Analysis Pool

The Pool is the main working area of Deepsona. It describes an analysis as a chain of five columns: you put data into Source, optionally condition it in Filters, choose what to compute in Analyses, reduce the result to numbers in Statistic, and decide where it goes in Destination. Pressing F5 runs the whole chain. The pool stays as it is, so the same processing can be repeated on new measurements at any time.

The Five Columns

Source: Adding Measurement Data

• Drag and drop files onto the pool. Dropping a folder adds every supported file it contains directly. Supported inputs include .hdf, .wav, .dpsa, .atfx and .ldsf.
• Or use Insert → Source (add HDF)…, or right-click the empty area of the Source tree and choose Add directory….
• Expand a file to see its channels. The file check box is tri-state: checked means the whole file participates, the intermediate state means only some channels are selected, unchecked means the file is excluded from the run.
• Only checked files and channels take part in a run. This is the quickest way to restrict a long measurement to the two channels you actually care about.
• Double-click a channel to play it back.
• Right-click a file for Player, Mark Editor, Edit with → Channel Editor or RPM Generator, and for Copy as path, Locate in Navigator and Locate on disk. The red × button removes the file from the Source pool (the file on disk is untouched).

The Insert Dialog

All columns except Source are filled through the Insert dialog. Open it with Insert → Filter… / Analysis… / Statistic… / Destination…, or by right-clicking the column and choosing Insert….

• Category list on the left. For the Analyses pool the categories are ALL, Native, Native Only, ASX, ASX Only and My Favorites. The other pools have their own categories, for example Basic Filters / Filters, or Documents / Deepsona Files / Other Files.
• Filter box at the top (Ctrl+F). Typing filters the list live and matches both the displayed name and the original English name, so either spelling finds the item.
• Three columns in the list:
  • Name — the item name in the interface language.
  • Original name — the English name. Result pages, chart titles and exports use this name, so the column lets you match a chart back to the item that produced it.
  • License — for analyses, the accuracy rating (see below); for the other pools, the option the item belongs to.
• Multi-select insert. Ctrl-click or Shift-click to pick several items and insert them in one go; you can also drag across a run of rows to select them, and starting on empty space gives a rubber-band selection instead. A double-click inserts the item under the cursor. The Filters, Analyses, Statistic and Destination insert dialogs all behave the same way.
• Favorites. Right-click a row and choose Add to Favorites. Anything you mark appears under My Favorites in every future session, which saves a lot of scrolling when you always run the same handful of analyses.
• Show unlicensed items at the bottom controls whether items that your license does not cover are listed. They cannot be inserted; the dialog reports which ones were skipped and inserts the rest.
• Insert keeps the dialog open for the next pick, Insert & Close inserts and closes, Close leaves without inserting.

Accuracy Ratings and Paired Rows

In the Analyses pool the License column shows how far an analysis has been verified:

Many analyses appear twice: once under Native and once with the suffix _ASX under ASX. The plain row computes the result with the built-in Deepsona engine. The _ASX row sends the same analysis through an external calculation path, which requires that path to be installed and licensed. Inserting both puts the two results into the same result page so their curves can be overlaid and compared; result titles carry a prefix that identifies which path produced them. Analyses that exist only as a built-in engine are grouped under Native Only and have no paired row.

Two Properties Shared Across Analyses

Control channel — the x-axis reference

For “versus something” analyses such as FFT vs RPM, the horizontal axis is no longer restricted to rotational speed. The control channel setting in the analysis property panel lets you use any channel as the x-axis reference — torque, throttle position, temperature, or anything else that varies with the operating condition.

• Leave it empty for automatic detection (by channel name and unit). The panel shows (auto).
• Order-based analyses use a separate setting — the order speed channel. An order is defined relative to rotational speed, so it cannot be referred to anything else; the two settings are kept apart and do not interfere.
• Direction: a falling reference works too. If “all directions” produces an empty plot, the reference is usually not monotonic over that section — bin advancement is one-way, so only the monotonic part of an oscillating section is collected.
• The step size follows the axis range rather than being fixed.

Output speed trace

The Output speed trace option in the analysis property panels (off by default) decides whether an additional speed-versus-time plot is produced.

The speed channel never takes part in the spectral computation — it is a reference input to the analysis, not a measured quantity. This option only affects whether the extra plot appears. Producing it for every analysis wastes a cell and squeezes the spectrum you actually came to look at, which is why it is off by default. It is available on the 1/n octave analyses, the CPB family and the autospectrum.

Running the Pipeline

• Press F5, click the Run button in the toolbar, choose Run → Run pipeline, or right-click the empty area of any pool column and choose Run pipeline.
• The run covers every checked source file and channel and every enabled pool item, in column order. Unchecked items stay in the pool and are skipped.
• If the pool cannot be run — for example when Source or Destination is empty — a message explains what is missing and the focus returns to the column concerned.

Results

• Results open in a Data Viewer tab next to the Pool tab. If the Destination column contains Mark Analyzer, the Mark Analyzer tab opens instead; these two destinations are mutually exclusive.
• Run → Overwrite result page on re-run is enabled by default: pressing F5 again reloads the page that the previous run opened instead of adding another tab. Switch it off if you want one tab per run — but note that many open pages consume a lot of memory. To keep one result and still re-run, duplicate the page from the tab menu first.
• Each run also writes a result bundle to C:\Vlisten Analysis Results, named <source>_Analysis_<date>_<time>_<id>.dpres. These files are listed in the Analysis Results section of the sidebar and can be reopened later. If that folder is not writable, the bundle goes to %LOCALAPPDATA%\AnalysisWorkbench\Dvlisten instead and a note is added to the event log.
• Run → Open results folder opens the folder in Windows Explorer.
• Files produced by an export destination (WAV, XLSX, HDF and so on) are written to the folder configured on that destination node.

Working in the Pool

• Pool lists support extended selection: Ctrl-click, Shift-click and marquee dragging in the blank area. Ctrl+A selects all, Del deletes the whole selection.
• Right-click an item for Properties, Cut, Copy, Paste, Rename (F2) and Delete. Renaming is worthwhile: the name is carried into result titles and reports.
• Selecting an item shows its parameters in the properties panel. Changes apply on the next run.
• The Filters and Statistic columns can be collapsed into a narrow vertical strip when they are not in use; click the strip to expand them again. Column widths are adjustable by dragging the separators.

Practical Tips

• Build the pool once, then keep re-using it: drop a new measurement into Source, uncheck the old one, and press F5.
• Mark the analyses you use daily as favorites — the My Favorites category is usually the fastest route to a working pipeline.
• Use the Original name column when a chart title is unclear; it is the exact string that appears in the result page.
• Add a Single Values Table or an Excel destination alongside the Data Viewer when you need numbers for a report rather than a picture.
• Keep the number of open result tabs small. Overwriting the result page on re-run is on for exactly this reason.

© 2026 Deepsona. All Rights Reserved.