Record one scene.
Keep a result you can revisit.

01 / SET UP A REPEATABLE RUN

Choose the scene before changing a setting.

1. Pick a repeatable segment. Use a built-in benchmark or a fixed replay. Choose a duration you can repeat; the example below records 60 seconds. Avoid menus, loading screens and switching applications during the recorded segment.
2. Write down the setup. Keep the game version, Windows build, driver, resolution, graphics quality, frame cap and capture tool in your notebook. Keep frame-generation and upscaling settings unchanged.
3. Plan five runs per condition. Warm up the game first. Use the same warm-up, start point and duration for five baseline runs and five runs after one change. Keep every raw capture and recheck the baseline later for drift.

You do not need to buy or install Radiant Optimizer to use the notebook. The workflow also works for a game setting, driver change or another tool, provided you can record the exact change and how to reverse it.

02 / CREATE A COMPATIBLE CSV

Use PresentMon's console export.

The example uses the standalone PresentMon 2.3.1 console release and its documented options. This is a specific documented version, not a claim that it is the newest. Select the console executable named PresentMon-2.3.1-x64.exe from the official release assets for an x64 Windows PC.

1. Put the console executable in a writable folder. Open PowerShell in that folder. Keep its exact filename; this example is for the console executable, not the graphical capture app.
2. Start your game and identify its process. In Task Manager's Details view, find the running game's executable and PID. Use that number below instead of 1234. Do not select the launcher. A restart may change the PID.
3. Prepare the scene, then run this command. Return to the game. At the same starting point each time, press Alt + Shift + F11 to begin the 60-second capture.

& ".\PresentMon-2.3.1-x64.exe" --process_id 1234 --v1_metrics --hotkey "ALT+SHIFT+F11" --timed 60 --terminate_after_timed --output_file ".\before-01.csv"

Repeat with output names before-02.csv through before-05.csv. After one documented change, repeat with after-01.csv through after-05.csv. If you choose a different duration, use it for every run.

Keep the command's export settings unchanged. Do not add --qpc_time, --date_time or --exclude_dropped: this importer needs the legacy time column and the complete present stream. Other CSV formats require a different analysis path.

03 / SAVE THE BASELINE, THEN THE COMPARISON

Bring the files back to your notebook.

1. Import the baseline. Open Your tests, complete the setup fields and choose Add before CSV captures. Select your five separate baseline files, then Save baseline & continue.
2. Record one change and its recovery path. Capture the after runs with the same process, scene, duration and settings apart from that change. Use Add after CSV captures in the second step.
3. Compare and save. Confirm the matching conditions, choose Compare runs, then save the comparison with a keep, revert, retest or undecided conclusion. Export a notebook backup and retain the raw CSVs separately.

Each CSV supplies one mean present rate. FrameLedger compares repeated captures and shows their spread, plus present-interval details. Those intervals describe application Present calls; they do not establish displayed FPS, input latency, a stutter diagnosis or a causal benefit from the change. A small difference amid wide variation is a reason to retest.

The notebook accepts 3–20 captures per condition; five or more is preferred. Each file may be up to 8 MB, with a 24 MB limit per selection. Use smaller selections to add a large set. A file needs at least 30 positive intervals and no more than 200,000 data rows.

Files stay on your device. Saved records contain filenames and calculated summaries; the original CSVs are not uploaded or included in the notebook backup. Local records do not sync between browsers or devices.

ALTERNATIVE / NO COMPATIBLE CSV

Enter repeated run averages.

If your game's benchmark or existing measurement tool reports average FPS, record that value after each complete run. Enter five separate baseline averages, then five after-change averages in the notebook, separated by commas or new lines. Use the same tool, metric, scene and duration for both conditions.

Keep screenshots or original reports with your raw evidence. Do not mix displayed FPS with present rate, treat several moments from one run as separate runs, or enter illustrative example numbers as your own measurement. Manual comparisons summarize only the averages you supply; they cannot reconstruct frame-time details.