Kymogram Beta

Changelog

You are using version 0.4.0.

0.4.0 2026-10-10

  • In a file written segment after segment by a logging loop, plotting one whole channel also summarises the other channels of the same length from the same reading: they then plot in a fraction of a second instead of reading the whole file again.
  • The file is read on a thread of its own while the plot is computed, with SIMD where the browser has it: the first plot of a 3 GB channel takes about 2.6 seconds on a MacBook Air M2.
  • A Parquet export no longer reads the whole file again before it starts.

0.3.0 2026-10-09

  • Opens files of 100 GB and more. A file written segment after segment by a logging loop no longer stops at a limit of the reader.
  • Opens large files at the speed of the disk, and in seconds with the .tdms_index dropped next to them: a 100 GB file opens in under 2 seconds with its index on a MacBook Air M2.
  • Zooming into a channel already plotted reads only the edges of the view: about 0.2 seconds on a 100 GB file. The points are those a full read gives.
  • The first plot of a large channel is about three times faster, and a long plot is drawn as it is read.
  • Opening or plotting a large file shows how far it has got.
  • A Parquet export of a file opened with its .tdms_index no longer reads the whole file again first.

0.2.0 2026-10-08

  • Compare two TDMS files: groups and channels in one file only, properties changed, table rows changed, added or removed, lined up so that an inserted step is one added row, and channels whose values differ, with the first difference, the largest one, and both curves on one plot.
  • A group can be read as a table, one column per channel and one row per value, with a tab per group like the sheets of the Excel add-in. A group of records, text or timestamps in rows of the same length, opens as a table; a group of measurements opens as a plot.
  • Exports the whole file as an Excel workbook, laid out as the Excel add-in lays it out: a sheet of properties, then one sheet per group, one column per channel. Up to two million values.
  • Copies a table to the clipboard as cells, to paste into Excel, a mail or a chat.
  • Two files dropped on the viewer can be compared in one click.
  • Plot and table controls on one line, timestamps written as a date and a time in UTC, and the properties already shown above the property list no longer repeated in it.

0.1.0 2026-10-03, first beta

  • Opens .tdms files in the browser, with their .tdms_index if you have one. Nothing is uploaded.
  • Tree of groups and channels, with each channel’s type and number of values, and a filter by name.
  • Plots every value of up to eight channels together, with one Y axis per unit. A peak one sample wide stays visible at any zoom.
  • Statistics of the visible range: minimum, maximum, mean, standard deviation, RMS, number of points.
  • Reads big-endian, interleaved and multi-chunk files, and files truncated by a bench crash. What cannot be read is reported with its segment.
  • Exports a group to CSV, with a decimal comma option for Excel in French, German and other locales, split at Excel’s row limit.
  • Exports a group to Parquet, with exact column types, read back by pandas in the tests.
  • Works offline after a first visit.
  • Dark theme by default, with a light theme one click away. The choice is remembered.
  • Legal notice and privacy policy.
  • Cursors on the plot: two positions, the time between them and its frequency, and the value of each channel at each cursor. Positions can be typed.
  • Save the plot as a PNG image, with its legend.
  • Batch Extract: drop a folder of TDMS files and get one table, one row per file. Minimum, maximum, mean, standard deviation, RMS, first or last value, duration, value at a given time, total and longest time above a threshold, over the whole run or a time range, and file, group or channel properties.
  • Batch Extract limits: a lower or upper limit per value, values out of limits marked high or low, and a PASS, FAIL or NO DATA verdict per file. Limits can be changed after reading without reading again.
  • Batch Extract exports to CSV and to an Excel workbook with failures in red, plots any column across files, and overlays a channel from up to eight runs on one time axis.
  • Batch Extract can watch a folder in Chrome and Edge, and adds each new file once the bench has finished writing it.
  • Batch Extract column sets, limits included, are saved to a file and loaded back.
  • Beta: every feature is free, with no account and no limit on the number of files. Pricing may change after the beta.
  • A CSV export also writes the properties of the file, the group and each exported channel, with their TDMS types, to a second CSV.
  • Repair: a file damaged in its middle, by a bad disk for example, is read on past the damage, keeping only what can be attached to a channel with certainty. The skipped bytes are reported. A repaired copy can be saved as a new .tdms file.
  • In Chrome and Edge, the site can be installed as an application that opens .tdms files from the file explorer, with a double-click once it is chosen in Open with.
  • Several files dropped on the viewer can be opened in Batch Extract in one click.
  • Sample files: a bench run for the viewer, and five endurance runs that fill a Batch Extract table with its verdicts in one click.
  • A plain site: a drop zone on each tool page, the detail folded under How it works, and a page for addresses that do not exist.
  • Counts a few anonymous events, such as a file opened or an export finished, to know whether the tool is used. No identifier and nothing from your files. The list is on the Security and privacy page.
  • Opens a 1 GB file in about a fifth of a second, and zooms in it in under a quarter of a second, on a MacBook Air M2.