defaid

This will be long and wordy but first... @hjwalter Hans, to answer your original question, the only inbuilt mechanism for specifically an image is the computer's graphics engine, which has constantly to resize every texture file in order to create the illusion of distance. BGLs are a different matter and I can imagine that someone has managed to compile one one that does create mips from base textures. Seems redundant though -- the processing power to create mips on the fly and then render them would be exactly the same as simply resizing and rendering the base image. I won't be at all offended if nobody reads this. I won't know unless some masochist points out a mistake... @jgf Thank you for adding that. I thought my explanation was reasonably clear but perhaps it wasn't. It could be a language barrier, I suppose. For anyone who just happens along and has the fortitude, I'll reiterate... First, image compression. DXT3 degrades image quality, right from the outset. And it gets worse with every hard save. I'll come back to that. All compression degrades image quality. Not just game textures and not just the DXT format. Every time you edit and save one of your photographs as a jpeg the image quality will diminish. Zoom right out from your newly compressed photo and it probably won't look so bad. Zoom in and you will see more of the degradation. The same thing happens with music (except for a few formats like .flac) though most of our ears can't hear the difference between a wav and an mp3. Decompressing a compressed image doesn't repair the damage. It is permanent. Converting DXT3 to 888-8 doesn't improve the image quality. I do it to downloads because 1) I'm likely to edit in the future and 2) I'll forget what format it is. I resave downloads as 888-8 to ensure that my future self doesn't mess up. There is some good news though. Image editing software doesn't actually save the file when you click 'Save'. What it does is make a temporary file of what's essentially metadata comprising the sum of all the editing that you've done. The compression part of the metadata is only applied to the image file itself when you 'Save As' or when you save and close. You can save as often as you like as long as you keep the image open, and only one compression will be done, right at the very end. DXT images are compressed. 32 bit are not. So, for the same number of pixels, DXT3 uses up much less disk space. They are arguably quicker to load into FS9 just because they are fewer MB in size but I've never noticed a difference, not even on my old XP machine. On a slight tangent, resizing is also a form of compression: 2048 × 2048 down to 512 × 512 means a file 1/16 of the kB size but the different compression formats we have been discussing don't alter the image dimensions, only the quality. Second, mip mapping. It's one of the more interesting aspects of mathematics that just as you can map a piece of a flat image to a point on the surface of an imaginary 3d model (atlas to globe), so you can map a different piece of the same image to the concept of 'further away'. Referring to the ISS images in my earlier post (which as usual came from Wikipedia), mips don't affect near objects because for close-ups, the highest-res base image is always used. The explanation for mips being necessary dates back to the 20th century. The size of a building, a tree, an aircraft on your monitor is not 30 metres wide by 40 metres long. Its real size changes as it seems to move closer or further from your virtual viewpoint. Sometimes the object may be 1000 pixels wide. A few minutes later it could be 30 pixels. Graphics hardware and software have to edit textures to create the illulsion of distance and perspective. Imagine an FS9 building falling behind you as you taxi past it. As you go by, looking out of your cockpit window, it is a rectangle covering 500 × 500 pixels of your monitor. Taxi onward and look back over your virtual shoulder. The object is now a trapezium 50 pixels tall and 10 pixels wide. The graphics stuff has to shrink the texture to make it look further away and has to skew it to account for perspective. Shrinking has to be done on a per-pixel basis so that's 262 000 pixels that have to be processed for a 512 × 512 texture. It has to be done for every texture in every frame of your fps. For a 2048 × 2048 it's nearly 4.2 million. However, if the texture has mips, then much of the work has already been done. While you are near the virtual building, the graphics engine will still select the high-res, unmipped base image in the texture file. As you taxi onward, the graphics engine will select one of the pre-shrunk images from a different piece of the texture file and then only has perhaps 64 × 64 pixels to process. This is the purpose of mip mapping: reducing the load. Fifteen or more years ago this was absolutely essential: small ram, weak processors, onboard graphics... Nowadays, small textures like those in FS9 -- a maximum of 2048 × 2048 -- are no trouble for modern equipment and mip mapping is not necessary. If you have an old PC and you find your framerate dropping then mip mapping textures may help. It's worth being aware that it may not: the framerate could be affected by the number of polygons in the frame, online weather or even by complex systems in the player aircraft. Lots of things affect it. But... modern PCs don't struggle with unmipped textures in FS9. D

Except possibly from the subjective perspective of personal preference, that's absolutely not the case. As I wrote above, mip mapping has no effect within a certain view distance because up to that point, the largest image will be used and not any of the mips. In that circumstance, mipped is exactly the same as unmipped.. Compressing to DXT-3 degrades the crispness by introducing dithering: blurring of edges and blending of colours as an artifact of the compression process. On top of that it seems also to reduce the bit depth of each channel, which alters the colours themselves. That happens at all draw distances and by the nature of textures, is most obvious at close range. Here's a snippet of a downloaded Shanghai livery for a Fruit Stand plane. I opened the 32-bit version (the repainter was farsighted enough to include it in the zip) and saved a couple of times as DXT3. Once was enough to introduce the artifacts but I made them more obvious by saving a couple more times. Whether or not one or the other looks better in game to any specific person is entirely subjective but compressed formats are lossy and degrade the crispness. D First the 32-bit original: Second, the DXT3 format:

I don't think that's how mips work. Each grade of mipmap is just a smaller, lower-resolution copy of the base texture in an image. Just as different areas of a single-file aircraft texture are mapped to different parts of a 3d model, so different areas of a mipped texture file are mapped to different distances. That is to say, each draw distance will use a different part of the mipped image. If there are no mips, the high-res base image is used at all draw distances so you effectively get full texture detail right out to the horizon. Image editing software such as DXTbmp doesn't show the whole set of mips in one image but mostly they would look like this: Unpipped texture used at all distances: Mipped texture. Larger file but easier on the graphics memory because for distant objects, the graphics hardware doesn't have to load the large hi-res image, just a small blurry version: I strip the mips from all my aircraft textures because high-resolution modern PCs, graphics hardware & drivers, and monitors don't really need them. (As an aside, I also convert them to 32-bit as I sometimes edit them and each hard save of a compressed image degrades it). Stripping them makes the file slightly smaller. (Decompressing them makes them bigger...) The fundamental resolution of a modern monitor is good enough that -- mostly -- the sparkling and the moiree pattern just don't happen. Modern processors and memory don't struggle with the high-res, unmipped textures even when handling hundreds of them at a complex airport with heavy AI. In the past I've also stripped them from other textures but occasionally came across problems so I leave those as the developer made them. D

This, combined with the jitter in your original post, suggests that you have a dirty or worn USB plug or socket. It could even be a broken solder joint in that port on the Acer. The supposition is lent weight by the fact that the joystick works ok when plugged into a different machine. I guess you've tried different ports on the same machine. If they all gave the same result then it's likely to be the USB plug on end of the joystick's lead. (If one machine has slightly tighter sockets then it wouldn't show the problem). Either that or a dirty pot but I'd expect that to give the same result on all machines. VP2's suggestion is good. Have you tried resetting the sensitivities in FS9 or even -- temporarily, so make a backup first -- creating a new cfg file? I didn't know W11 had a prefetch folder. In any case, it won't affect USB or joysticks. It's only a close-to-hand folder in which Windows can put copies of stuff it expects to need later in a hurry. D

Richie, if you know your way around PCs and still haven't had any success, try running FS9 again and, after it's failed to launch, take a look at the windows event log. There should be more info there. D

Just read the first investigation report in this month's AAIB bulletin and had flashbacks to flying into Billy Bishop in Toronto before Christmas, when I forgot I had a backup AI and magnetic compass. https://assets.publishing.service.gov.uk/media/6617eb186b4cf65594d1ea4b/AAIB_Bulletin_4-2024.pdf No 28V dc ess bus, no 28 V dc stby bus, no 28V dc emer bus. No option for Ctrl, Alt, Del... or even a pause to make a cup of tea.

This suggests the gear was still rolling despite the brakes being set. I've forgotten which sim you use but is it possible to alter the effectiveness of the brakes (for FS9 either aircraft.cfg or the air file)? It would allow you to restore the more cooperative wind when landing.

Who needs realism?

Hi Hans. The file in question is taxiway_marks.bmp in the folder ../fs2004/Texture/ If there is a fault with the file then it's probably in the alpha channel. Attached a zip with the original for you to compare. D Here's the zip: taxiway_marks.zip Here's the normal file: Here's a modified file with black transparency bottom left:

Can I just point out that if you're circumnavigating England and Scotland, you can't follow the coast all the way? D

*Sigh* I'm definitely in the wrong plane. How long to fly the entire coast? This happened in 1986 around the UK. If the BBC are cooperating, the video should be embedded in their page but 1986... low res. The target time was apparently half an hour, though given 500 kt, I think that's just the BBC making stuff up. https://www.bbc.co.uk/archive/qed--round-britain-whizz/zms7rj6

I should clarify that I meant the fourth stage would be d j's last. I'll be without her from Sydney onward. The maps for the fourth stage show the terrain rising to between 3000 - 5000 feet so we're a little higher than we were for the rest of the challenge. Makes a change from being drunk. Where we went. Smallfry is an easy plane to fly but, looking at some of the other contestants' times, I wish again that we'd brought the Mosquito. I'll be doing the east coast run solo hence the change of method described below. I expect it to take around nine hours and who knows yet how many fuel stops. We adopted a slightly different approach to the timing this time. Hitherto I'd set the throttle and left it fixed so that each leg of each stage was planned at a slightly faster IAS. This time I made constant throttle adjustments in order to maintain the same IAS throughout the flight. It worked well, adding a minor task to the cruise but much simplifying the planning & timing. Our overall result for the entire fourth stage was 24 seconds late. No prizes for guessing who has a warehouse at the end of Bankstown's 11R. * * * * * * We leave Melbourne at 0002z, two minutes past our intended departure time. Off the ground at Moorabbin at 0004Z We're back to a world of two halves, all blue and gold, though as we progress generally eastward, the gold becomes increasingly green. The horizon a knifecut. Mangalore and Tocumwal go by without much disturbing us and, at 5000 feet, we probably didn't disturb them either. YMNG, YTOC and YNAR Somewhere between Narrandera and Parkes, as we cross a pond NE of West Wyalong, the south east horizon creases and we know we're coming to the last stretch of the stage. Over Parkes, eyeing the creased horizon. Passing overhead Forbes town, there's a Skylane on the radio. I watch the airport slide under my wing but never manage to spot the Cessna. There's a marked change of direction at Parkes: a right turn through 60° that puts higher ground right in front of us. d j tells me it's Orange but I'm sure that colour's called green. The ground a little closer. Orange town, Orange airport, left and right respectively. As the ground starts to rise, we fly into RAAF Richmond's weather which, aside from the wind, starts as more of the same: clear smooth air. Different terrain, different vegetation, different wind: there's a distinct sideways shuffle to our motion now with 10 kt of wind coming over the right wingtip and it's not long before we spot a dark line of hills ahead, complete with attendant cloud. Clouds and a dark line on the horizon. Bathurst passes less than half a mile below us but the ground is still rising. The timing is definitely easier when maintaining a steady airspeed. Seconds from Bathurst. The magazine New Scientist has a readers' questions section to which other readers can reply. I recall one letter asking why the bottoms of clouds are flat while the tops are bumpy. It was the only one I was ever able to answer. Blackheath & Katoomba's plateau ahead. At the smooth cloudbase, the air was all bouncy little thermals. d j points out similarities between the local and Welsh topography: a lot of the high ground tops out at the same altitude, with valleys cut into it. In English, it's "hilly" but the French call this ground "vallonnée": valleyed - approaching it from the opposite direction as it were. Try vallonnée on Google Translate though, and you'll find they give a context-based translation, offering 'hilly'... We aren't sure of the reason for the shape here but, assuming that it's still sandstone and other sedimentary stuff, we decide it must be rivers, which we confirm later with a map. It's accepted that in a lot of Wales most of the valleys were cut into what was originally flattish ground by glaciation and subsequent river erosion, and that the glaciers weren't around long enough to finish the job. Blackheath and Katoomba are on the plateau in the middle of the map, immediately south of the obvious scar of the Grose Valley. We cross the A32 between Blackheath and Katoomba, perfectly situated on the very edge of the plateau and we pass overhead a rather neglected Katoomba airfield. If you have FS9 with VOZ, keep an eye out for it. Mine may be missing something. Looking online later for the Grose Valley, the canyon that delimits the north eastern edge of the plateau, I found we'd arrived at much younger rocks -- Triassic sandstone around 250 million years young. Not much in Wales of that age; just a small sliver in the north east corner. If I ever visit NSW, Katoomba and the Grose Valley will both be on the list. Looking south over Katoomba perched on the edge. From there it was all down hill. We'd been intending to make a straight in approach to one of the 11s but at the last minute I decided to call Tower. They gave us 29L. "Oh, you idiot." "Yeah. Sorry." Better open the taps. Duly opened. Left downwind. It was a somewhat stressful approach and landing, fast and with only inches of tarmac to spare. Full stop at YBTH at 0408Z. We thought Katoomba was a possibility. I see it now

Here's the plan for our run over to Sydney. Sedate ain't the word.

Surely it could never happen to a lady like that?

d j and I spent a couple of hours studying the last stage of the southern route. After staring at maps for a while, we both looked up and said "Screw that." The original route had again skirted all the interesting terrain in favour of unrestricted landing in case of a failure. In fact, looking at maps, it drew a good approximation of a huge DME arc around Australia's highest peak. Thinking back, we realised that the previous leg had also very adroitly avoided everything that deviated from the horizontal, sneaking between Mt Gambier and the Grampians before taking the flat ground between Warrnambool and Geelong. We weren't going to let that go. *** *** *** We both felt cheated by the absence of interesting terrain. Calculation for the final southern stage told us we'd probably have fuel for a deviation totalling around 260 nm. It added a couple of extra hours which was fine but the marginal fuel situation was unappealing. Further discussion, and knowing that this would be the last stage for d j, led to a modified plan. We both wanted something more challenging than a heading and a time. "Day off?" I nodded. "Let's be tourists." The plan was laid. Day trippers. We wandered off to find some lunch. I was disappointed not to have had a meat pie floater while I was in Adelaide. The owner of the nearest greasy spoon to the airport offered me a bung fritz but I wasn't sure what he meant so we left. We looked it up later. Bung fritz. #YouCanHaveMine *** *** *** 28th January. The following morning, the new race number is done so we take a taxi into town to pick it up. The smartarse printer has added a decimal point. We don't complain -- it matches the ever-increasing sortie stickers under my window. A rather more aggressive race number. By half past ten we're back in the Arrow and ready to do a little sightseeing. The weather has improved: fluffy white cumulus to break up the flat blue and a couple of degrees cooler with the wind blowing in over the bay. Battery on and the sweet sound of the gyro spinning up. The day involves 60 miles to Eildon Weir VOR (the 100 m waterfall doesn't appear in FS9) then double that to Corryong NDB, then south of east and sneak into the mountains looking for the big one. We've planned a return along the south side of the hills. Terrain, for a change, requires up to 6000'. Lumpy air at the cloud base. What it looks like and what it feels like. Eildon Weir VOR looks across to the 300 foot Snobs Creek Falls in the real world. Judging from photos it's not in a single fall but it does look like a nice spot. FS9 has a rather dull stream but at least the mesh is right. FS9 on a recce and what Google showed. Lake Eildon with a grillion campgrounds and marinas. Terrain rises ahead so we do too, to 5500' and smooth air once we've cleared the edge of the cloud. "Well, you've had your squint up the Snobs. It's time to climb." d j's view of the north side of the Alpine National Park. Yesterday, while we were deciding where to go, d j found reference to The Man from Snowy River. The statue at Corryong caught her imagination and it's enough to persuade us to stop there for lunch. The poem concerns a young chap, the only one of his band brave enough to take big risks for big wins. He became the prototypical national hero emerging from the dying days of British colonialism - a straight-up and courageous no-bullshit bush cowboy. Being British, we'd obviously never heard of him. Corryong town and airport are in the valley to the right and the NDB sits on the forested hill directly ahead. Wind 271@05 so probably a right downwind for runway 24. d j makes the call. Corryong NDB on the hill ahead. Gosh that happened in such a rush. Then I realised, 2× sim rate... The man and his horse, courtesy of Wikipedia After lunch we have a short hop around the north end of a little hill, over the border into NSW, through a gap to Khancoban hydro electric reservoir and through another to a small grass strip that lies on a longer and probably more interesting way up Australia's highest, Mount Kosciuszko, than the iron road & chairlift from the south. The mountain's name has had a difficult time in getting itself applied to the right lump. Wikipedia says "The confusion was straightened out in...". Not over here, it wasn't. The before start checklist says "Flightplan: review." d j says "We're departing in the wrong direction. That Cessna did a left pattern over the town. You can make a climbing right turn and head back onto track between the airfield and the NDB. "Flying outbound the NDB on 108° should take us north of the hill that's in front of us now, and through a gap to Khancoban. Hang a right over the reservoir and another saddle will spit us out right on final. "Don't mess up-- it's in a deep hole." Khancoban with transmission line wayleaves and Khancoban Pondage, the hydro-electric reservoir. Over the saddle. That's YGHI right on the nose. Time to slow down -- and drop all the flaps. And recheck the sim rate. It's a steep approach to a small field. Mount Kosciuszko is just behind the skyline and looks just the same. Steep approach to a small field. Backtracking. Quite a hole. After a very brief stroll, we make a rather claustrophobic departure from Geehi camp. All I could see was high ground. We've had Wagga Wagga's cavok for much of the northern half of the trip but Canberra's cloud adds spice for just a couple of miles either side of Geehi. We re-enter the Alpine National Park, immediately southwest of the Kosciuszko National Park. Forested hills up to 1800 m and sub-alpine habitat above that. I swear I can smell the air from this side of the monitor. When you're crossing the hills, it's very obvious why the B&H took the line it did. There's absolutely nowhere to let down. Homeward bound. RW time is getting late so we go back up to 6000' for the southern side of the Alpine NP and fly Livingstone radial 33 inbound. We skip Mount Hotham airport, which is just north of the VOR and miles from the mountain of the same name, and race back over the flattening and sinking ground following R-246 outbound until Melbourne comes up. Direct to ML 114.10 for a while then back to Moorabbin's NDB, to park up in time for drinks before dinner and the start of the final stage. Lining up on 17L.