TFStudio 1.7.3

[aai2k]

Changelog

v1.7.3

Added

• Merit Function Editor: the table works like a spreadsheet. Cells select by dragging or with Shift and Ctrl and rows from the row-number column, Ctrl+C and Ctrl+V move a single value or a block of them, every row operation is on a right-click menu with its shortcut, and Type and Pol are cells you type into.
• Explorer: project folders hold folders as well as designs, to any depth. A folder or a design moves by being dragged onto another folder, onto the space below the tree to bring it back to the top level, or from its right-click menu.

Changed

• n,k Characterization: the Sample setting is gone, since the window only ever fitted a film on a substrate polished on both faces. Drude is gone from the model list too, because Drude-Lorentz covers the same metals and takes the number of oscillators from the data. Documentation now describes how to work with this feature more clearly.
• Report: rebuilt from the ground up as a docked window that follows the design as it is edited. Blocks take their settings from the analysis window they mirror, templates replace presets, several designs print as one lettered comparison instead of a report each, and branding, paper size and language are set once for every report.
• Merit Function Editor: the wizard is a three-box form that keeps its shape whichever preset is picked, folds away behind a chevron and remembers its settings; saved merit functions load and save from the table's own bar.
• Material and operand pickers open on the full list with the current entry centred and its category marked, instead of filtered to that category.
• Explorer: a design name only has to be free in the project folder it goes into, so the same name can be used in more than one folder. New designs number from 1 in each folder.
• Settings: one data folder replaces the separate paths for designs, materials and presets, and changing it moves your files there. Paths set in earlier versions are no longer used.

Fixed

• Light themes: panel dividers, control outlines and operand row colours are more visible.
• Keyboard shortcuts work on a non-Latin keyboard layout and with Caps Lock on; Ctrl+Shift+Z redoes.
• Every wavelength axis spans exactly the range computed instead of starting at a round number below it.
• Merit Function Editor: a table of thousands of operands opens, scrolls and redraws as quickly as a short one; typing goes where the cursor is; a single number in a target row sets a flat target instead of leaving the old end value behind as a ramp; and preset block headers name the bands the preset was actually built from.
• Material picker: the search box keeps its focus and keeps up with typing, however many catalogs are installed.
• n,k Characterization: a metal, and a film thick enough that several thicknesses reproduce the measurement, come back as the film that was measured instead of a matching spectrum over the wrong index. A measurement that cannot see the thickness says the value was assumed rather than reporting it as measured, and a spectrum of the coated surface alone is named as such instead of failing to invert.
• Material Editor: a material saved from n,k Characterization opens instead of taking the application down.
• Optical Evaluation: a merit function of thousands of targets draws and resizes without stalling the window, and a range covered by many of them keeps one shade so the curves behind stay readable. A drawn target is set to the angle picked in the drawing bar instead of always landing at normal incidence.
• Torn-off windows: dropdowns and menus open in the right place, close when you click away and follow the theme; a dialog a tool asks for opens over that window instead of the main one. On a Wayland session one can be dragged, though docking it back has to go through the Dock button on its strip.

---

This discussion was created from the release TFStudio 1.7.3.

[xluo3166-pixel]

软件现在的设计已经很强大了，有没有更多的优化方案支持WDM的设计，现在使用WDM功能设计出来的膜系，对角度的变化敏感度比较高，软件的设计思路是，设计-制澡-反演吗，现在这个软件越来越强大了，我看你之前的设计思路里制造是在考虑中还是不考虑，主要是停镀算法的问题。

[aai2k]

The software is already very capable. Are there more optimization options for WDM design? The stacks the WDM feature produces right now are fairly sensitive to angle changes. Is the software's intended flow design → fabrication → reverse engineering? This software keeps getting stronger. In your earlier plans, was manufacturing in scope or not? Mainly the deposition-stop algorithm.

The flow is design → fabricate → reverse engineering, a closed loop like that? I typed too fast above, there were typos.

About the first question about wdm filter design: i'd need to do some more research and testing before I can act on it, but I think I know the cause. Global integer search scores against 0 deg only and it generates cavities of multiple orders like this

Cavities of different orders shift at different rates as filter is tilted so they drift out of tune with each other and the band distorts. If cavities were of the same order that should keep the band's shape further off normal. Don't know. Maybe I can do something about this. Put that in my plans.

The second thing. Well, design - fabrication - reverse engineering is the intended path now. It already can fit measured curves for example. More features for manufacturing are being considered, but this is tricky and hard to get right.

The mono monitoring simulator already uses strategies to cut layers, turning point, level and time. If
there is something specific you need, or something another program does that you want here, I'd take a look

[xluo3166-pixel]

好的，非常感谢

[xluo3166-pixel]

设计思路是，设计-制作-反演，这样闭环来的吗
上面打字太快有错别字。