J-20 5th Gen Fighter Thread VI

[Inst]

latenlazy: do you own a DSLR? Did you ever notice that your camera has a "distortion correction" feature on by default, and if you turn it off, distortion pops up? This does not need to be done in post, although PP often does a more thorough job. It affects the JPEG or camera-processed image, and is very useful since the JPEG is already pre-processed and each time you reprocess an image, information is lost (try sharpening then unsharpening repeatedly in Lightroom with a JPEG).

As to evidence, we're back to epistemological debates. At the end of the day, none of us will have access to the gold standard, which is detailed test information compiled by Chengdu, so we are always fudging to some extent. I think your problem is that you think in terms of binary; i.e, there's legitimate evidence and there's illegitimate evidence. That's not how PLA-watching works; i.e, we have evidence that is more credible and evidence that is less credible, alongside evidence that is not credible at all. Uncredible evidence simply can't be used, but when we're looking at videos we're working with less credible evidence. As with my canteloupe example, less credible evidence that has been binned together is as good as more credible evidence.

 

[Inst]

@ougoah: The F-22 STR claim is relatively worthless, to be frank, you don't know what altitude or airspeed the STR occurs at. The gold standard for maneuverability information is an E-M diagram, such as this:

.This shows maneuverability at different speeds at a given altitude. Note STR is about 14 degrees / second at 4572 meters, and that at lower altitudes, aircraft maneuverability increases but speeds do not. For instance, we have 20-24 degree/sec STR figures for the Eurocanards and the Flankers, but we do not know at which altitude and which speed this occurred at; i.e, they could have higher turn rates at 500 meters at corner speed, or the figure could actually be higher due to the stated ITR occurring at non-peak maneuvering speeds and at altitude.

 

[latenlazy]

latenlazy: do you own a DSLR? Did you ever notice that your camera has a "distortion correction" feature on by default, and if you turn it off, distortion pops up? This does not need to be done in post, although PP often does a more thorough job. It affects the JPEG or camera-processed image, and is very useful since the JPEG is already pre-processed and each time you reprocess an image, information is lost (try sharpening then unsharpening repeatedly in Lightroom with a JPEG).

Yes. I’m a hobby photographer. Not all cameras have a distortion correction feature, and not all built in camera distortion correction fixes all distortions. Do you even understand how the software works?

As to evidence, we're back to epistemological debates. At the end of the day, none of us will have access to the gold standard, which is detailed test information compiled by Chengdu, so we are always fudging to some extent. I think your problem is that you think in terms of binary; i.e, there's legitimate evidence and there's illegitimate evidence. That's not how PLA-watching works; i.e, we have evidence that is more credible and evidence that is less credible, alongside evidence that is not credible at all. Uncredible evidence simply can't be used, but when we're looking at videos we're working with less credible evidence. As with my canteloupe example, less credible evidence that has been binned together is as good as more credible evidence.

Some “fudging” is worse than others. I don’t think in terms of binary. I think in terms of error bars. How do you assess what is credible and what isn’t? Ideally you should base that on standards of evidence and analysis. You can concede being less rigorous on that front, but that doesn’t absolve that your methods fall below some standard. Your cantaloupe example shows that you have a very weak grasp of this. Law of large numbers has its limits. It can’t get around flawed measuring instruments. If your instrument is fundamentally inadequate it doesn’t matter how many samples of data you collect. The statistical aggregate of that data is not measuring what you want it to measure.

 

[Inst]

For DSLRs, my camera from 2011 had lens correction built into the camera. This is from 7 years back; i.e, any DSLR built in the last 7 years has lens correction. I don't claim that lens correction is perfect, but it cuts out most of the lens distortion given that the lenses and cameras are being produced by the same manufacturer, who input the necessary information into the DSLR.

As to epistemology, you just said you think in terms of binaries, i.e, what is credible and what isn't.

Second, law of large numbers fails when there's a systemic bias; i.e, the canteloupe scale is always going to be -1 kg as opposed to having an equal chance of being wrong in both directions. If we're looking at a large pool of different videos shot using different lenses, where is the systemic bias? Basically, using your argument, lens distortion means that our satellite photography image that we used to arrive at 20.8-20.88 length for the J-20 is not credible due to distorting factors.

Your problem is that your obsession is with being "right or wrong". If you want to avoid being wrong absolutely, you can't really make any judgments at all. I'm happy having a degree of error in assessing evidence and working with "less credible" evidence because otherwise we have no evidence at all. It's intellectually dishonest to say that from the evidence provided, we can't tilt in the direction of the J-20 being less stealthy than its American counterparts by just saying "all of the evidence provided is not credible" as opposed to "all of the evidence provided is less credible".

 

[latenlazy]

For DSLRs, my camera from 2011 had lens correction built into the camera. This is from 7 years back; i.e, any DSLR built in the last 7 years has lens correction. I don't claim that lens correction is perfect, but it cuts out most of the distortion of

That doesn’t mean all videos or photos we see are taken with it. Also the DSLR stuff is kind of a red herring. Most of the videos we get seem to be taken with smartphones. Lens distortion is also only ancilllary to my main point (and I should apologize a bit here for not making this clear in my initial response).

You just said you think in terms of binaries, i.e, what is credible and what isn't.

Where did I say I think in terms of binaries, or weigh credibility based on binaries?

Second, law of large numbers fails when there's a systemic bias; i.e, the canteloupe scale is always going to be -1 kg as opposed to having an equal chance of being wrong in both directions. If we're looking at a large pool of different videos shot using different lenses, where is the systemic bias? Basically, using your argument, lens distortion means that our satellite photography image that we used to arrive at 20.8-20.88 length for the J-20 is not credible due to distorting factors.

In a large pool of videos the systemic bias is in the lack of commonly shared flight parameters and shooting conditions. How do you know two different videos of the same plane involve the same flight settings? Or are shot at the same focal lengths with the same distance to the camera? You don’t.

Read what I actually said. I wasn’t just talking about lense distortion. Space is rendered differently for objects shot at different distances with lenses of different focal lengths.

Your problem is that your obsession is with being "right or wrong". If you want to avoid being wrong absolutely, you can't really make any judgments at all. I'm happy having a degree of error in assessing evidence and working with "less credible" evidence because otherwise we have no evidence at all. It's intellectually dishonest to say that from the evidence provided, we have to tilt in the direction of the J-20 being less stealthy than its American counterparts by just saying "all of the evidence provided is not credible" as opposed to "all of the evidence provided is less credible".

It’s intellectually dishonest to say that some conclusions are likely or certain based on evidence that doesn’t substantiate such claims. This is not an obsession with being right or wrong. This is basic intellectual rigor. Either way, whether I have “an obsession with being right” is really no defense for the objective quality of your analysis, and seems to be a real bizarre thing to fault in a hobby that’s all about trying to get the most accurate information to the best of our abilities.

 

[Air Force Brat]

That doesn’t mean all videos or photos we see are taken with it. Also the DSLR stuff is kind of a red herring. Most of the videos we get seem to be taken with smartphones. Lens distortion is also only ancilllary to my main point (and I should apologize a bit here for not making this clear in my initial response).


Where did I say I think in terms of binaries, or weigh credibility based on binaries?


In a large pool of videos the systemic bias is in the lack of commonly shared flight parameters and shooting conditions. How do you know two different videos of the same plane involve the same flight settings? Or are shot at the same focal lengths with the same distance to the camera? You don’t.

Read what I actually said. I wasn’t just talking about lense distortion. Space is rendered differently for objects shot at different distances with lenses of different focal lengths.

It’s intellectually dishonest to say that some conclusions are likely or certain based on evidence that doesn’t substantiate such claims. This is not an obsession with being right or wrong. This is basic intellectual rigor. Either way, whether I have “an obsession with being right” is really no defense for the objective quality of your analysis, and seems to be a real bizarre thing to fault in a hobby that’s all about trying to get the most accurate information to the best of our abilities.

Gentlemen,,, I'm very happy that you both share my interest in aerodynamics and aircraft performance, really, I do appreciate both of you, I really do! but we are far afield here, and distracted from our very honest search for the truth about the Chengdu J-20, Master Deino will be disappointed in us, and I for one hate that as well, you have both communicated some very helpful information, and I have learned a few things,, but now, lets let it rest here??

the video that Deino posted does indeed show the J-20 doing a "power loop", (Max's term). The video does indeed show that the J-20 is very powerful in the vertical, and has great pitch authority even at very low energy/airspeeds, and I for one am very grateful that we were able to share this video on Sino Defense..

I would encourage each of you to read Deino's article on the J-20, it has been cited as a source by others already,, I'm looking for a hard copy, I missed it in the bookstore???

 

[Inst]

By the way, one correction about lens distortion. In the photography field, there's a very common misunderstanding about wide vs telephoto vs normal lenses. People seem to think any lens that is not a normal lens creates additional perspective distortion or reduces perspective distortion. That's a myth. Lenses don't create or remove perspective distortion; it's the photographers who do so. Wide and telephoto have different FOV, but photography usually involves a subject filling up much of the image. This means the photographer needs to get closer or further to take a photograph of the subject, creating the perspective distortion or lack-of associated with wide angle and telephoto lenses. The only real distortion we need to consider is the classical lens distortions, which are usually software-corrected either by the camera or in post.

 

[latenlazy]

By the way, one correction about lens distortion. In the photography field, there's a very common misunderstanding about wide vs telephoto vs normal lenses. People seem to think any lens that is not a normal lens creates additional perspective distortion or reduces perspective distortion. That's a myth. Lenses don't create or remove perspective distortion; it's the photographers who do so. Wide and telephoto have different FOV, but photography usually involves a subject filling up much of the image. This means the photographer needs to get closer or further to take a photograph of the subject, creating the perspective distortion or lack-of associated with wide angle and telephoto lenses. The only real distortion we need to consider is the classical lens distortions, which are usually software-corrected either by the camera or in post.

Between a long lens and a wide lens, for an object of the same size in the frame, the object in the wide angle lens is closer than the long lens, and it’s relational distance to the background is different. That relational distance of an object in a camera frame to its background also changes at different distances from the camera. In other words, you *need* to know both the focal length the image was shot at *and* how far away the image was from the camera in order to glean any useful information about depth.

 

[Inst]

In a large pool of videos the systemic bias is in the lack of commonly shared flight parameters and shooting conditions. How do you know two different videos of the same plane involve the same flight settings? Or are shot at the same focal lengths with the same distance to the camera? You don’t.

Read what I actually said. I wasn’t just talking about lense distortion. Space is rendered differently for objects shot at different distances with lenses of different focal lengths.

It’s intellectually dishonest to say that some conclusions are likely or certain based on evidence that doesn’t substantiate such claims. This is not an obsession with being right or wrong. This is basic intellectual rigor. Either way, whether I have “an obsession with being right” is really no defense for the objective quality of your analysis, and seems to be a real bizarre thing to fault in a hobby that’s all about trying to get the most accurate information to the best of our abilities.

The lack of commonly shared flight parameters is NOT systemic bias and actually is its precise opposite. Systemic bias would instead be, say, if you're aiming to take a random sample to determine the incidence of childhood malnutrition in a population at large, but you end up setting up your sampling station in a poor village. The figure shown would be representative of the poor village, but not of the population at large, defeating your purpose.

As to conclusions being likely or certain, we're in an equivocation game again. You're using likely or certain to indicate that both are the same, but they're not. Nor am I even saying that likelihood or certainty matters when we're running individual video analyses; what matters is that the accuracy is statistically significant for the total sample set.

===

The point of video analysis currently is basically to catch Chengdu screwing up. We know what EM-envelopes for aircraft tend to be, what we are looking for is somewhere where the J-20 is ecxeeding normal EM-envelopes. What we look for right now is a shot of the J-20 exceeding normal parameters so we can say that the J-20 does have unusual agility. Or, alternately, with enough videos, we can conclude that the available evidence suggests that the J-20 is not exceptionally maneuverable, corroborating the claims of retired PLA pilots.that its subsonic maneuverability is merely "good".

 

[Inst]

Between a long lens and a wide lens, for an object of the same size in the frame, the object in the wide angle lens is closer than the long lens, and it’s relational distance to the background is different. That relational distance of an object in a camera frame to its background also changes at different distances from the camera. In other words, you *need* to know both the focal length the image was shot at *and* how far away the image was from the camera in order to glean any useful information about depth.

This is incorrect. You can test it yourself. If you have a zoom lens, shoot a photograph of an object, centered to minimize barrel distortion, at your longest focal length. Then shoot one of the object at your longest focal distance, from the same position as you shot it from the telephoto, at your shortest focal length. Then crop the wide angle until its FOV is the same as that of the telephoto picture. You should see about the same damn thing.