> The result was visually striking. Repeating tile patterns disappeared, and artists could paint unique detail across large environments without concern for reuse. The primary cost was not GPU throughput, but latency elsewhere in the system.
No, the primary "cost" was artists having to fill a world with unlimited textures instead of just filling memory and then having to make due.
The constraint of "limited texture memory budget" also puts a constraint on how much work the artists can do. Remove that constraint lets artists do unlimited work. It might sound like a plus because "freedom!" but it turns into a minus trying to actually ship on time and at budget.
I get that wasn't the point of the article's "cost", but thought it was worth mentioning.
Probably just Aspie, judging by some of their other writing (including their About page). I've seen Aspie writing misidentified as LLM output surprisingly often.
“Good portion of the world” is probably a handful of people.
Her full quote btw:
“Once upon a time, I was the centerfold of Playboy,” says the former model in the new documentary Losing Lena. “But I retired from modeling a long time ago. It’s time I retired from tech, too.”
Well, for instance, it's the official policy of the IEEE to not allow this image in new publications. And they're far from the only journal (or set of journals) that have this policy.
Given that it's use is banned in most academic journals dealing with imaging/graphics, you'd be wrong.
And as several journals have brought up in the banning, it's not even good at what it purports to be for these use cases. It's a pretty poor quality image to start off with due to being scanned to a digital file with 1970s technology.
At this point the ones defending its continued use are the vocal minority on some weird anti-woke crusade that doesn't even make sense on technical grounds.
You’re using vocal minority framing right now. When I care about it, I’m a weird crusader for caring and noticing. But then you organize a campaign to change it.
There is a large body of literature using these images so it’s helpful to have a comparison which is persistent through time and familiar.
> Given that it's use is banned in most academic journals dealing with imaging/graphics, you'd be wrong.
A copyrighted image of a nude model elected for no obvious reason has a test image in the University of South California by some pervs and then used in a lot of papers as a test image.
Or, a standard cropped image of a playgirl used in the field of image processing.
"elected for no obvious reason" isn't quite right, as a test image for computer graphics it has regions of very high frequency detail and regions of very low frequency detail which make it easier to spot various compression artifacts, and it makes a good study for edge detection, with both very clear edges along the outline, but more subjective edges in the feathering.
It's redish. Ok it has a blur and details on the foreground but could have been any image with blurred background and a face.
"very low frequency detail", we are talking about a 512x512 picture here, it has low and high frequency details (FFT speaking) like most photos.
"Good for edges detection" doesn't mean anything. Like, is the image good for edge detection or the algorithm is good at detecting edges ? What does "subjective edges" even mean ? Does it mean hard to spot ?
That looks like technical reasons but it just noise. They literally grab a playboy magazine and decided it was well enough (and indeed, it wasn't that bad, yes). Still not professional. The message is "We have playboy magazines at work and we are proud of it".
No I think the social context is inappropriate. However I do not think possessing or liking such a picture is perverted. I also do not thinking a cropped version of the picture which has no sexual content is inappropriate.
> The result was visually striking. Repeating tile patterns disappeared, and artists could paint unique detail across large environments without concern for reuse. The primary cost was not GPU throughput, but latency elsewhere in the system.
No, the primary "cost" was artists having to fill a world with unlimited textures instead of just filling memory and then having to make due.
The constraint of "limited texture memory budget" also puts a constraint on how much work the artists can do. Remove that constraint lets artists do unlimited work. It might sound like a plus because "freedom!" but it turns into a minus trying to actually ship on time and at budget.
I get that wasn't the point of the article's "cost", but thought it was worth mentioning.
Time budgets still exist.
This just let you do things like many layers of baked in multi texturing in places where artists had previously ran into engine constraints.
Honestly, having to "make do" when your budget was full probably took more time trying to find neat hacks.
> Texture binds multiply. Draw calls explode. Bandwidth usage spikes. You spend more time feeding the GPU than rendering.
Is this AI?
This is a common rhetorical device for humans named parataxis.
Its just a casual writing style, written like how you might describe it verbally to imply the list of ill effects goes on and on.
Probably just Aspie, judging by some of their other writing (including their About page). I've seen Aspie writing misidentified as LLM output surprisingly often.
Like Vulcans and Data.
A good portion of the world and Lenna herself have asked that image be retired.
Best replacement I've seen: https://mortenhannemose.github.io/lena/
“Good portion of the world” is probably a handful of people.
Her full quote btw:
“Once upon a time, I was the centerfold of Playboy,” says the former model in the new documentary Losing Lena. “But I retired from modeling a long time ago. It’s time I retired from tech, too.”
Well, for instance, it's the official policy of the IEEE to not allow this image in new publications. And they're far from the only journal (or set of journals) that have this policy.
Of course. Most people don’t care and a few vocal ones do.
As in most organizations that would know about it and come into contact with it.
I think my comment is true of the graphics programming/research community.
Given that it's use is banned in most academic journals dealing with imaging/graphics, you'd be wrong.
And as several journals have brought up in the banning, it's not even good at what it purports to be for these use cases. It's a pretty poor quality image to start off with due to being scanned to a digital file with 1970s technology.
At this point the ones defending its continued use are the vocal minority on some weird anti-woke crusade that doesn't even make sense on technical grounds.
You’re using vocal minority framing right now. When I care about it, I’m a weird crusader for caring and noticing. But then you organize a campaign to change it.
There is a large body of literature using these images so it’s helpful to have a comparison which is persistent through time and familiar.
> Given that it's use is banned in most academic journals dealing with imaging/graphics, you'd be wrong.
Critical thinking caps required for this one.
Now at least parts of it are paged out...
Who is Lenna?
https://en.wikipedia.org/wiki/Lenna
The eponymous woman in the Playboy photograph.
A copyrighted image of a nude model elected for no obvious reason has a test image in the University of South California by some pervs and then used in a lot of papers as a test image.
Or, a standard cropped image of a playgirl used in the field of image processing.
"elected for no obvious reason" isn't quite right, as a test image for computer graphics it has regions of very high frequency detail and regions of very low frequency detail which make it easier to spot various compression artifacts, and it makes a good study for edge detection, with both very clear edges along the outline, but more subjective edges in the feathering.
It's redish. Ok it has a blur and details on the foreground but could have been any image with blurred background and a face.
"very low frequency detail", we are talking about a 512x512 picture here, it has low and high frequency details (FFT speaking) like most photos.
"Good for edges detection" doesn't mean anything. Like, is the image good for edge detection or the algorithm is good at detecting edges ? What does "subjective edges" even mean ? Does it mean hard to spot ?
That looks like technical reasons but it just noise. They literally grab a playboy magazine and decided it was well enough (and indeed, it wasn't that bad, yes). Still not professional. The message is "We have playboy magazines at work and we are proud of it".
It’s perverted now?
It's literally cropped pornography.
Is a nude picture perverted?
No. That is not the question. The question is "do you hang out with an erotic magazine at work ?" and "Is it normal ?"
No I think the social context is inappropriate. However I do not think possessing or liking such a picture is perverted. I also do not thinking a cropped version of the picture which has no sexual content is inappropriate.
Which image?