways that a generative model based on all of that would not be.
Greg! I love this!!! Just last night I was trying to rewatch the x-files and was telling Luna that I would need to get a TV filter/shader/overlay thingy to see it the way it was meant to be seen.
I hope this leads to people being interested in more and larger public gatherings. Seeing things with our own eyes and having fun with other real people.
Time to wheel out one of my favorite quotes about the signature of a medium:
"Whatever you now find weird, ugly, uncomfortable and nasty about a new medium will surely become its signature. CD distortion, the jitteriness of digital video, the crap sound of 8-bit - all of these will be cherished and emulated as soon as they can be avoided. It’s the sound of failure: so much modern art is the sound of things going out of control, of a medium pushing to its limits and breaking apart. The distorted guitar sound is the sound of something too loud for the medium supposed to carry it. The blues singer with the cracked voice is the sound of an emotional cry too powerful for the throat that releases it. The excitement of grainy film, of bleached-out black and white, is the excitement of witnessing events too momentous for the medium assigned to record them." - Brian Eno
Idle thought: I don't think I've ever seen one of these TV emulator things implement the situation where the vertical oscillator was slightly wrong and you get the picture slowly looping up the screen.
This one does. You can configure the noise injected into the signal and when it gets too much, it loses sync and the picture starts rolling. It's actually a software NTSC modulator/demodulator, not just an effect to simulate it.
https://github.com/LMP88959/NTSC-CRT
I sincerely appreciate this fidelity to fidelity
I actually posted ntsc-rs as it came up in my research - I'm also looking for something like what you're describing..!
I was also looking into https://codeberg.org/fsphil/hacktv which generates a variety of different analog tv signals (meant to be broadcast using HackRF) - but yes, I want the opposite - an analog-receiver-emulator...? And one that would be "ok" with incorrect signals // fail like an analog TV would... :-)
You're not getting the full experience of analogue telly artifacts until you emulate colour subcarrier phase shift and colour burst detection failure. (-:
And of course PAL and Hanover bars.
Which is why NTSC was often said to mean Never Twice the Same Color!
Good thing they fixed it in Always The Same Color.
In my decades working in TV it was always "Never The Same Color."
See also: Picture At Last!
See also: System Essentially Contrary to the American Method
I once tried to fully analyze the amazing NTSC emulation used in OpenEmulator. I went down a rabbit hole that involved losing motivation several lessons in to a signal processing class on YouTube, but for those interested, I did at least pull quite a lot of it apart here: https://observablehq.com/@zellyn/apple-ii-ntsc-emulation-ope...
I also ported it to JavaScript (linked from above page)
I educated anyone who asked about the NTSC filter over the years because I wanted to see less-optimized implementations of it, given how much faster hardware is than the mid 2000s (it precalculated the kernels for every color at every phase offset and did the signed RGB math during rendering). There's something satisfying about being able to recreate the peculiarities of old hardware we grew up with, as a way of demystifying it.
I was immediately reminded of the fake VHS line artifacts for Stranger Things - A Bad Lip Reading[0], which I assume are sort of a bit about the fake film grain things during the opening titles in the Stranger Things show.
[0] https://www.youtube.com/watch?v=S-4rhjO6xYg
I do love that this is an area of such active development. But I'm curious to see what the artifact simulation crowd thinks of it. I most often encounter them as shaders for emulators and such, but of course this kind of structure degradation of a pristine video is also in high demand these days for video production. Producers want that 90s-camcorder look but crews can't actually use the clunky 90s-camcorder hardware and formats.
I'm actually surprised there isn't much of a scene for authentic camcorder footage - directors love to bust out real black and white film cameras for stuff?
Film is a fun, interesting, authentic, and useful medium for filmmakers, and there are established workflows for it. A camcorder writing interlaced video to miniDV may have its charms (I still have a great old Panasonic 3CCD one) but as a filmmaking tool it would be really inconvenient. Shooting in an ordinary digital workflow and adding the effect later is a no brainer production-wise.
That said, I would not be surprised to see camcorders, DV or VHS or whatever, rise up as a Polaroid-like alternative to smartphone cameras! Old digital point and shoots are already popular that way.
In 2009, I recorded a video of the after effects of a torrential downpour in Toronto on a Sony HDV camera. I also called up a few news stations to see if I could sell it.
I ended up reaching CFTO (CTV Toronto), and took the footage over to Channel 9 Court. What happened next took me by complete surprise.
The flagship station of a national network had no deck in the building that would play HDV mini DV tapes. I hadn’t brought my camcorder or my MBP either, so I couldn't quickly convert it into a format that they could use.
I ended up going home, and exporting via FCP and burning onto a DVD. It worked, I got to see the inside of a news station and I got $135 for it. The news broadcast later that day showed about 10 seconds of my footage, which by extrapolation, was the highest-ever hourly rate I’ve ever earned: ~$48,600/hour.
The lesson here was that DV and DV-adjacent workflows were difficult in a pro context even when they were mainstream in the consumer market.
I've never been a smartphone user, and have moved from a Flip Camcorder, to various point-and-shoots in video mode (never liked very much), and just in the last 3 years, have discovered that Sony handicams are now pocket-sized, I never considered carrying around one before, but it's actually completely reasonable.
The model (HDRCX405) is wonderful, 30x optical zoom a real value-add over smartphones, but also I just love the ergonomics in general, very easy to pick it up, and start a video within a second.
That said, Sony discontinued the low-end handicam line last year (this model went from $200 new to $800 used), which is really unfortunately, right as I hope this niche might gain momentum.
To have true VHS effect, I think we should train AI for this, examples from digital videos to record on true VHS tape, on multiples VHS devices then digitalize and calculate difference between original and digitalized from VHS.
Then even we could have filter like: VHS Panasonic, VHS Sony...
This would be very interesting project.
Can't you just cut out the AI from this pipeline by recording footage onto VHS and then digitizing that?
I think the idea would be allowing arbitrary videos to be converted without the need for a working tape drive
I might argue that generating and decoding an actual NTSC signal, as the OP project does, would be true in ways that a generative model based on all of that would not be.
heres a test output it looks convincing
https://x.com/AgentifySH/status/2063351105162224119
You need the 80s soundtrack for the full effect: https://www.youtube.com/watch?v=8oVfIFrpslI
Greg! I love this!!! Just last night I was trying to rewatch the x-files and was telling Luna that I would need to get a TV filter/shader/overlay thingy to see it the way it was meant to be seen.
You mind reader you
Rafał!!! U+1F62D U+1F62D U+1F62D!!!! haha
I'll email you. sorry everyone, just two pal's pall'ing around xx
I wonder if any of this was used to produce Backrooms
It looks quite unusual, I will definitely try it.
never expected valadaptive to be on front page of HN
pretty cool!
Great, now I won't be able to trust that old videos aren't AI slop either.
I hope this leads to people being interested in more and larger public gatherings. Seeing things with our own eyes and having fun with other real people.
This will be the thing that does it. That sounds reasonable.