Troubleshooting EthoVision Paste Track: Common Issues and FixesTroubleshooting EthoVision’s Paste Track feature can be challenging when recordings don’t behave as expected. This article covers the most common problems users encounter with Paste Track in EthoVision XT, explains likely causes, and gives practical steps to resolve them. It assumes familiarity with basic EthoVision terms (arena, detection settings, filters) and that you have access to your experiment files and the Paste Track data.

---

What is Paste Track and when to use it

EthoVision’s Paste Track lets you import pre-recorded position data (for example, from another tracking system or manual annotations) and play it back as if it were tracked by EthoVision. This is useful for combining external tracking data with EthoVision’s analysis pipeline, running custom visualizations, or reprocessing older datasets with modern analysis settings.

---

Common issue: Imported track won’t load or shows “invalid file” errors

Symptoms:

• Paste Track import dialog rejects your file.
• EthoVision displays “invalid file format” or a generic read error.

Likely causes and fixes:

• File format mismatch — EthoVision expects a specific text/CSV layout. Check that your file follows the format required by your EthoVision version (time stamp/frame, X, Y, optional headings). Export again from the source with a plain CSV or tab-delimited text file.
• Encoding problems — ensure the file is UTF-8 or ANSI without BOM. Re-save using a plain text editor (Notepad++/VS Code) and choose UTF-8 (no BOM).
• Decimal and separator mismatch — EthoVision expects decimals and separators matching its locale (comma vs. period and comma vs. semicolon). Replace separators or adjust OS/EthoVision locale settings.
• Hidden characters or headers — remove extra header lines or non-numeric metadata. The first data row should be numeric values or a single header row matching EthoVision’s expected column names.
• Wrong coordinate origin or scale — if the file contains negative or extremely large values, verify units (pixels vs. mm) and convert if necessary.

Quick checklist:

• Save as CSV (comma) or tab-delimited plain text.
• Use consistent decimal mark (.) unless your EthoVision locale requires comma.
• Remove extra headers and footers.
• Ensure timestamps are monotonic and start at zero or the expected frame.

---

Common issue: Track appears but is offset, rotated, or scaled incorrectly

Symptoms:

• The animal’s path appears outside the arena or in the wrong orientation.
• Path is too small or too large relative to the video frame.

Causes and fixes:

• Coordinate origin differences — external tracking may use top-left origin; EthoVision may use center-origin or vice versa. Apply an offset transformation: translate coordinates so the arena origin aligns. In many cases, subtract/add a fixed X/Y offset to all points.
• Axis inversion — Y-axis may be inverted between systems (increasing downward vs. upward). Flip the Y coordinate: Ynew = frameHeight − Yold (if using pixels).
• Rotation mismatch — if the tracking system used a rotated camera or the arena was defined differently, rotate the track points by the needed angle around the arena center. Use a simple 2D rotation matrix:

  
  x' = cosθ * (x − xc) − sinθ * (y − yc) + xc y' = sinθ * (x − xc) + cosθ * (y − yc) + yc 

• Scale mismatch — convert units (pixels ↔ mm) by multiplying coordinates by the appropriate scale factor. Confirm scale from the video calibration or arena size.
• Check arena definition — confirm EthoVision arena coordinates and video calibration match the transformed track.

Example transformations:

• Flip Y: y_new = height − y_old
• Translate: x_new = x_old − offset_x; y_new = y_old − offset_y
• Scale: x_new = x_old * scale_factor

---

Common issue: Jittering, sudden jumps, or noisy tracks

Symptoms:

• Track shows rapid small jitter around a general position.
• Sudden unrealistic jumps between frames.

Causes and fixes:

• Sampling or rounding issues — imported data may have lower temporal resolution or rounding. Interpolate missing frames or upsample by linear interpolation to match EthoVision frame rate.
• Coordinate precision — low-precision coordinates lead to quantization jitter. Re-export with higher precision (more decimal places) when possible.
• Units mismatch causing apparent jumps — ensure consistent time base (timestamps vs. frame indices). If timestamps are irregular, resample to fixed frame intervals.
• Filtering settings — apply EthoVision smoothing/filters (median/low-pass) or apply an external filter (moving average, Savitzky–Golay) before import.
• Outliers — detect and remove spurious points that jump far from the previous position, then interpolate over them.

Suggested filtering workflow:

1. Identify outliers (e.g., >3× SD displacement or >threshold speed).
2. Replace outliers with NaN and interpolate linearly.
3. Apply a 3–5 frame median filter or a low-pass filter tuned to the expected animal speed.

---

Common issue: Missing frames or gaps in the track

Symptoms:

• Track has segments with no data or sudden blanks.
• EthoVision playback pauses or misaligns after gaps.

Causes and fixes:

• Frame numbering mismatch — ensure file uses continuous frame indices or proper timestamps. If file uses timestamp in seconds with fractional parts, convert timestamps to frame indices using the known frame rate: frame = round(timestamp × fps).
• Exported with missing frames — re-export ensuring all frames are included, or fill gaps by interpolation.
• Different frame rates — if your source tracking used a different fps, resample to the EthoVision video fps to align frames.

Filling gaps:

• Short gaps: interpolate linearly (or spline) between surrounding points.
• Long gaps: mark as missing and avoid interpolating across behavior-critical periods; consider re-tracking for those segments.

---

Common issue: Mismatched timestamps — track plays faster or slower than video

Symptoms:

• Paste Track playback runs ahead or behind the video.
• Behavioral events do not align with video frames.

Causes and fixes:

• FPS mismatch — confirm the video frame rate and the data timestamps/frame indices. Convert timestamps to match the video FPS. If your file uses frame numbers from another FPS, scale frame numbers by fps_target / fps_source.
• Timebase starting point — ensure both video and track start at the same time reference (e.g., both start at 0 s). If not, apply a time offset.
• Non-uniform sampling — if the track timestamps are irregular, resample to uniform intervals matching the video frame times.

Example conversion: If source fps = 25 and video fps = 30, then new_frame = round(old_frame * 30 / 25).

---

Common issue: EthoVision gives low detection quality or unexpected behavior when analyzing pasted tracks

Symptoms:

• Calculated measures (speed, distance) are wrong or noisy.
• Zone-enter/exit events occur at incorrect times.

Causes and fixes:

• Coordinate/scale mismatches affecting distance calculations — ensure units are correct (convert pixels to mm if needed).
• Sampling interval incorrect — speed is distance / time, so wrong timebase will produce wrong speeds. Resample timestamps to correct fps.
• Arena definition — zones must be defined in the same coordinate system as the pasted track. Re-check zone positions after coordinate transforms.
• Smoothing not applied — apply the same smoothing/filtering used for native EthoVision tracks.

Verify calculations by comparing simple metrics:

• Distance: sum of Euclidean distances between consecutive points.
• Speed: distance between frames divided by frame interval.

---

File-specific tips and formatting examples

Minimal CSV example (frame, x, y):

frame,x,y 0,102.34,88.12 1,103.11,88.40 2,103.80,89.05 

Timestamp-based example (time_s, x, y):

time_s,x,y 0.000,102.34,88.12 0.033,103.11,88.40 0.067,103.80,89.05 

• Include a single header row if EthoVision expects it; otherwise start with numeric rows.
• If including additional columns (orientation, confidence), ensure EthoVision’s Paste Track parser accepts them or remove extra columns.

---

Diagnostic checklist to run before importing

• Confirm file encoding (UTF-8/no BOM).
• Verify delimiter and decimal separator.
• Ensure timestamps or frame numbers match video fps and start time.
• Check coordinate origin, axis direction, rotation, and scale.
• Inspect for missing frames or NaNs; interpolate or mark appropriately.
• Preview a small sample (first 1000 rows) after transformation to confirm alignment.

---

When to re-track vs. fix the Paste Track file

Fix when:

• Problems are limited to scale, rotation, offsets, minor gaps, or sampling mismatches.
• The raw positional data is otherwise complete and accurate.

Re-track when:

• Large portions are missing or contain frequent outliers that cannot be reliably interpolated.
• Source tracking resolution or accuracy is too low for desired analyses.
• The video is available and re-tracking would be faster and more accurate than repairing exported data.

---

Useful scripts and tools

• Use Python (pandas, numpy) or MATLAB for batch transformations (scaling, rotation, interpolation).
• Example Python snippet for flipping Y, scaling, and resampling: “`python import pandas as pd import numpy as np

df = pd.read_csv(‘paste_track.csv’) frame_height = 480 # set to your video height in pixels scale = 0.264 # mm per pixel if applicable fps_src = 25 fps_target = 30

Flip Y

df[‘y’] = frame_height – df[‘y’]

Scale

df[[‘x’,‘y’]] *= scale

Resample timestamps to target fps

df[‘time_s’] = df[‘frame’] / fps_src new_frames = np.arange(0, df[‘time_s’].max(), 1.0/fps_target) interp_x = np.interp(new_frames, df[‘time_s’], df[‘x’]) interp_y = np.interp(new_frames, df[‘time_s’], df[‘y’]) out = pd.DataFrame({‘frame’: np.arange(len(new_frames)), ‘x’: interp_x, ‘y’: interp_y}) out.to_csv(‘paste_track_converted.csv’, index=False) “`

---

When to contact support

Contact Noldus/EthoVision support when:

• You suspect a bug in EthoVision’s Paste Track import (provide sample files and version info).
• The Paste Track parser fails on files that meet documented format specifications.
• You need assistance mapping advanced file formats or batch-processing large datasets.

Provide support with:

• EthoVision version, OS, and video/frame rate.
• A short sample of the paste track file (first 200 lines).
• A brief description of expected vs. observed behavior.

---

Summary

Most Paste Track problems come from mismatches in format, coordinate systems, sampling rate, or units. Systematically checking encoding, delimiters, timestamps, origin/axis orientation, scale, and applying simple transformations (translate/rotate/scale/resample) will fix the majority of issues. Use smoothing and outlier detection to address noise, and re-track only when data are fundamentally compromised.

---