This reminds me of a scene in "A Fire Upon the Deep" (1992) where they're on a video call with someone on another spaceship; but something seems a bit "off". Then someone notices that the actual bitrate they're getting from the other vessel is tiny -- far lower than they should be getting given the conditions -- and so most of what they're seeing on their own screens isn't actual video feed, but their local computer's reconstruction.
Was that the same book that had the concept of (paraphrasing using modern terminology) doing interstellar communications by sending back and forth LLMs trained on the people who wanted to talk, prompted to try and get a good business deal or whatever?
Fascinating. Vinge is about the furthest from “soft” sci-fi I can think of. We must have very different definitions of what makes something soft.
It’s certainly true that Vinge doesn’t spend much time on the engineering details, but I find him unusually clear on “imagine if we had this kind of impossible-now technology, but the rest of what we know about physics remained, how would people behave?”
He was, after all, a physics professor.
Rainbow’s End is much clearer on this than his distant future stuff, of course.
> Fascinating. Vinge is about the furthest from “soft” sci-fi I can think of. We must have very different definitions of what makes something soft.
That award goes to Greg Egan who has full list of citations on his website for each of his novels, as well as a list of mathematicians and physicists he requested help from.
If you want to read books that occasionally delve into pages of equations, Greg Egan is the author for you! (Seriously though, really good books, and the implications of his "what-ifs" are pretty damn cool)
Soft vs hard is based on how closely the world tracks with modern physics/science. As such even just FTL is soft, let alone everything else that doesn’t fit.
> Soft vs hard is based on how closely the world tracks with modern physics/science
Maybe it's not productive to quibble about definitions like this, but FWIW I don't agree with this criteria. I would argue Greg Egan's work, for example, is just about the "hardest" sci-fi there is, and yet much of that work takes place in universes that are entirely unlike our own.
Personally, I think what makes for "hard" sci-fi is that the rules of the universe are well-laid-out and consistent, and that the story springs (at least in some significant part) out of the consequences of those rules. That may mean a story set in the "future", where we have new technology or discover new physics, or "alternate universe" sci-fi like Egan's.
If changing the laws of the universe is fine, then nothing gets excluded even Harry Potter. It’s one of those definitions that allows anything and ultimately only feels fine because you’re adding some other criteria.
In defense of hard science fiction, it’s a meaningful category to talk about even if it’s not something you personally care about. People often want to weaken it but that just opens a door for a new category say “scientific science fiction” and we are back to square one.
Asking questions like what does AGI look like when they can’t just magically solve all issues can be fun. Hand waving the singularly as some religious event can also make interesting stories but so is considering how chaos theory limits what computation can actually achieve.
> If changing the laws of the universe is fine, then nothing gets excluded even Harry Potter.
Greg Egan's law changes are on the level of "I consulted with a bunch of theoretical physics professors and asked them what the implication of tweaking this one fundamental constant would be, then I spent years meticulously crafting a world that takes into account those implications, and I had others physics professors check over my work to make sure it was within the bounds of actuality, and then I wrote a story about characters in this new world."
> Asking questions like what does AGI look like when they can’t just magically solve all issues can be fun.
Greg Egan actually has a great book about this! Permutation City. CPU cycles aren't unlimited, and there are tons of ethical problems being confronted with the entire "simulate a person" thing.
Harry Potter isnt typically considered scifi because it doesn't critically examine its own premise and because the rules of the universe are yoked to the needs of the plot.
> the rules of the universe are yoked to the needs of the plot
It’s common for the rules of the universe to be adapted to fit the plot of random Star Trek episodes.
HP is not considered science fiction because of the trappings of the story. People use spells and enchanted objects for telekinesis, teleportation, and time travel not psychic abilities and technology to do the same things.
> critically examine its own premise
A great deal of science fiction doesn’t do that while plenty of fantasy does.
> If changing the laws of the universe is fine, then nothing gets excluded even Harry Potter
the laws of the universe in Harry Potter are so fickle and ever changing with the plot line that to me, it can only be considered soft. compare with Egan who takes a given cosmology and then works 100% within that world. that's hard.
That’s not a question about the underlying rules of a fictional work but your perception of how they are created. It’s possible to have a completely well defined fantasy setting with exact rules without the reader being aware of what those rules are or even knowing it’s using well defined rules.
Consider The Martian, early versions where posted online and the author changed what resources the character had to work with at the beginning. So what feels like a creative solution to limited resources was really giving the character exactly what they needed after a solution was found. Only examining a work we can’t distinguish ‘soft’ physics updated as the plot demands from a story based around fixed rules.
You seem to confuse the creative process with the final product. The rules can change during the creative process. It's the final product that I judge as a reader - I won't bother going over the inconsistencies in Harry Potter here, it's been done ad nauseam elsewhere. The physics doesn't change over the course of the story of the Martian.
What you view as inconsistencies are based around assumptions for how the underlying rules work and what happened that don’t necessarily apply.
One of the more interesting science fiction short stories I read seemed to have very inconsistent time travel, but on closer reading you find the two different methods involved had two different sets of rules. It’s easy to say something is inconsistent, but any possible story has a corresponding set of rules that work.
It’s rather similar to considering what characters may have been lying in a story.
It’s a classic definition. Soft/hard science fiction has two meanings either the topic is focused on hard sciences (physics) vs soft sciences (sociology) or “It can also refer to science fiction which prioritizes human emotions over scientific accuracy or plausibility.[1]”
So it’s not universal but is an accepted definition that any deviation from the possible or probable (for example, including faster-than-light travel or paranormal powers) to be a mark of "softness."
Popular science fiction is generally extremely soft, but occasionally you get stuff like The Cold Equations where the plot is driven by real world constraints. Even then it included FTL so a purest would call it soft.
A friend of mine and I both read it about the same time and discussed it afterwards. I thought it was pretty good, he thought it was not that great. What we agreed on was that in spite of there being many fantastic aspects to the book, on the whole it failed to be an awesome novel.
Definitely worth giving it a try if you're a programmer, just for the fact that it's written by another programmer: the opening scene where they find a bunch of rules written down and just follow them reminds me of ACPI; the discussion of public-key cryptography and shipping drives full of one-time-pad around the galaxy; the "compression scheme" with the video.
I agree that it was good but not particularly great. A Deepness in the Sky, however, is fantastic -- similar in many aspects but just flat out better all around.
It uses technological differences as key plot and setting components not just space as sea, so it is sci fi but it is improbable in many ways so yea “soft” sci fi or more speculative fiction
I think I agree both books were good and "A Deepness In The Sky" was better, but I would warn everyone that I thought both books used dramatic irony (showing us that characters were evil while hiding this from main characters) to hold attention to a degree that I kind of hated. And in "A Deepness In The Sky" sexual violence was used repeatedly to illustrate how evil the main characters were. I found it unnecessarily and a bit in poor taste.
On the other hand I think both books developed ideas wonderfully and there are bits of them I keep coming back to, even if I'll probably never reread them
These sorts of models pop here quite a bit, and they ignore fundamental facts of video codecs (video specific lossy compression technologies).
Traditional codecs have always focused on trade offs among encode complexity, decode complexity, and latency. Where complexity = compute. If every target device ran a 4090 at full power, we could go far below 22kbps with a traditional codec techniques for content like this. 22kbps isn't particularly impressive given these compute constraints.
This is my field, and trust me we (MPEG committees, AOM) look at "AI" based models, including GANs constantly. They don't yet look promising compared to traditional methods.
Oh and benchmarking against a video compression standard that's over twenty years old isn't doing a lot either for the plausibility of these methods.
This is my field as well, although I come from the neural network angle.
Learned video codecs definitely do look promising: Microsoft's DCVC-FM (https://github.com/microsoft/DCVC) beats H.267 in BD-rate. Another benefit of the learned approach is being able to run on soon commodity NPUs, without special hardware accommodation requirements.
In the CLIC challenge, hybrid codecs (traditional + learned components) are so far the best, so that has been a letdown for pure end to end learned codecs, agree. But something like H.267 is currently not cheap to run either.
It may sound like marketing wank, but it does a appear to be an established term of art in academia as far back as 1997 [1]
It just means that a person can't readily distinguish between the compressed image and
the uncompressed image. Usually because it takes some aspect(s) of the human visual system into account.
I read “perceptually lossless” to be equivalent to “transparent”, a more common phrase used in the audio/video codec world. It’s the bitrate/quality at which some large fraction of human viewers can’t distinguish a losslessly-encoded sample and the lossy-encoded sample, for some large fraction of content (constants vary in research papers).
As an example, crf=18 in libx264 is considered “perceptually lossless” for most video content.
Can you propose a better term for the concept then? Perceiving something as lossless is a real world metric that has a proper use case. "Perceptually lossless" does not try to imply that it is not lossy.
I work in graphics. Calling this transparency would be a terrible idea and make a lot of discussions around compression of videos and images with actual transparency very confusing.
Does the compression algorithm work well for transparency? Yes, it's effect on transparency is totally transparent! In fact the transparency is fully transparently compressed by our codec.
Yeah, don't do this please. Perceptually lossless is a term I've heard lots of times before and companies developing codecs usually have a fairly strong technical basis for making the claim. As in, it's not like they just glance at the results and say "yep, looks good to me". Rather, they'll be looking and spectral curves and image diffs - probably also motion diffs for videos - and checking whether they the losses are small enough to be undetectable to human eyes.
why not? if you change one pixel by one pixel brightness unit it is perceptually the same.
for the record, I found liveportrait to be well within the uncanny valley. it looks great for ai generated avatars, but the difference is very perceptually noticeable on familiar faces. still it's great.
For one, it doesn't obey the transitive property like a truly lossless process should: unless it settles into a fixed point, a perceptually lossless copy of a copy of a copy, etc., will eventually become perceptually different. E.g., screenshot-of-screenshot chains, each of which visually resembles the previous one, but which altogether make the original content unreadable.
Perceptual closure under repeated iterations is just a stronger form of perceptual losslessness, then, after k generations instead of the usual k=1. What you’re describing is called generation loss, and there are in fact perceptually lossy image codecs that have essentially no generation loss; jpeg xl is one https://m.youtube.com/watch?v=FtSWpw7zNkI
There is "Is identical", "looks identical" and "has lost sufficient detail to clearly not be the original." - being able to differentiate between these three states is useful.
Importantly the first one is parameterless, but the second and third are parameterized by the audience. For example humans don't see colour very well, some animals have much better colour gamut, while some can't distinguish colour at all.
Calling one of them "perceptually lossless" is cheating, to the disadvantage of algorithms that honestly advertise themselves as lossy while still achieving "looks identical" compression.
It's a well established term, though. It's been used in academic works for a long time (since at least 1970), and it's basically another term for the notion of "transparency" as it relates to data compression.
I honestly don't notice this anymore. Advertisers have been using such language since time immemorial, to the point it's pretty much a rule that an adjective with a qualifier means "not actually ${adjective}, but kind of like it in ${specific circumstances}". So "perceptually lossless" just means "not actually lossless, except you couldn't tell it from truly lossless just by looking".
It is in no way the definition of lossy. It is a subset of lossy. Most lossy image/video compression has visible artifacting, putting it outside the subset.
It means what it already says for itself, and does not need correcting into incorrectness.
"no perceived loss" is a perfectly internally consistent and sensible concept and is actually orthogonal to whether it's actually lossless or lossy.
For instance an actually lossless block of data could be perceptually lossy if displayed the wrong way.
In fact, even actual lossless data is always actually lossy, and only ever "perceptually lossless", and there is no such thing as actually lossless, because anything digital is always only a lossy approximation of anything analog. There is loss both at the ADC and at the DAC stage.
If you want to criticize a term for being nonsense misleading dishonest bullshit, then I guess "lossless" is that term, since it never existed and never can exist.
Similar to your points, i also expect `perceptually lossless` to be a valid term in the future with respect to AI. Ie i can imagine a compression which destroys detail, but on the opposite end it uses "AI" to reconstruct detail. Of course though, the AI is hallucinating the detail, so objectively it is lossy but perceptibly it is lossless because you cannot know which detail is incorrect if the ML is doing a good job.
In that scenario it certainly would not be `transparent` ie visually without any lossy artifacts. But your perception of it would look lossless.
Why don't you think it's a thing? A trivial example is audio. A ton of audio speakers can produce frequencies people cannot hear. If you have an unprocessed audio recording from a high end microphone one of the first compressions things you can do is clip of imperceptible frequencies. A form of compression.
As there are several patents, published studies, IEEE papers and thousands of google results for the term, I think it's safe to say that many people do not agree with your interpretation of the term.
"As a rule, strong feelings about issues do not emerge from deep understanding." -Sloman and Fernbach
It is definitely a thing given a good perceptual metric. The metric even doesn't have to be very accurate if the distortion is highly bounded, like only altering the lowermost bit. It is unfortunate that most commonly used distortion metrics like PSNR are not really that, though.
But that's mathematically impossible, to restore signal from extremely low bitrate stream with any highly bounded distortion. Perhaps only if you have highly restricted set of posible input, which online meetings aren't.
> Perhaps only if you have highly restricted set of posible input, which online meetings aren't.
Are you sure? After all, you can effectively summarize meetings in a plain text which is extremely restricted in comparison to the original input. Guaranteed, exact manner of speech and motions and all subtleties should be also included to be fair, but that information is still far limited to fill the 20 kbps bandwidth.
We need far more bandwidth only because we don't yet have an efficient way to reconstruct the input faithfully from such highly condensed information. Whenever we actually could, we ended up having a very efficient lossy algorithm that still preserves enough information for us human. Unless you are strictly talking about the lossless compression---which is however very irrelevant in this particular topic---, we should expect much more compression in the future even though that might not be feasible today.
Ability to tell MP3 from the original source was always dependent on encoder quality, bitrate, and the source material. In the mid 2000's, I tried to encode all of my music as MP3. Most of it sounded just fine because pop/rock/alt/etc are busy and "noisy" by design. But some songs (particularly with few instruments, high dynamic range, and female vocals) were just awful no matter how high I cranked the bitrate. And I'm not even an "audiophile," whatever that means these days.
No doubt encoders and the codecs themselves have improved vastly since then. It would be interesting to see if I could tell the difference in a double-blind test today.
I find generic ABX tests not great, personally, because I generally don’t know what to be listening for. However, with songs I’ve listened to lossless my whole life, it’s much easier to spot encoding failures - an intuitive “wait, that cymbal crash sounded different” or “that multi-instrument harmonic should be cleaner/dirtier.”
That being said, 320Kbps AAC encoded by Core Audio I’ve found to be pretty much transparent with anything I’ve thrown at it. Anything less than that (256Kbps AAC, 320Kbps MP3, etc) I can ABX sometimes, as long as I’m familiar with the source material, and usually only with quality headphones. Although no streaming services provide that, so I’m stuck with ALAC through Apple Music for streaming (which is more convenient than my old solution, which was transcoding and transferring to an iPod ~20k songs selected from ~90k in my library based on a variety of rules than never gave me the song I’m looking for). And really, ~900Kbps lossless is pretty easy to justify these days with 5G data speeds and generally much higher data transfer limits.
The other downside to storing losing encodings these days is the fact that almost everyone uses Bluetooth for their listening, which is an additional lossy encoding. While 256Kbps AAC/320Kbps MP3 might be transparent in some cases, when it’s re-encoded it very rarely is (in my experience)
iirc there's "easy" (though i don't know them) tests to validate if the signal is lossless or not. When played over speakers for humans, at least.
I always intend to figure out how that works, because i don't feel a lot of audiophiles are actually speaking truth in many cases lol. Still, i don't know - i can't remember my sources to figure it out for myself :/
Lossy audio formats suddenly become very discernible once you subtract the left channel from the right channel. Try that with Lossless audio vs MP3, Vorbis, Opus, AAC, etc. You're listening to only the errors at that point.
A family member of mine didn't see the point of 1080p.
Turned out they needed cataract surgery and got fancy replacement lenses in their eyes.
After that, they saw the point.
Needing to define "perception" is a much weaker criticism than "isn't a thing and doesn't make sense".
It's easy enough to specify an average person looking very closely, or a 99th percentile person, or something like that, and show the statistics backing it up.
I like how the saddle in the background moves with the reconstructed head; it probably works better with uncluttered backgrounds.
This is interesting tech, and the considerations in the introduction are particularly noteworthy. I never considered the possibility of animating 2D avatars with no 3D pipeline at all.
Well, this isn't probably a problem with the model, but the source frame having wrong eye gaze. Besides, perceptually lossless need not be defined in a side-by-side comparison context. If you were only viewing the right hand side video, how could you tell the eye gaze is off? The point was more on that the movement looks natural, unlike almost all neural avatars up to this year.
Your argumentation does make sense to me; but it also makes the term lossless pull a lot of weight. Lossless in video encoding is usually defined by zero difference between source and target.
> But one overlooked use case of the technology is (talking head) video compression.
> On a spectrum of model architectures, it achieves higher compression efficiency at the cost of model complexity. Indeed, the full LivePortrait model has 130m parameters compared to DCVC’s 20 million. While that’s tiny compared to LLMs, it currently requires an Nvidia RTX 4090 to run it in real time (in addition to parameters, a large culprit is using expensive warping operations). That means deploying to edge runtimes such as Apple Neural Engine is still quite a ways ahead.
It’s very cool that this is possible, but the compression use case is indeed .. a bit far fetched. A insanely large model requiring the most expensive consumer GPU to run on both ends and at the same time being limited in bandwidth so much (22kbps) is a _very_ limited scenario.
One cool use would be communication in space - where it's feasible that both sides would have access to high-end compute units but have a very limited bandwidth between each other.
Increasingly mobile networks are like this. There are all kinds of bandwidth issues, especially when customers are subject to metered pricing for data.
Staying in contact with someone for hours on metered mobile internet connection comes to mind. Low bandwidth translates to low total data volume over time. If I could be video chatting on one of those free internet SIM cards that's a breakthrough.
One use case might be if you have limited bandwidth, perhaps only a voice call, and want to join a video conference. I could imagine dialling in to a conference with a virtual face as an improvement over no video at all.
130m parameters isn’t insanely large, even for smartphone memory. The high GPU usage is a barrier at the moment, but I wouldn’t put it past Apple to have 4090-level GPU performance in an iPhone before 2030.
The trade-off may not be worth it today, but the processing power we can expect in the coming years will make this accessible to ordinary consumers. When your laptop or phone or AR headset has the processing power to run these models, it will make more efficient use of limited bandwidth, even if more bandwidth is available. I don't think available bandwidth will scale at the same rate as processing power, but even if it does, the picture be that much more realistic.
The second example shown is not perceptually lossless, unless you’re so far on the spectrum you won’t make eye contact even with a picture of a person. The reconstructed head doesn’t look in the same direction as the original.
However is does raise an interesting property in that if you are on the spectrum or have ADHD, you only need one headshot of yourself staring directly at the camera and then the capture software can stop you from looking at your taskbar or off into space.
Now that you mention it, it never occurred to me that Snake's radio transmitted video as well. "Did you like my new sunglasses?"
If you could reserve a small portion of the radio bandwidth to broadcast a thumbnail + low bandwidth compressed representation of the face movements, you could technically have something similar without encoding any video (think low res, eye + mouth movements).
We could run something like TensorFlow.js in a Chrome extension to identify the person in the image and replace it in the dom. A little resource intensive for inference on every image in but probably worth it in this case.
the only information that needs to be transmitted is the change in expression, pose and facial keypoints
Does anyone else remember the weirder (for lack of a better term) features of MPEG-4 part 2, like face and body animation? It did something like that, but as far as I know nearly no one used that feature for anything.
Now that we're moving towards context-specific compression algorithms, can we please use WASM as the file header for these media files, instead of inventing something new. :)
As there are several patents, published studies, IEEE papers and thousands of google results for the term, I think it's safe to say that many people do not agree with your interpretation of the term.
Lossiness definitely matters when you’re doing forensics. But not for consumers.
If you just want to bop to Taylor who the fuck cares. The iPod ended that argument. Yes I can be a perfectionist, or I can have one thousand songs in my pocket. That was more than half of your collection for many people at the time.
I am sure calling modern heuristics "AI" gets people excited, but it doesn't seem "Magical" when trivial implementations are functionally equivalent. =3
I know metahuman. As impressive as it is, when you judge by the standards of game graphics, if you are ever mislead into thinking metahumans are real humans or even real physically existing things it's time to see your eye doctor (and/or do MRI head scan).
On the other hand AI videos can be easily mistaken for people or hyper realistic physical sculptures.
There's something basic about how light works that traditional computer graphics still fails to grasp. Looking at its productions and comparing it to what AI generates is like looking at output of amateur and an artist. Sure, maybe artist doesn't always draw all 5 fingers but somehow captures the essence of the image in seemingly random arrangement of light and dark strokes, while amateur just tries to do their best but fails in some very significant ways.
"AI" videos make many errors all the time, but most people are not aware of what to look for... Undetectable CGI is done in film/games all the time, and indeed it takes talent to hide the fact it is fake.
One could rely on the media encoder to garble output enough to look more plausible (people on potato devices are used to looking at garbage content.) However, at the end of the day the "uncanny valley" effect takes over every-time even for live action data in a auto-generated asset, as the missing data can't be "Magically" recovered with 100% certainty.
BTW This is the best sci-fi book ever.
It’s certainly true that Vinge doesn’t spend much time on the engineering details, but I find him unusually clear on “imagine if we had this kind of impossible-now technology, but the rest of what we know about physics remained, how would people behave?”
He was, after all, a physics professor.
Rainbow’s End is much clearer on this than his distant future stuff, of course.
That award goes to Greg Egan who has full list of citations on his website for each of his novels, as well as a list of mathematicians and physicists he requested help from.
If you want to read books that occasionally delve into pages of equations, Greg Egan is the author for you! (Seriously though, really good books, and the implications of his "what-ifs" are pretty damn cool)
The short stories "Luminous" and "Dark Integers", the novels "Diaspora" and "Schild's Ladder". So good.
qntm (another author) hits somewhat similarly.
Actually, he was a mathematics and computer science teacher at San Diego State University.
https://en.wikipedia.org/wiki/Vernor_Vinge
Maybe it's not productive to quibble about definitions like this, but FWIW I don't agree with this criteria. I would argue Greg Egan's work, for example, is just about the "hardest" sci-fi there is, and yet much of that work takes place in universes that are entirely unlike our own.
Personally, I think what makes for "hard" sci-fi is that the rules of the universe are well-laid-out and consistent, and that the story springs (at least in some significant part) out of the consequences of those rules. That may mean a story set in the "future", where we have new technology or discover new physics, or "alternate universe" sci-fi like Egan's.
In defense of hard science fiction, it’s a meaningful category to talk about even if it’s not something you personally care about. People often want to weaken it but that just opens a door for a new category say “scientific science fiction” and we are back to square one.
Asking questions like what does AGI look like when they can’t just magically solve all issues can be fun. Hand waving the singularly as some religious event can also make interesting stories but so is considering how chaos theory limits what computation can actually achieve.
Greg Egan's law changes are on the level of "I consulted with a bunch of theoretical physics professors and asked them what the implication of tweaking this one fundamental constant would be, then I spent years meticulously crafting a world that takes into account those implications, and I had others physics professors check over my work to make sure it was within the bounds of actuality, and then I wrote a story about characters in this new world."
> Asking questions like what does AGI look like when they can’t just magically solve all issues can be fun.
Greg Egan actually has a great book about this! Permutation City. CPU cycles aren't unlimited, and there are tons of ethical problems being confronted with the entire "simulate a person" thing.
It’s common for the rules of the universe to be adapted to fit the plot of random Star Trek episodes.
HP is not considered science fiction because of the trappings of the story. People use spells and enchanted objects for telekinesis, teleportation, and time travel not psychic abilities and technology to do the same things.
> critically examine its own premise
A great deal of science fiction doesn’t do that while plenty of fantasy does.
the laws of the universe in Harry Potter are so fickle and ever changing with the plot line that to me, it can only be considered soft. compare with Egan who takes a given cosmology and then works 100% within that world. that's hard.
Consider The Martian, early versions where posted online and the author changed what resources the character had to work with at the beginning. So what feels like a creative solution to limited resources was really giving the character exactly what they needed after a solution was found. Only examining a work we can’t distinguish ‘soft’ physics updated as the plot demands from a story based around fixed rules.
One of the more interesting science fiction short stories I read seemed to have very inconsistent time travel, but on closer reading you find the two different methods involved had two different sets of rules. It’s easy to say something is inconsistent, but any possible story has a corresponding set of rules that work.
It’s rather similar to considering what characters may have been lying in a story.
> but on closer reading
does not make the inconsistencies go away, but they multiply.
Again based on specific assumptions. The universe of possibilities includes very strange places.
that's an unjustified assumption.
If nothing else pure randomness is a unsatisfying possibility as is a full branching search of every possible state for a universe.
You don’t claim to be definitive?
So it’s not universal but is an accepted definition that any deviation from the possible or probable (for example, including faster-than-light travel or paranormal powers) to be a mark of "softness."
https://en.wikipedia.org/wiki/Soft_science_fiction
Popular science fiction is generally extremely soft, but occasionally you get stuff like The Cold Equations where the plot is driven by real world constraints. Even then it included FTL so a purest would call it soft.
Definitely worth giving it a try if you're a programmer, just for the fact that it's written by another programmer: the opening scene where they find a bunch of rules written down and just follow them reminds me of ACPI; the discussion of public-key cryptography and shipping drives full of one-time-pad around the galaxy; the "compression scheme" with the video.
On the other hand I think both books developed ideas wonderfully and there are bits of them I keep coming back to, even if I'll probably never reread them
Traditional codecs have always focused on trade offs among encode complexity, decode complexity, and latency. Where complexity = compute. If every target device ran a 4090 at full power, we could go far below 22kbps with a traditional codec techniques for content like this. 22kbps isn't particularly impressive given these compute constraints.
This is my field, and trust me we (MPEG committees, AOM) look at "AI" based models, including GANs constantly. They don't yet look promising compared to traditional methods.
Oh and benchmarking against a video compression standard that's over twenty years old isn't doing a lot either for the plausibility of these methods.
Learned video codecs definitely do look promising: Microsoft's DCVC-FM (https://github.com/microsoft/DCVC) beats H.267 in BD-rate. Another benefit of the learned approach is being able to run on soon commodity NPUs, without special hardware accommodation requirements.
In the CLIC challenge, hybrid codecs (traditional + learned components) are so far the best, so that has been a letdown for pure end to end learned codecs, agree. But something like H.267 is currently not cheap to run either.
Agreed hybrid presents real opportunity.
Someone was just having fun here, it's not as if they present it as a general codec.
It just means that a person can't readily distinguish between the compressed image and the uncompressed image. Usually because it takes some aspect(s) of the human visual system into account.
[1] https://scholar.google.com/scholar?hl=en&as_sdt=0%2C22&q=per...
As an example, crf=18 in libx264 is considered “perceptually lossless” for most video content.
Does the compression algorithm work well for transparency? Yes, it's effect on transparency is totally transparent! In fact the transparency is fully transparently compressed by our codec.
Yeah, don't do this please. Perceptually lossless is a term I've heard lots of times before and companies developing codecs usually have a fairly strong technical basis for making the claim. As in, it's not like they just glance at the results and say "yep, looks good to me". Rather, they'll be looking and spectral curves and image diffs - probably also motion diffs for videos - and checking whether they the losses are small enough to be undetectable to human eyes.
for the record, I found liveportrait to be well within the uncanny valley. it looks great for ai generated avatars, but the difference is very perceptually noticeable on familiar faces. still it's great.
There is "Is identical", "looks identical" and "has lost sufficient detail to clearly not be the original." - being able to differentiate between these three states is useful.
The other two are variations of lossy.
Calling one of them "perceptually lossless" is cheating, to the disadvantage of algorithms that honestly advertise themselves as lossy while still achieving "looks identical" compression.
It's also used in the first paragraph of the Wikipedia article on the term "transparency" as it relates to data compression.
"no perceived loss" is a perfectly internally consistent and sensible concept and is actually orthogonal to whether it's actually lossless or lossy.
For instance an actually lossless block of data could be perceptually lossy if displayed the wrong way.
In fact, even actual lossless data is always actually lossy, and only ever "perceptually lossless", and there is no such thing as actually lossless, because anything digital is always only a lossy approximation of anything analog. There is loss both at the ADC and at the DAC stage.
If you want to criticize a term for being nonsense misleading dishonest bullshit, then I guess "lossless" is that term, since it never existed and never can exist.
In that scenario it certainly would not be `transparent` ie visually without any lossy artifacts. But your perception of it would look lossless.
The future is going to be weird.
"As a rule, strong feelings about issues do not emerge from deep understanding." -Sloman and Fernbach
Are you sure? After all, you can effectively summarize meetings in a plain text which is extremely restricted in comparison to the original input. Guaranteed, exact manner of speech and motions and all subtleties should be also included to be fair, but that information is still far limited to fill the 20 kbps bandwidth.
We need far more bandwidth only because we don't yet have an efficient way to reconstruct the input faithfully from such highly condensed information. Whenever we actually could, we ended up having a very efficient lossy algorithm that still preserves enough information for us human. Unless you are strictly talking about the lossless compression---which is however very irrelevant in this particular topic---, we should expect much more compression in the future even though that might not be feasible today.
No doubt encoders and the codecs themselves have improved vastly since then. It would be interesting to see if I could tell the difference in a double-blind test today.
https://abx.digitalfeed.net/
https://www.npr.org/sections/therecord/2015/06/02/411473508/...
That being said, 320Kbps AAC encoded by Core Audio I’ve found to be pretty much transparent with anything I’ve thrown at it. Anything less than that (256Kbps AAC, 320Kbps MP3, etc) I can ABX sometimes, as long as I’m familiar with the source material, and usually only with quality headphones. Although no streaming services provide that, so I’m stuck with ALAC through Apple Music for streaming (which is more convenient than my old solution, which was transcoding and transferring to an iPod ~20k songs selected from ~90k in my library based on a variety of rules than never gave me the song I’m looking for). And really, ~900Kbps lossless is pretty easy to justify these days with 5G data speeds and generally much higher data transfer limits.
The other downside to storing losing encodings these days is the fact that almost everyone uses Bluetooth for their listening, which is an additional lossy encoding. While 256Kbps AAC/320Kbps MP3 might be transparent in some cases, when it’s re-encoded it very rarely is (in my experience)
I always intend to figure out how that works, because i don't feel a lot of audiophiles are actually speaking truth in many cases lol. Still, i don't know - i can't remember my sources to figure it out for myself :/
It's easy enough to specify an average person looking very closely, or a 99th percentile person, or something like that, and show the statistics backing it up.
This is interesting tech, and the considerations in the introduction are particularly noteworthy. I never considered the possibility of animating 2D avatars with no 3D pipeline at all.
I feel like I’m taking crazy pills.
Smells like rationalization to me.
> On a spectrum of model architectures, it achieves higher compression efficiency at the cost of model complexity. Indeed, the full LivePortrait model has 130m parameters compared to DCVC’s 20 million. While that’s tiny compared to LLMs, it currently requires an Nvidia RTX 4090 to run it in real time (in addition to parameters, a large culprit is using expensive warping operations). That means deploying to edge runtimes such as Apple Neural Engine is still quite a ways ahead.
It’s very cool that this is possible, but the compression use case is indeed .. a bit far fetched. A insanely large model requiring the most expensive consumer GPU to run on both ends and at the same time being limited in bandwidth so much (22kbps) is a _very_ limited scenario.
Though, I somewhat doubt even 22kbps is available generally.
However is does raise an interesting property in that if you are on the spectrum or have ADHD, you only need one headshot of yourself staring directly at the camera and then the capture software can stop you from looking at your taskbar or off into space.
I don't know. I think you'd be surprised.
That's already kind of an issue with vloggers. Often they're looking just left or right of the camera at a monitor or something.
Reminds me of the video chat in Metal Gear Solid 1 https://youtu.be/59ialBNj4lE?t=21
If you could reserve a small portion of the radio bandwidth to broadcast a thumbnail + low bandwidth compressed representation of the face movements, you could technically have something similar without encoding any video (think low res, eye + mouth movements).
Maybe there is a custom web filter in there somewhere that could block particular people and images of them.
https://news.ycombinator.com/item?id=22907718
Does anyone else remember the weirder (for lack of a better term) features of MPEG-4 part 2, like face and body animation? It did something like that, but as far as I know nearly no one used that feature for anything.
https://en.wikipedia.org/wiki/Face_Animation_Parameter
and in the worst, trust on the internet will be heavily undermined
...as long as the model doesn't include data to put a shoe on one's head.
Lossiness definitely matters when you’re doing forensics. But not for consumers.
If you just want to bop to Taylor who the fuck cares. The iPod ended that argument. Yes I can be a perfectionist, or I can have one thousand songs in my pocket. That was more than half of your collection for many people at the time.
24fps * 52 facial 3D marker * 16bit packed delta planar projected offsets (x,y) = 19.968 kbps
And this is done in Unreal games on a potato graphics card all the time:
https://apps.apple.com/us/app/live-link-face/id1495370836
I am sure calling modern heuristics "AI" gets people excited, but it doesn't seem "Magical" when trivial implementations are functionally equivalent. =3
https://www.unrealengine.com/en-US/metahuman
The artifacts in raster image data is nowhere near what a reasonable model can achieve even at low resolutions. =3
On the other hand AI videos can be easily mistaken for people or hyper realistic physical sculptures.
https://img-9gag-fun.9cache.com/photo/aYQ776w_460svvp9.webm
There's something basic about how light works that traditional computer graphics still fails to grasp. Looking at its productions and comparing it to what AI generates is like looking at output of amateur and an artist. Sure, maybe artist doesn't always draw all 5 fingers but somehow captures the essence of the image in seemingly random arrangement of light and dark strokes, while amateur just tries to do their best but fails in some very significant ways.
One could rely on the media encoder to garble output enough to look more plausible (people on potato devices are used to looking at garbage content.) However, at the end of the day the "uncanny valley" effect takes over every-time even for live action data in a auto-generated asset, as the missing data can't be "Magically" recovered with 100% certainty.
Bye =3
In movies it can be done with enough of manual tweaking by artists and a lot of photographic content around to borrow sense of reality from it.
"Potato" devices by which I assume you mean average phones, currently have better resolutions than PCs had very recently and a lot still do (1080p).
And a photo on 480p still looks more real than anything CGI (not AI).
Your signature is hilarious. I won't comment about the reasons because I don't want this whole thread to get flagged.
https://www.youtube.com/watch?v=vJG698U2Mvo
Several 8bit games had their own aesthetic charm, but were at least fun...
Cheers, =3
- Arthur C. Clarke