2 hours ago, Wobbie said:

We sadly lack any firm definition of what exactly is meant by a 'light year', as used by Ianbourne'. 

A light-year is a unit of length used to express astronomical distances and measures about 9.46 trillion kilometres or 5.88 trillion miles. As defined by the International Astronomical Union, a light-year is the distance that light travels in vacuum in one Julian year.

So, to use 'light year' to compare graphics is simply nonsensical!

A better word to use is 'better'

 

Thank you. That was my whole point. But the usual crowd just jump on you in here and try to dismantle you. 

17 minutes ago, Ianrivaldosmith said:

Thank you. That was my whole point. But the usual crowd just jump on you in here and try to dismantle you. 

You’re absolutely right. Maybe we can ask the mods to ban him for such flagrant use of the term ‘light year’ personally I’m appalled.

Good job you guys were here to correct him who knows where we may of ended up. Thanks again! 

33 minutes ago, Ianrivaldosmith said:

Thank you. That was my whole point. But the usual crowd just jump on you in here and try to dismantle you. 

Okay, the new MFS isn't "light years" ahead of the current sims as a light year is a unit of measurement.  Got it.

 

However, the new MFS is a huge generational LEAP over any current sim out there today.  In stock default form.  The DLC for the new MFS will also be a generational leap over the current DLC for today's sims.

Any sims we own today are about to be vastly outclassed by the end of this year.

Edited June 28, 20205 yr by Mengy

I like how everyone spends 2 pages discussing one phrase I used while completely ignoring the actual topic I raised, namely that the giant tree problem reveals potential difficulties with AI recognition of trees and ground shrubs from 2D aerial photos, which in turn has an effect on the scenery depiction in the sim.

Perhaps Asobo is picking sensible posters on here for admission to the alpha, meaning we're left with the squabbling silly billies?

Just now, lambourne said:

 

 , meaning we're left with the squabbling silly billies?

Not all but some,  yes 😄.   

The vegetation rendition is not bad but not perfect, or rather not as good as it should/could be. And it is not a minor detail  which shouldn't obscure "the big picture" as some enthusiasts say.  Your remarks on London City were absolutely justified in this regard. But the expertise of the AI (if it is a true AI) will progress over time as AI learns if properly tutored. So I am still an optimist on the issue, thanks to people like you who care to say what they think in a well  documented post. 

Don't be discouraged by the silly billies 😁 !

Edited June 28, 20205 yr by Dominique_K

Spot the seven differences ...

Bottom pic is Google Earth.

 

Ok, I'll start ....

1. There is an aeroplane in the top one 😇

Dave.

 

1 hour ago, Dominique_K said:

Not all but some,  yes 😄.   

The vegetation rendition is not bad but not perfect, or rather not as good as it should/could be. And it is not a minor detail  which shouldn't obscure "the big picture" as some enthusiasts say.  Your remarks on London City were absolutely justified in this regard. But the expertise of the AI (if it is a true AI) will progress over time as AI learns if properly tutored. So I am still an optimist on the issue, thanks to people like you who care to say what they think in a well  documented post. 

Don't be discouraged by the silly billies 😁 !

I'm sure, that vegetation will be presented as good as today's technicaly possible. But we have to know that there doesn't exist any relevant database of vegetation type and height for every small place on Earth and today's technology isn't able correctly recognize vegetation type and height from satelites image with top view. Earth surface is much bigger, than we can imagine ever and it's absolutely impossible manualy edit vegetation on whole Earth. So unfortunately we must accept, that they must use some similar to landclasses for vegetation placing. There isn't any better automatic way. But I'm sure, that vegetation in MSFS will be represent reality much better, than previous type of sim.

Lol, yes Dave ... and clouds too.

Even thought the FS2020 image looks already stunning in its own right, what we can say about FS trees in general is that:

• There often are many of them, also in places where in reality there are none, or there are bushes.
• They sometimes are too large, like bottom left they are taller than the 7 story buildings.
• They tend to have just one dark green color while in reality many trees tend to be lighter green and in a forest more different shades of green are mingled.

Also water tends to always have this one very dark blue color, while in reality rivers often are more brownish. And have less ripples.

Also I wonder why AI houses are generated, which makes them deviate from the real houses, while the real houses are available in 3D. Or maybe London is not available in 3D in Bing, I haven't checked.

But hey, didn't I just say myself not to look too much at all these details ... I'm already very happy with how FS2020 looks as it is ... to me it looks quite immersive. There's always room for some fine tuning later.

Tree sizes

 

Edited June 28, 20205 yr by RudyB24

31 minutes ago, RudyB24 said:

Or maybe London is not available in 3D in Bing,

It isn’t.

To those who have taken a look at Google Earth, especially in VR, trees are, upon inspection, clumps of faceted polygons that up close seem more like a type of green obelisk, than an actual tree.

Of course, GE and BING Photogrammetry as a general rule fail at close range, but trees especially lose any resemblance to the real thing under direct scrutiny.

To LM or LR, I suspect this would be an insurmountable problem that we would simply have to live with, either permanently, or at least for several years.

Instead, almost certainly via the Azure AI, Microsoft/Asobo have managed to convert the vast majority of those trees worldwide into a reasonable close-up facsimile of an actual tree.

To me, it seems hard to overstate the impressiveness of the accomplishment, matched by Asobo also somehow applying window lighting to nighttime photogrammetry,

Given that type of proven competence, its hard to believe that the further step of correcting tree heights is not within the developers capability, and in fact was probably on their internal road-map before we ever started complaining about it.

For myself, I'm not particularly worried about the "tree problems" eventual resolution, and honestly we've gone on about it long enough that hammering further on the issue, I suspect, is unlikely to move things along any faster.......

 

15 hours ago, Mengy said:

That would be me, my ISP is utterly terrible.

Invite me to the alpha Microsoft, I'll gladly test the bejeezus out of that offline mode for you!!!!  😉

I'm an Alpha tester and I also live in the Slow-Internet-Universe but I have never even tried the offline mode in the new sim, always used full online mode. It works a lot better than I expected.

3 hours ago, lambourne said:

Perhaps Asobo is picking sensible posters on here for admission to the alpha, meaning we're left with the squabbling silly billies?

I think the reason why no one initially really picked up on your post is that the tree height topic has been brought up for months in what feels like every single thread and I personally started to viscerally blank these posts out. The devs know and I don't think there's any conceivable scenario in which they don't do anything about it.

Edited June 28, 20205 yr by threegreen

Just now, ludekbrno said:

I'm sure, that vegetation will be presented as good as today's technicaly possible. But we have to know that there doesn't exist any relevant database of vegetation type and height for every small place on Earth and today's technology isn't able correctly recognize vegetation type and height from satelites image with top view. Earth surface is much bigger, than we can imagine ever and it's absolutely impossible manualy edit vegetation on whole Earth. So unfortunately we must accept, that they must use some similar to landclasses for vegetation placing. There isn't any better automatic way. But I'm sure, that vegetation in MSFS will be represent reality much better, than previous type of sim.

The strength of AI seems to me its ability to dig into enormous amounts of hugely varied sources ( "big data")  to discern patterns and predict. Not just automate a process by looking into a database. Wloch (I think) said that they were running the AI on a network of 20 000 computers (twenty thou). He also said they tagged 3 trillions trees... 

Predicting variety and height of trees will become an important tool to  manage forests. And I'd bet that the people doing the job for FS20 are not far from other people working in these matters.  

 Asking for trees to be  at the right scale, colors, varieties within limits is  no way derogotary, on the contrary it pays tribute to the work already done. 

Not an expert myself, just a fascinated old man.  The last site I visited  https://www.20tree.ai/ . You will find many more.

14 hours ago, LHookins said:

I think once you've seen the actual sim you might change your mind.

 When it comes to him, I wouldn't bet on it!

While trees may be too big in some places, I also think that the buildings are too small.  Compare the building heights to the width of the runway in @RudyB24 s  shots above.  (and I really think we should have actual approach photos and not Google Earth) and the Google Earth shot needs to be closer to runway which would make the buildings to the left on this side of the river appear even larger.

 

and the air molecules are too big as well...and not properly polarized.

Edited June 28, 20205 yr by sightseer