Templot Club Archive - Search Results
To search again, click the Back button in your browser.
Search results for: algorithms
16 results found.
... Windows 95/98/ME is going to struggle with bitmaps on the sketchboard and will be limited to the native vector graphics. I'm getting a bit concerned that I'm creating a major support load for myself from users saying that stuff doesn't show on the printed output. The image zoom levels in Templot are far beyond anything normally needed in other software.) I'm not too sure what you are asking me about the grating panels? Obviously as you zoom out the smaller details are going to be lost. There are several resampling algorithms available for the sketchboard, but the very highest quality is just too slow to be practical. Here are a couple of sketchboard views. I set the trackplan in red so that the rails would show over the table. They line up with your rails perfectly: 2_022223_010000001.png 2_022223_010000002.png Here's the workpad view. Your dimension of 28ft-8.5in radius for the brick ring gives 229.67mm diameter in 4mm scale. I measured it with the ruler tool and you can see that it is spot on. 2_022223_000000000.png Many thanks again for your work ...
... more hours before it is anywhere near fit to be released. Even then, for every template correctly chaired, there will be dozens loaded from existing plans showing chairs in the wrong place or the wrong size. Just to get that far it needs a massive new dialog to allow users to specify prototype chair dimensions and locations, and a new data structure in the box file to contain the information. For example, I have spent 2 days -- so far without sufficient success for customized templates -- trying to create an algorithm to put the CCL and CCR chairs on the ends of the check rails. Bearing in mind that the timbers under them can be shoved, and equalized or square-on, and that the check rail lengths, flare lengths, and flare angles are all infinitely adjustable. Not to mention that even P4 has an oversize flangeway, so the prototype chair sizes don't fit. In 00 and EM and other scales, the oversize flangeways and end gaps make a mockery of the prototype chair sizes and flare angles. The only ...
... together, but is only at 72 dpi resolution, and as stated is quite small in size. See second image below. I was wondering if, the algorithm's, mentioned below, that you used to modify the sketchboard images might be of help. Lanczos3 Triangle Mitchell Bspline Bell If so, where can I get these algorithm's so that I can experiment with them? All the best, Brian Nicholls. 1853_271020_290000000.png 1853_271021_340000000.png posted: 27 May 2012 23:23 from: Martin Wynne Brian Nicholls wrote: where can I get these algorithms so that I can experiment with them? Hi Brian, Unfortunately you have missed the boat with the old-maps site. Until recently it was possible to zoom in and display the original scans for screen capture. This option has now been removed, presumably to encourage you to pay £16 for their PDF files. There has been some discussion about this on the OS maps Yahoo group. You would probably do better to order a photocopy of the map from a reference library or the Warwickshire county archives. The full ...
... this is causing so much confusion. I would have thought the situation is blindingly obvious. "Is this right?" And then the two options given are Yes and No. I'm puzzled, Now you say, there are only two intersections, in reality, there is only one true intersection where the centre lines of the two tracks cross each other, since the two roads diverge away from each other, in either direction, to infinity. I can only assume that the other is a factious one calculated mathematically by your algorithm and the maths formula it uses, presumably using a general case calculation which takes into account both positive and negative numbers. Which is why when you deliberately reproduce the offset notch position, it will always appear at the same point in space, since it apparently selects the negative number calculation to display the notch, unless this is the calculation that is done last just before the display appears. Not knowing the algorithm you use, I cannot comment further on the mathematics involved. However, I can only say, that if ...
... are unusual on the prototype -- designs vary a lot and don't use conventional chairing for the vees. If you have a prototype drawing (there is a 1:1 drawing in the GWSG GWR track book, p.54 -- see below) you can create one using partial templates and much timber shoving. By the way, a 90-degree crossing is 1:0 angle -- a 1:1 RAM crossing is 45 degrees. The current bottom limit in Templot is 1:2. Below that the timbering algorithm falls over. I intend to reduce the limit to 1:1.5 in a future pug, but it will still need manual timbering intervention. If you are familiar with geometry and computer programming I'm sure you will see the problem with a 1:0 angle and hear the alarm bells ringing! regards, Martin. posted: 17 Sep 2007 20:03 from: Jamie92208 Hi Martin I've got a short angle crossing to construct. Is it possible for you to post a sample timbering for say a 45 degree one so ...
... topic: 481 Rail rolling problem posted: 9 Jul 2008 01:15 from: Chris Mitton I've come across what might be a snag in the algorithm for rolling rail-joints along a series of templates (or else I'm doing something wrongly!). If a section of plain track is made up of a series of templates of differing curves/ transitions/ whatever, and one of them is a very short fill-in, the joints don't roll across the short length properly. What seems to happen is that if the rolled-in length is longer than the template, ie the template is too short to contain any rail-joints, it is ignored; consequently the next template along has no information about how much to "roll-in" so it starts again, leaving you with an odd length in the middle. Regards Chris posted: 9 Jul 2008 03:08 from: Martin Wynne Hi Chris, Many thanks for reporting this bug. Sorry about that. I will get it fixed. For a workaround in the ...
7. iPlot
... when touched. The iPad can be used in either landscape or portrait modes, automatically switching as you rotate the iPad. This is particularly useful for turnouts that are vertical on your layout. iPlot can guess the formation you want if you overlay 2 simple templates. Eg a curve and a straight or to curves coming together can be used to define the basis of a turnout and two crossing templates can be used to define a diamond crossover. But I've been waiting for Martin to bring out the next pug with the curved crossover algorithms, so the current iPad version is a bit ropey here. One function that has been missed out is the use of scans. But it is replaced by the photos function. Simply take a photo with the built in camera, use hold pinch-zoom to correct any perspective errors and pinch-zoom to get it to the same scale as you template. One final advantage is you don't need to print out. You can simply put a clear protective layer over your iPad, select the hold mode, and build ...
... in many cases were supplied undrilled, to be cut and drilled on site to fit as required. You have probably noticed that the relevant menu entry is disabled and marked (nyi) -- not yet implemented. It would be easy to insert the data I have here into the program as defaults. The difficulty is that you can't just mark a rail joint. The timbering has to be adjusted accordingly so that the timbers each side are at the proper joint spacing of about 24"- 26" centres. A full algorithm to timber half-diamonds needs a lot of work, allowing for 3 joints between the K-crossing and V-crossing for all angle combinations. Unlike the prototype we also have to do that for 00, EM, TT, etc., not to mention narrow-gauge. Already there are two options for the way the timbering supports the point rail tips. You can add extra joint marks to a template manually using severely shoved timbers -- add a bonus timber and then reduce its width and length to ...
... Click the real> K-crossing options> fixed K-crossings menu option to override the automatic setting. regards, Martin. posted: 17 May 2012 04:20 from: Martin Wynne Paul Boyd wrote: I remember you said once that one reason for not having very short diamonds in Templot is that the timbering gets complex- not if it's turned off, it's not! Hi Paul, Well yes and no. I do feel that the current 1:1.5 limit* is the sensible minimum for the present generator algorithm. It will go shorter, but would need a knowledgeable user to understand what they are seeing and adjust things accordingly. There are check rail conflicts, truncated wing rails, and other artefacts to correct for. And the prototype varies so much that it's impossible for Templot to know what to do about them by itself. I'm afraid a lot of my design decisions nowadays are based not on providing extra functionality but on minimizing support requests. You have only to look at this very topic to see how even the essential functions ...
... where the S&D crossed the GWR main line. This is an irregular diamond of around 1 in 1.9 105_211505_400000000.jpg More pictures here. Cheers posted: 21 Jul 2013 20:29 from: Martin Wynne That's excellent Paul. Arthur, diamond-crossings shorter than about 1:4 are usually built on waybeams (longitudinal timbers) as in Paul's example. Designs varied between companies and each one was normally created as a special. For details of a GWR 1:1 diamond-crossing, see: message 812 The timbering algorithm in Templot falls over below about 1:4, so you need to shove the timbers to match the drawing by widening them, twisting them round and crabbing them to lie under the rails as waybeams. regards, Martin. posted: 22 Jul 2013 13:49 from: Brian Nicholls Hi All, Does anyone have prototype photographs of such an obtuse crossing, showing the actual waybeams? If so please post here. All the best. Brian Nicholls. posted: 22 Jul 2013 19:01 from: LSWRArt Thanks ...
... grid lines in a much enlarged view I can see short breaks where the shape lines would cross them. In order to make them "visible" do I simply have to thicken the lines of the shapes? Hi Stuart, You beat me to it. Yes. output> line thickness> set any line thickness... menu item. Also of course they need to be set to some colour other than white. They are probably going faint because you are viewing the image at less than 100% using a resampling algorithm. The default settings are intended to be only 1 dot wide at 100% size, which is fine for printing, and should show on the screen as thin lines in PNG format. Your files are looking ok here at 100% size. (If you use the JPG format, all bets are off. Don't do it.) See also: topic 2848 regards, Martin. posted: 5 Mar 2016 21:51 from: Martin Wynne p.s. Stuart, This is a section of what I'm seeing on ...
... , timber lengths are the same in EM as P4 just the spacing that alters, my question is "why" John posted: 18 Apr 2017 15:35 from: Martin Wynne Hayfield wrote: As for the shorter timbers on Templot, timber lengths are the same in EM as P4 just the spacing that alters, my question is "why" Hi John, There is no prototype timber length for bullhead turnouts. Each railway company had their own standards, which they didn't always stick to. Templot has its own internal algorithm for deciding the timber lengths, based on its best guess of a typical size. If the result is not what you want, that's what the shove timber functions are there for. One click on the lengthen or shorten buttons will change each timber to match your timber base. I will make you an EM A-5 template which matches the Exactoscale A-5 base. regards, Martin. posted: 18 Apr 2017 15:41 from: Hayfield Martin Thank you posted: 20 Apr 2017 00:28 from ...
... .. my B-12 was daft and unreasonable but you pull out an A-20 and call it just unusual! The idea of branching left to come across the main track( or two if branched with a tandem) is nice. I have often seen it used( although usually only with a single turnout) on tight formations in the USA. Locations like access to cramped buildings around New York harbour when the cars come off the carfloats. Tim Warris probably used them somewhere. Good work with updating the SST algorithm. Thank you. Shall we confuse the new users even more by banging in loads of acronyms? Regards Rob posted: 20 Jan 2018 22:04 from: Trevor Walling Hello, Shall we confuse the new users even more by banging in loads of acronyms? That could lead to even more questions. Best wishes Trevor. posted: 20 Jan 2018 22:21 from: Tony W Hi Martin. Your A-7, A-20 Tandem is on a par with one which I built recently for my storage ...
... unless you post a more recent photo of yourself, the likeness or otherwise will remain a mystery to us. That photo has a lovely fine grain. Scanning my negs at 3200dpi I've become aware of just how grainy they are. Much more so than I remember when printing them. Perhaps I had a soft enlarger lens, or is it possible for the graininess to increase with age? It's only gelatin holding the grains of silver apart, they've had plenty of time to clump together. For these I've tried the softer Bell algorithm for resampling down to web size, instead of Lanczos: This sad scene of desolation is Ludlow on the same day in 1969 as the Clee Hill pics: ludlow_1969_1280x640.jpg And getting back to the scenics, Dinham Bridge and weir in Ludlow: ludlow_dinham_1969_1280x800.jpg It is little changed today. Google's view from the bridge: http://goo.gl/maps/PnYYXLJZpGp cheers, Martin. posted: 12 Apr 2018 07:09 from: FraserSmith Hi Martin In relation to your mountain road photo, I looked along a few roads in ...
... it used to be as Martin has automated some of the processes. Phil. posted: 21 Jun 2019 15:18 from: Martin Wynne Julian Roberts wrote: I had the no doubt naive idea that Templot would be able to look at an individual turnout (or slip or diamond or whatever) on the map and zap out a model size template to construct that turnout. Hi Julian, If only. But nowadays the Templot code is available open source, so if you or anyone else feels able to write the AI algorithms to do this, do please get on with it. Templot is 40 years old this year, and we can't wait for ever! You could of course simply print the map at model size and build track directly onto it, without regard to the niceties of prototype track design. No doubt some modellers have done that. Templot will print the map at full model size if you put a short bit of dummy template on each page. But do I in fact have to design it, and adjust it to fit ...
... working over them, which makes it a lot easier to see what you are doing -- here on the left: 2. low-contrast converted original: 2_241521_120000001.png But it still doesn't match the screenshot captures from smooth browser resampling. So I have found some old resampling code from Anders Melander which could be used in Templot, as a one-time conversion on the background shapes. It's very slow, several seconds per tile, but only needs to be done once. The difficulty is that there are 7 different resampling algorithms available, with varying degrees of smoothing and speed, and each one can have an adjustable edge-effect setting. Mostly they are optimised for colour photographs, rather than scans of historic maps, so the default settings are not necessarily the best for our purpose. And then in each case there is the degree of zooming to apply. An utter minefield of different settings which would take me forever to evaluate. From which I need to select 1 or maybe 2 options to offer in the background shapes. Here are some ...
Search took 0.085 seconds
Back to Templot Club