EPD2HexCompress.exe 2.00 - Deprecated

Deprecated / superseded: EPD2HexCompress.exe has been superseded by PositronPixelFoundry.

This page is retained for users of the earlier application and for historical reference. For new image-conversion work, use Positron Pixel Foundry and its detailed User Manual.

The retained legacy material below reflects the final 2.x documentation page and may identify its last maintenance build as 2.01.

PositronPixelFoundry overview  |  Compression Explained

BW/BWR/BWY Picture to HEX & Compression Tool for EPD EPD2HexCompress

(Current version 2.01)

This tool converts any picture media supported by FFmpeg into a HEX data table for Positron.
It is specifically designed for pictures destined for Black/White, Black/White/Red, or Black/White/Yellow EPD displays.

This is not a general picture-to-black-white-red/yellow converter — those are readily available online with many options.
The purpose of EPD2HexCompress.exe is to take control of exactly how a picture is reduced to your panel's actual color set, then pack the result into a compressed HEX flash table, with full visibility into the size and quality trade-offs along the way.

Version 2.00 is a complete redesign from 1.00: a live preview window, a full image-adjustment pipeline, thirteen dithering algorithms, adjustable classification thresholds, three EPD panel types instead of one, per-method compression comparison, and settings profiles.
Everything below reflects the current tool.

Panel types - Select the target panel from the dropdown:

• BW — Black/White only.
• BWR — Black/White/Red.
• BWY — Black/White/Yellow.

The panel type determines whether a color plane is generated at all, and which color the Red/Yellow classification targets.

Classification - adjustable, not fixed:

• BW Threshold (0–100%) — the luminance cutoff between black and white. Pixels at or above this brightness become white; below it, black.
• R/Y Threshold (0–100%) — how strictly a pixel must resemble red or yellow before it's assigned to the color plane, rather than falling through to the black/white decision.

Both sliders update the live preview immediately, so you can watch the classification change in real time rather than guessing and re-exporting.
Every slider in the tool also responds to the mouse wheel — hover over one and scroll to nudge its value by 1, without needing to click it first.

Transparency handling

Transparent pixels are no longer discarded or left to chance. Pick a background color (via the color swatch) and every transparent pixel is flattened to that exact color before classification — cleanly, with a hard edge (no half-transparent fringing), regardless of how soft or anti-aliased the source image's alpha channel is.

Image adjustments

• Brightness (-50 to +50) — darken or lighten. 0 = no change.
• Contrast (-50 to +50) — flatten or increase contrast. 0 = no change.
• Gamma (-50 to +50) — reshape midtones without touching pure black or pure white. 0 = no change.
• Saturation (-50 to +50) — desaturate toward grayscale, or boost color intensity. 0 = no change.
• Sharpening (-50 to +50) — sharpen edges at positive values, blur at negative values. 0 = no change.
• Auto Levels (checkbox) — automatically stretches each color channel to use its full range before the other adjustments are applied.
• Edge Preservation (checkbox) — a bilateral smoothing filter that reduces noise while keeping edges crisp, applied before classification. Most useful on real photographs; has no benefit on already-flat source art.
• Invert (checkbox) — swaps black and white after classification.

Before classification, the source image can be run through a full adjustment chain — genuinely useful once you're feeding it photographs or gradient artwork rather than only flat icons.
Every adjustment updates the live preview instantly.

Dithering

• N/A — no dithering (the original 1.00 behavior — best for icons and flat line art).
• Bayer, Blue Noise, Random, Halftone — ordered/positional methods. Fast, and threshold changes stay quick to re-preview.
• Floyd Steinberg, Atkinson, Jarvis Judice Ninke, Burkes, Stucki, Sierra, Two-Row Sierra, Sierra Lite — error-diffusion methods, generally the best gradient/photographic quality. Each threshold change triggers a full reclassification pass rather than a fast partial update, since error diffusion depends on the whole image.
• Riemersma — error diffusion traced along a Hilbert space-filling curve rather than raster order, avoiding some of the directional artifacts the raster-order methods can show.
• Dithering Strength slider (0–100%) - control how strongly it's applied.

Where 1.00 had no dithering at all and expected near-flat-color source art, 2.00 adds a full dithering engine for anything with genuine gradients or photographic tone — thirteen algorithms plus an off setting, selectable from the Dithering dropdown.

Live preview window

Clicking PREVIEW converts and classifies the current image with all current settings and opens a live preview window:

• 1x–4x zoom buttons, for checking individual pixels — the currently active zoom level is highlighted, and the window title shows the zoom factor alongside the source filename and processed dimensions.
• Adjusting any slider, checkbox, or dropdown while the preview is open updates it immediately — no need to click Preview again after each tweak.
• The preview reflects exactly the RGB565, 2-or-3-color result your panel will actually receive.

Output transformations:

• -90 / +90 — rotate the preview 90° counter-clockwise or clockwise. Width and height swap accordingly, and the width/height boxes update to match.
• 180 — rotate the preview 180°. Width and height are unchanged.
• H ↔ — flip the preview horizontally (left-to-right mirror). Width and height are unchanged.
• V ↕ — flip the preview vertically (top-to-bottom mirror). Width and height are unchanged.
• X/Y ⤢ — diagonal mirror (transpose): reflects the image across its corner-to-corner diagonal. Like the 90° rotations, width and height swap.

Once a preview is open, the six buttons let you reorient the classified result directly — geometry only, applied instantly to the already-processed image with no reconversion from the source picture.
Each of these carries through to the HEX table and to Save BMP — whatever orientation the preview shows is exactly what gets exported.
They are not saved in settings profiles, since orientation is treated as a per-picture choice rather than part of a reusable adjustment "look."

Live statistics

Below the preview controls, three figures update continuously as you adjust settings:

• Black / White / Red-or-Yellow pixel counts, each shown with its percentage of the total image (the three percentages always sum to 100%).
• Compression result for the currently selected method: total pixels vs. resulting HEX byte count, expressed as a percentage.

What this percentage means: it's the combined byte count from both planes divided by the total pixel count (width × height of the resized image) — in effect, "what fraction of a standard, saved BMP file's size does the compressed result use." 

The 1-byte-per-pixel baseline isn't arbitrary: a standard BMP saving three discrete colors (black, white, and red or yellow) would naturally use 8-bit indexed color — one byte per pixel, giving access to a 256-color palette even though only 3 of those slots are actually used — since that's the smallest standard BMP pixel format capable of holding more than two colors.

So this percentage is a genuine, real-world comparison: how much smaller is the compressed HEX table than simply saving the same image as an ordinary 8-bit, 256-color BMP.

• HEX byte count for all three compression methods (Auto / Bitmap / RLE) shown side by side, regardless of which one is currently selected — so you can compare before committing to one.

These figures reflect whatever is currently in memory and remain visible even after closing the preview window; they're cleared only when a new picture is loaded.

Compression methods

Data in the table is compressed using an adaptive per-plane scheme. The Black/White plane and the Red/Yellow plane are compressed independently, so each can use whichever mode suits its own content:

• RLE — stores each run of same-colored pixels as a (color, length) pair; most efficient for large uniform regions.
• Bitmap — stores one bit per pixel with no run compression; most efficient for noisy or fine-detail content where runs are too short for RLE to help.
• Constant — a single flag byte for a plane that's entirely one color (all-black, all-white, or an empty color plane); the smallest possible result when it applies, and always used automatically whenever a plane turns out to be uniform, regardless of which method is selected below.

Choose how each plane's mode is picked, via the radio buttons:

• Auto (default) — each plane is encoded both ways (RLE and Bitmap), the actual encoded sizes are compared, and whichever is smaller is kept — independently per plane. This is almost always the smallest result, since it lets the Black/White plane and the color plane each pick their own best encoding.
• Bitmap — any plane with mixed content is encoded as bitmap.
• RLE — any plane with mixed content is encoded as RLE.

The three HEX-byte-count figures in the live statistics area let you see exactly what each choice would cost before picking one.

Settings profiles

Once you've dialed in an adjustment chain you like, SAVE PROFILE writes every slider, the dithering type, and the three checkboxes above to a plain text .ini file — not the panel type, and not the compression method, since those are treated as per-session/per-panel choices rather than part of an image-processing "look."

LOAD PROFILE restores a saved profile in one step. If a picture is already loaded, the preview updates once, after every value has been restored — not once per slider — so loading a profile doesn't flash through a dozen intermediate states.

Saving the classified result

SAVE BMP writes the currently previewed image to disk as a native RGB565 BMP — the exact classified colors, at full native resolution regardless of what zoom level the preview window happens to be showing.
Useful as a reference file, or for reloading later to pick up exactly where you left off.

Converting a picture to HEX

1. Click ADD PICTURE and select a source image. This also resets every adjustment slider and checkbox to its default and closes any previously open preview, so you're starting clean for the new picture.
2. Select the panel type (BW / BWR / BWY).
3. Set the desired output width and height, and enable the crop checkbox if the source image's aspect ratio doesn't already match your target.
4. Click PREVIEW and adjust the BW/R-Y thresholds, image adjustments, and dithering to taste, watching the live preview and pixel-count statistics update as you go.
5. Rotate or flip the preview if needed, using the six transformation buttons.
6. Choose a compression method (or leave on Auto), using the per-method byte counts to compare.
7. Assign an alias for the HEX data table.
8. Click CREATE HEX. If you've already previewed the image, this uses that result directly; if not, it runs the same conversion and classification automatically first.

The resulting HEX table is written to a text file in the TEMP subfolder (which opens automatically after conversion) and is also displayed directly in the tool, ready to COPY to the clipboard.
Paste this data into your Positron program file to be used by the library commands.

Created with the Personal Edition of HelpNDoc: Maximize Your Reach: Convert Your Word Document to an ePub or Kindle eBook