RawScope

RAW 9 quality evidence

Quality claims need context. This page records what we tested, what we measured and what the result does not prove.

Fine detail

The retained test

The source was DSC09859.ARW from a Sony A7R V with the Sony FE 70-200mm F2.8 GM II. The exact focal length was not recovered in the retained test record.

The final quantitative pilot compared RawScope RAW 8, RawScope RAW 9 and darktable 5.6.1. Each output was exported at full processed resolution as a 16-bit Adobe RGB TIFF, without export resizing or export sharpening.

Affinity 3.3.0 was included in earlier comparison work but was not part of this final quantitative pilot, so no quantitative claim here is made against Affinity.

What we measured

Thirteen physical ColorChecker edges were accepted across all three outputs.

MTF50 results
OutputMedian target-normalised MTF50
(cycles per patch spacing)
RAW 927.703
RAW 825.985
darktable 5.6.124.824

RAW 9 measured higher than both other outputs on all 13 accepted edges.

The RAW 8 and RAW 9 outputs registered directly on the same pixel grid. The lens-corrected darktable output differed in geometry, so it was not silently warped or resampled to force a same-pixel comparison.

What the result means

In this test, using this RAW file and these recorded processing conditions, RawScope's RAW 9 output showed higher measured fine-detail response than the RAW 8 and darktable outputs.

That is the claim this evidence supports.

What it does not mean

This was one RAW file and an exploratory target-based comparison, not a universal RAW-converter benchmark or an ISO sharpness laboratory test.

The darktable rendering was materially darker than the RawScope renderings and its lens-corrected output had different geometry.

This evidence does not show that RAW 9 is always sharper, that it always produces less noise, or that RawScope is better than every other RAW converter.

Why RawScope exposes RAW 9

Where Apple reports RAW 9 as available for the loaded file, RawScope lets you use it directly, inspect the rendering at true 1:1 and export the prepared result for finishing in another editor.

RawScope is designed to make that RAW stage visible and deliberate, not to replace the editor you use afterwards.

Methodology and provenance

The retained record identifies the source RAW, compared applications, export conditions and accepted edge measurements supporting the result above.

The publication provenance for a comparison crop is not yet complete. In particular, the final public comparison requires an exact source RAW hash, the exact RawScope build/source associated with the compared exports, crop coordinates, registration record and final crop hashes.

For that reason, the measured result and its limits may be published without a comparison image. A comparison visual will be added only after the provenance record is complete and the final crop is approved.