{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreibk3mu4lxvn7mkujkvka42byxisen3np6ngfasirkfsjc5e5yoxbu",
    "uri": "at://did:plc:ayfizlbmvprdzzpxhepnqrns/app.bsky.feed.post/3mj6d7dvxf6e2"
  },
  "path": "/f/technology/14802",
  "publishedAt": "2026-04-10T16:39:15.000Z",
  "site": "https://jstpst.net",
  "tags": [
    "plenoptic \"light field\" cameras",
    "Google has already used this technology to good effect",
    "patented new ideas for microlens arrays",
    "A COMPETITION FOR AN AI DIAGNOSIS APP, BAFFLINGLY NAMED 'MICROLENS'"
  ],
  "textContent": "# smartphone cameras are doing their best\n\n> **tldr** : You carry your phone with you all the time, and you can practice many of the skills of photography on it. Consider saving money on a camera and spend it on a better phone.\n\nThe resolution of film cameras is _low_ , actually. Some of the best photos of all of history were taken on cameras we'd understand as a complete piece of shit.\n\nSmartphone camera lenses and sensors, however, are _tiny_. They've reached about the limit of what physics can do with their lens and sensor.\n\nTo account for this, \"computational photography\" is employed. It's not some shitty AI sharpening filter, although there is a lot of that in recent days. Proper computational photography mostly traditional algorithms which combine time-series data from the camera sensor, gyroscope, accelerometer, lidar (and depth from other lenses too maybe!)\n\nComputational photography is how we get \"panorama mode\", combined exposures (\"HDR\"), night-mode, etc. We usually exclude the ADC, debayering, and raw-to-jpg conversion from the term, but it's worth stressing that digital photography requires a ton of computation for even the bare minimum. It's why digital cameras chew through batteries like there's no tomorrow.\n\n# consider buying a fancier smartphone instead of a dedicated camera. i am not joking!\n\nAnd an aspiring \"serious\" photographer is someone who (1) probably needs a cellphone, and (2) maybe wants to buy a real camera\".\n\nIf you're in the market for a cellphone and a camera, you might spend $400 on a decent used smartphone, and at least $600 more on the lowest-end used mirrorless camera with three lowest-end used lenses, which you might find too onerous to ever carry with you. And with that camera, you're going to want some accessories at minimum: A blower to get dust off your camera without damaging it, extra caps to cover the lenses, a shoulder strap, a case to put it in, and at least one extra battery.\n\nUsing a Real Camera means carrying around a ton of bullshit.\n\nOr, you could spend $1000 to get a new high-end smartphone. (Or less on a used one!) \"Flagship\" phones come with\n\n  * The choice of three or more fixed (\"prime\") lenses,\n\n  * The ability to control two of three legs of photography (shutter-speed and ISO) as well as focal point\n\n  * The ability to access \"raw\" photos with full dynamic range, allowing you to practice editing, and\n\n  * The ability to photograph with the computational photography disabled. (Mostly- again, you don't get your .jpgs without some computing.)\n\n\n\n\n> **side note** : Of the iPhones, I would argue that the iPhone 16 Pro is currently the best iPhone to use as a camera, and will be for some time. This is for two virtues: The first virtue is the telephoto zoom. The higher the better IMO, and the 16 Pro has the highest of any in its form factor, and will for some time. The second virtue is the \"camera control\" button. This simulates a knob and shutter button, but I've disabled the knob and I use it as a shutter button. This lets me open the camera quickly, and then take a photo, with the same button.\n\nBut it begs to the question... Why are these the two choices?\n\n# the space between \"camera phone\" and \"dedicated camera\"\n\nThere have been a lot of products between the chasm of \"dedicated camera\" and \"smartphone with cameras\". The Samsung Galaxy Camera was just a point-and-shoot (one lens) with a screen, Yongnuo released a mirrorless camera running Android, etc. But they all emphasize providing smartphone apps on a camera, which doesn't really strike me as something anyone would want.\n\nThere are also accessories for using your phone as a better camera: Attachable lenses, tripods and mounts, battery grips, and thunderbolt SSDs that can store video faster than on internal storage.\n\nThere are also smartphone cameras that dedicate a ton of body space to providing a larger sensor, which naturally comes with sacrifices of quality. And with a larger sensor, you preclude the possibility of a proper telephoto lens, since you'll need to dedicate far more internal space for that lens.\n\n# Minimum requirements\n\n> **tldr** : Water-resistant, pocket-proof. Accelerometer+gyrometer+LIDAR on the device, and it links to your smartphone for storing and processing photos. The only thing lacking is a reason to own it.\n\nI think Apple or Google should create a _separate_ camera which pairs _to your smartphone_ , for more than just integration into the native Photos app.\n\nThat separate camera would not need its own screen, would not need as much body (as the processing happens on your smartphone), and would not need much storage (except as a buffer to send data to your phone). It could be little more than a lens and a sensor, while also being able to take advantage of computational photography for nice features like panoramas.\n\nConnectivity would benefit from a short-range, very-high-bandwidth link. Wi-Fi 7, at ~20Gbit speeds, is probably good enough, but we still have a 60GHz \"Wi-Gig\" that's still underutilized.\n\nCrucially, this would need to **fit in our pockets** without risking destroying the whole thing. It should be waterproofed and pocket-proof like a smartphone.\n\nYou might be thinking, \"This product already exists, it's called the Point and Shoot. The only difference is the on-phone processing and lack of screen.\"\n\n# We're ignoring half the information light rays give us: Direction\n\nThere are so many types of cameras out there. Tilt-shift, telecentric, etc. The coolest of these are the plenoptic \"light field\" cameras, such as Raytrix or Lytro.\n\nLytro already made this, providing a pocket-size camera. Rather than use LiDaR or mutiple lenses to simulate depth estimation, these cameras record photo density _and_ the direction of light and depth. This allows the physically-accurate simulation of depth-of-field after the fact (rather than the cheap imitation you get on a smartphone), as well as depth information (very useful for 3D displays, like Apple's trying to do with spatial photos).\n\nLytro had a real product, but is now defunct, with Google having had acquired their engineers. Do you see where this is going?\n\n# WHAT IS GOOGLE DOING WITH LIGHTFIELD CAMERAS\n\n> **tldr** : We almost had a proper 3D camera in everyones pocket recording fine-grained information about the world, and Google threw it away\n\nGoogle produces the smartphones with the best photographic capacities, in no small part due to their capacity for computational photography!!\n\nThe most important thing to happen to photography was the gradual, decades-long process of getting photography into the hands of billions of people, and Android was a significant driver of that!\n\nGoogle has already used this technology to good effect, and they've patented new ideas for microlens arrays, AND THEN THEY DID NOTHING WITH IT\n\nTHE LAST THING GOOGLE DID WITH MICROLENS WAS A COMPETITION FOR AN AI DIAGNOSIS APP, BAFFLINGLY NAMED 'MICROLENS'\n\nA Lytro-like companion camera for smartphones would at once:\n\n  * Satisfy needs for a \"real camera\" for many consumers,\n\n  * Expand accessibility to artistic expression for many consumers, but also\n\n  * Expand the richness of the recorded history that comes with the saturation of access to cameras,\n\n  * Increase the interest in lightfield cameras, and the exploration in ways microlens arrays can be utilized.\n\n\n\n\nNow, the dozen of fans of Lytro cameras are seeing no way this can come back, as \"AI\" can simulate what it did _decently_ enough.\n\n# TLDR?\n\nThere's a product that is obviously aligned with Google and Apple's interest that would also be _really cool_ , and Google owns it and is just SITTING on it. And they have been for YEARS.\n\nI am complaining about wishing capitalism would provide a cool gadget, I know. But understand: The gadget would be _very cool_.",
  "title": "I'm surprised neither Google nor Apple have come out with a dedicated digital camera",
  "updatedAt": "2026-04-10T16:39:15.000Z"
}