ELA Explained: How Compression Reveals Photoshop Traces

When people talk about image authenticity, ELA (Error Level Analysis) is one of the most common forensic techniques. But why does it work? And why do edited areas often "light up" in an ELA view? This article explains the core idea in plain language.

The Core Idea: Compression Leaves Fingerprints

JPEG Compression in Simple Terms

JPEG is a lossy format. Each time you save a JPEG, the image is re‑compressed and some subtle detail is lost.

Think of it like painting:

• First save: the whole image is painted using one brush size (compression quality)
• Later edit: you replace or modify a portion, then save again using the current brush

If the edited area has a different compression history, its "brush strokes" won’t match the rest of the image.

How ELA Finds Those Differences

ELA works by re‑compressing the image at a fixed quality (e.g. 95%), then comparing pixels between the original and the re‑compressed version.

• Original, untouched areas tend to show low, consistent error
• Edited areas show noticeably different error levels
• The tool then amplifies those tiny errors so we can see them clearly

In short: an unedited image behaves like a single cloth, while edits look like patches under a microscope.

Why Edits Leave Traces

Scenario 1: Compositing / Splicing

If you paste a person into a background:

• The background might already be heavily compressed
• The pasted element could be from a different source or device

After saving, both are re‑compressed together, but their histories differ, so the inserted object stands out in ELA.

Scenario 2: Local Retouching

Clone stamps, blurs, or object removal create new pixels. Even if they look natural, their compression signature usually differs.

ELA often reveals these regions as unusually smooth or overly bright blocks.

Limitations to Know

ELA is powerful, but not magic.

Cases It Struggles With

1. Full re‑capture (screenshots or re‑photos)
   • Everything is uniformly re‑compressed, so ELA can’t isolate edits
2. AI‑generated images
   • There is no "original vs edited" area; the compression is globally consistent
3. Expert‑level forgery
   • Skilled editors can add noise or re‑compress to hide traces

Common False Positives

• Images that went through social media compression
• Heavy filters or global color grading
• Very low‑quality or over‑compressed files

Practical Tips

1. Use the original file when possible
2. Compare against a known‑clean image to learn the normal pattern
3. Adjust parameters (compression quality / error scale)
4. Combine with other signals (metadata, source checks, reverse image search)

---

ELA gives you a useful lens into image tampering. It won’t prove authenticity on its own, but it’s a fast, reliable first step when used carefully.