Hacker News Reader

Comment by ShadowBanThis01 | Hacker News Reader

Is this yet another DJI drone that's not compatible with DJI's own goggles?

If so, it's a baffling blunder they've made multiple times. This one would even be more embarrassing, since they call the secondary camera a "POV" camera.

Another issue: DJI continues to rip people off for a CinemaDNG "license." CinemaDNG is a free codec that is not encumbered by licensing fees. At $16,500, this scam adds insult to injury: "The format is unencrypted and free from intellectual property encumbrances or license requirements" https://en.wikipedia.org/wiki/CinemaDNG

The thing looks like a very cool professional tool. But these repeated offenses are getting old.

View on HN

From the same wiki page

The adoption of CinemaDNG among camera manufacturers appears to be hindered by Red Digital Camera’s patent US9245314, which covers in-camera recording of lossless compressed raw video. There had been an unsuccessful attempt to invalidate the patent.

DJI gets around this by moving the recording circuitry outside of the camera. It's in the body of the drone, not in the camera.

Are you implying that is sufficient to circumvent RED’s patent? IANAL but I would argue a camera-drone is still a camera.

Besides that, what you describe sounds like a decision made to reduce weight of the camera module, possibly not related to patent.

BS patents tend to have nonsensical loopholes.

"Are you implying that is sufficient to circumvent RED’s patent? IANAL but I would argue a camera-drone is still a camera."

Yes, the Inspire 2 does this, and it was widely reported that this is how they dodged the BS patent. The weight distribution may also be another bonus, or vice versa.

I actually don't know why there's this perception that the patent only applies to camera-internal compressed raw, because Apple/Atomos have to pay Red royalties but Atomos recorders are obviously external. But you hear it over and over.

The bar for novel/non-obvious seems really low if this patent has been successfully defended.

There is almost no bar now. The USPTO is derelict, and an ignorant ruling allowed what was supposed to be illegal: software patents.

Even worse is that Red simply ripped off JPEG2000 and got a patent on it.

The patent "system" has become a tool for exactly what it was supposed to prevent: the theft of people's work.

Thomas Edison Was a “Patent Troll” (2014) https://slate.com/technology/2014/05/thomas-edison-charles-g...

The only saving grace is the relatively short expiration of twenty years. if it was seventy years after the author's death progress would grind to a halt.

IP expiration dates are a "brain plasticity" parameter in society. Higher values mean lower rates of change, too low values mean too high rates of change. Twenty years is a sweet spot.

20 years is eternity in software. It may as well be forever.

If I remember correctly, the patent is simply compressing the RGBG bayer filter sub-pixels as separate R, G, and B images. That's almost the entirety of it, and some rubber stamper at the USPTO granted it.

Compressing it using JPEG2000. You can actually decompress the image using an off-the-shelf JPEG2000 decoder.

Here's hoping GPT-4 will be used to flood USPTO with trivially "novel" patents, until the right people are forced to admit that the system no longer works at all and needs serious reform.

Imagine an AI trying to defend it self from future better AI by jamming the patent system.

It’s wild (and scary) how many plot points from Accelerando have recently leapt into the realm of plausibility.

Sounds like an interesting book, I’ll check that out

AFAIK it costs something like $10,000 to file a patent, although most of that is legal fees that maybe GPT could reduce.

I imagine in this case a bulk discount would apply.

Do not need to register. Just generate a lot of text based on small derivations of existing patents and put them in a searchable public database (you will have good embeddings anyways) with a clear cryptographic time stamp on it.

Now if a new patent is filed the patent office cannot ignore it as prior work. While the bar for machine generated patents is high. There is IMHO no reason why prior art cannot be machine generated and if something very similar can equally generated by a machine it becomes even more questionable if something is patentable at all.

Nikon is fighting the good fight: https://ymcinema.com/2022/12/08/red-vs-nikon-jury-trial-is-s...

We're all rooting for them.

The Inspire has never been intended to fly with goggles. It has a dedicated "pilot view" FPV camera, but in normal usage it's typically operated by a crew of two, a pilot flying line-of-sight and/or via FPV camera and a dedicated main camera operator, both using similar/identical remote controllers with a display attached (this is typically either a large tablet or a dedicated "proper" camera monitor).

Why would it have to be "intended" for that? If someone wants to do it, why prevent it? They say outright that the thing uses the same generation of wireless transmission ("O3") as their current goggles and stand-alone camera & transmitter.

Not to mention that they make a big deal out of being able to send video to two destinations at once, ideal for a feed going to goggles.

Is the drone "intended" to be carried around by hand while filming? It doesn't have handles for that, and yet they show a guy doing it in their promo video.

Project requirements / intended use is important. This is like asking why Betaflight doesn't support automated flight profiles or loiter. It's not a bad question or unreasonable request, but can be answered as "out of scope / not a commonly-requested feature".

That answer would be hard to validate here. It already has the necessary wireless-transmission feature. And if the feature weren't desired, why would it exist on ANY drone?

Look up "cinewhoop" drones. DJI even makes a drone that addresses this market. It just doesn't have the image quality that this one does.

I'd say simply supporting an already-existing DJI video receiver is way more of an expected use case than carrying a drone around in your hands while filming.

> And if the feature weren't desired, why would it exist on ANY drone?

Different market segments. To draw earlier on my analogy, loiter and automated flight exist in iNav, PX4, and Ardupilot.

Are you the target market for this product? Do you plan to buy one? Have you talked to people who are? Do they want goggles support?

Yes, I am. If it supports goggles, I will most likely buy it. And I did get a response from B&H to a question about it today, saying that many customers have indeed inquired about this capability but there is no information from DJI about it yet. Here's the exact statement from B&H:

"Unfortunately DJI has not provided further insight as to whether these two will be compatible. We hope they are, with a provided firmware update in some months, as customers have really desired another non-racing fpv headset."

For what that's worth...

I don’t know. If the drone happens to have Bluetooth and a speaker, should it just obviously support being used as a Bluetooth speaker just in case someone wants to do that?

Um, yes, since it was built with those functions.

I’m not well-versed in the UAS laws globally, but in Canada you basically need to have a spotter if the PIC doesn’t have unaided line of sight to the aircraft[1]. DJI tends to go above and beyond the actual legal limitations when it comes to aviation safety, so this doesn’t surprise me a whole lot. It’s often frustrating; my Mavic Mini 1 often prevents me from doing perfectly legal things in my own yard because I live pretty close to an airport. I can unlock it, but last I checked it required using a laptop to fill out some forms and was quite a bit more complicated than the usual 30-second setup that I love.

[1] Aircraft less than 250g do not have the same regulations, which is why the TinyWhoop family of aircraft are super popular here for FPV racing.

Given the use case for this drone (medium-high budget media production) it's extremely unlikely that a person is going to be flying this drone without a spotter. Not having support for their own FPV googles on a $17,000 drone is kind of disappointing.

The RAW upgrade is still significantly cheaper than Arri's RAW codec. Not sure if the Arri price is still the same, but it used to be higher than the cost of this drone to unlock RAW on an Alexa. It was also just an image sequence rather than a wrapped container format.

Is this defending one rip-off by citing another?

Kind of, yeah, it is. In high end film/video gear, you know you're always getting over charged for anything. So this is one of those, just think, you could be paying more situations. Some of that pricing for hardware does tend to make a difference in quality in how rugged and durable something is to survive the rough and tumble environment of a production set. Can it survive the multiple remove from case, build, use, tear down, throw back in case, survive cases being tossed around like they are luggage loading onto an airline all in on day for each company move on the schedule and still be usable the next time you use it. This is software only type of pricing that is ridiculous

I love how some Arri shills down-modded that. Or are they DJI shills?

HMMmmmmmm

That was on the older Alexa's, there is no licence required to unlock raw on Arri cameras.

Do you know if they are still image sequences or have they chosen to wrap them up into an MXF or similar wrapper? It's been quite some time since I've been around Alexa footage.

Old guy story: the original RAW image sequence named the files by converting the time code to frame number and append that to the sequence name. So, midnight/0-hour timecode would be frame 0, and 23:59:59:23 would be whatever that number is that I'm too lazy to convert. However, if someone forgot to sync timecode to time-of-day as is pretty standard OR if someone was actually shooting late night and the clip was started before 23:59:59:23 the frame number conversion would see 00:00:00:00 flip around and dutifully use frame number 0. this would cause most NLE software that would recognize the image sequence to create a clip starting with frame 0 which was technically in the "middle" of the shot that starts in another clip based on standard importer logic. </rantOfAnOldDIT>

Reply on news.ycombinator.com