diff --git a/GPU/Common/DepalettizeCommon.cpp b/GPU/Common/DepalettizeCommon.cpp
index 70214fc3d5..185eecdbc4 100644
--- a/GPU/Common/DepalettizeCommon.cpp
+++ b/GPU/Common/DepalettizeCommon.cpp
@@ -147,10 +147,10 @@ Draw::SamplerState *DepalShaderCache::GetSampler() {
 	return nearestSampler_;
 }
 
-DepalShader *DepalShaderCache::GetDepalettizeShader(uint32_t clutMode, GEBufferFormat pixelFormat) {
+DepalShader *DepalShaderCache::GetDepalettizeShader(uint32_t clutMode, GETextureFormat textureFormat, GEBufferFormat bufferFormat) {
 	using namespace Draw;
 
-	u32 id = GenerateShaderID(clutMode, pixelFormat);
+	u32 id = GenerateShaderID(clutMode, textureFormat, bufferFormat);
 
 	auto shader = cache_.find(id);
 	if (shader != cache_.end()) {
@@ -171,7 +171,8 @@ DepalShader *DepalShaderCache::GetDepalettizeShader(uint32_t clutMode, GEBufferF
 	config.startPos = gstate.getClutIndexStartPos();
 	config.shift = gstate.getClutIndexShift();
 	config.mask = gstate.getClutIndexMask();
-	config.pixelFormat = pixelFormat;
+	config.bufferFormat = bufferFormat;
+	config.textureFormat = textureFormat;
 
 	GenerateDepalFs(buffer, config, draw_->GetShaderLanguageDesc());
 	
diff --git a/GPU/Common/DepalettizeCommon.h b/GPU/Common/DepalettizeCommon.h
index c394f47eb7..fbac86300d 100644
--- a/GPU/Common/DepalettizeCommon.h
+++ b/GPU/Common/DepalettizeCommon.h
@@ -49,7 +49,7 @@ public:
 	~DepalShaderCache();
 
 	// This also uploads the palette and binds the correct texture.
-	DepalShader *GetDepalettizeShader(uint32_t clutMode, GEBufferFormat pixelFormat);
+	DepalShader *GetDepalettizeShader(uint32_t clutMode, GETextureFormat texFormat, GEBufferFormat pixelFormat);
 	Draw::Texture *GetClutTexture(GEPaletteFormat clutFormat, const u32 clutHash, u32 *rawClut);
 
 	Draw::SamplerState *GetSampler();
@@ -63,8 +63,8 @@ public:
 	void DeviceRestore(Draw::DrawContext *draw);
 
 private:
-	static uint32_t GenerateShaderID(uint32_t clutMode, GEBufferFormat pixelFormat) {
-		return (clutMode & 0xFFFFFF) | (pixelFormat << 24);
+	static uint32_t GenerateShaderID(uint32_t clutMode, GETextureFormat texFormat, GEBufferFormat pixelFormat) {
+		return (clutMode & 0xFFFFFF) | (pixelFormat << 24) | (texFormat << 28);
 	}
 
 	static uint32_t GetClutID(GEPaletteFormat clutFormat, uint32_t clutHash) {
diff --git a/GPU/Common/DepalettizeShaderCommon.cpp b/GPU/Common/DepalettizeShaderCommon.cpp
index 9874e91d6e..11c3c50394 100644
--- a/GPU/Common/DepalettizeShaderCommon.cpp
+++ b/GPU/Common/DepalettizeShaderCommon.cpp
@@ -49,7 +49,7 @@ void GenerateDepalShader300(ShaderWriter &writer, const DepalConfig &config, con
 	const int shift = config.shift;
 	const int mask = config.mask;
 
-	if (config.pixelFormat == GE_FORMAT_DEPTH16) {
+	if (config.bufferFormat == GE_FORMAT_DEPTH16) {
 		DepthScaleFactors factors = GetDepthScaleFactors();
 		writer.ConstFloat("z_scale", factors.scale);
 		writer.ConstFloat("z_offset", factors.offset);
@@ -71,7 +71,7 @@ void GenerateDepalShader300(ShaderWriter &writer, const DepalConfig &config, con
 	writer.C("  vec4 color = ").SampleTexture2D("tex", "v_texcoord").C(";\n");
 
 	int shiftedMask = mask << shift;
-	switch (config.pixelFormat) {
+	switch (config.bufferFormat) {
 	case GE_FORMAT_8888:
 		if (shiftedMask & 0xFF) writer.C("  int r = int(color.r * 255.99);\n"); else writer.C("  int r = 0;\n");
 		if (shiftedMask & 0xFF00) writer.C("  int g = int(color.g * 255.99);\n"); else writer.C("  int g = 0;\n");
@@ -102,6 +102,17 @@ void GenerateDepalShader300(ShaderWriter &writer, const DepalConfig &config, con
 	case GE_FORMAT_DEPTH16:
 		// Remap depth buffer.
 		writer.C("  float depth = (color.x - z_offset) * z_scale;\n");
+
+		if (config.bufferFormat == GE_FORMAT_DEPTH16 && config.textureFormat == GE_TFMT_5650) {
+			// Convert depth to 565, without going through a CLUT.
+			writer.C("  int idepth = int(clamp(depth, 0.0, 65535.0));\n");
+			writer.C("  float r = (idepth & 31) / 31.0f;\n");
+			writer.C("  float g = ((idepth >> 5) & 63) / 63.0f;\n");
+			writer.C("  float b = ((idepth >> 11) & 31) / 31.0f;\n");
+			writer.C("  vec4 outColor = vec4(r, g, b, 1.0);\n");
+			return;
+		}
+
 		writer.C("  int index = int(clamp(depth, 0.0, 65535.0));\n");
 		break;
 	default:
@@ -135,7 +146,7 @@ void GenerateDepalShaderFloat(ShaderWriter &writer, const DepalConfig &config, c
 	const int shift = config.shift;
 	const int mask = config.mask;
 
-	if (config.pixelFormat == GE_FORMAT_DEPTH16) {
+	if (config.bufferFormat == GE_FORMAT_DEPTH16) {
 		DepthScaleFactors factors = GetDepthScaleFactors();
 		writer.ConstFloat("z_scale", factors.scale);
 		writer.ConstFloat("z_offset", factors.offset);
@@ -144,7 +155,7 @@ void GenerateDepalShaderFloat(ShaderWriter &writer, const DepalConfig &config, c
 	float index_multiplier = 1.0f;
 	// pixelformat is the format of the texture we are sampling.
 	bool formatOK = true;
-	switch (config.pixelFormat) {
+	switch (config.bufferFormat) {
 	case GE_FORMAT_8888:
 		if ((mask & (mask + 1)) == 0) {
 			// If the value has all bits contiguous (bitmask check above), we can mod by it + 1.
@@ -249,7 +260,7 @@ void GenerateDepalShaderFloat(ShaderWriter &writer, const DepalConfig &config, c
 	// index_multiplier -= 0.01f / texturePixels;
 
 	if (!formatOK) {
-		ERROR_LOG_REPORT_ONCE(depal, G3D, "%s depal unsupported: shift=%d mask=%02x offset=%d", GeBufferFormatToString(config.pixelFormat), shift, mask, config.startPos);
+		ERROR_LOG_REPORT_ONCE(depal, G3D, "%s depal unsupported: shift=%d mask=%02x offset=%d", GeBufferFormatToString(config.bufferFormat), shift, mask, config.startPos);
 	}
 
 	// Offset by half a texel (plus clutBase) to turn NEAREST filtering into FLOOR.
diff --git a/GPU/Common/DepalettizeShaderCommon.h b/GPU/Common/DepalettizeShaderCommon.h
index 5ce5ef88a8..74bb38d196 100644
--- a/GPU/Common/DepalettizeShaderCommon.h
+++ b/GPU/Common/DepalettizeShaderCommon.h
@@ -29,7 +29,8 @@ struct DepalConfig {
 	int shift;
 	u32 startPos;
 	GEPaletteFormat clutFormat;
-	GEBufferFormat pixelFormat;
+	GETextureFormat textureFormat;
+	GEBufferFormat bufferFormat;
 };
 
 void GenerateDepalFs(char *buffer, const DepalConfig &config, const ShaderLanguageDesc &lang);
diff --git a/GPU/Common/TextureCacheCommon.cpp b/GPU/Common/TextureCacheCommon.cpp
index af642a7df1..64a48537b6 100644
--- a/GPU/Common/TextureCacheCommon.cpp
+++ b/GPU/Common/TextureCacheCommon.cpp
@@ -525,6 +525,10 @@ TexCacheEntry *TextureCacheCommon::SetTexture() {
 	def.format = format;
 	def.bufw = bufw;
 
+	if (texaddr == 0x04710000) {
+		texaddr = texaddr;
+	}
+
 	std::vector<AttachCandidate> candidates = GetFramebufferCandidates(def, 0);
 	if (candidates.size() > 0) {
 		int index = GetBestCandidateIndex(candidates);
@@ -892,6 +896,7 @@ bool TextureCacheCommon::MatchFramebuffer(
 
 	const bool noOffset = texaddr == addr;
 	const bool exactMatch = noOffset && entry.format < 4 && channel == RASTER_COLOR;
+
 	const u32 w = 1 << ((entry.dim >> 0) & 0xf);
 	const u32 h = 1 << ((entry.dim >> 8) & 0xf);
 	// 512 on a 272 framebuffer is sane, so let's be lenient.
@@ -927,6 +932,7 @@ bool TextureCacheCommon::MatchFramebuffer(
 		// Check works for D16 too (???)
 		const bool matchingClutFormat =
 			(fb_format == GE_FORMAT_DEPTH16 && entry.format == GE_TFMT_CLUT16) ||
+			(fb_format == GE_FORMAT_DEPTH16 && entry.format == GE_TFMT_5650) ||
 			(fb_format == GE_FORMAT_8888 && entry.format == GE_TFMT_CLUT32) ||
 			(fb_format != GE_FORMAT_8888 && entry.format == GE_TFMT_CLUT16);
 
@@ -971,7 +977,9 @@ bool TextureCacheCommon::MatchFramebuffer(
 		// 3rd Birthday (and a bunch of other games) render to a 16 bit clut texture.
 		if (matchingClutFormat) {
 			if (!noOffset) {
-				WARN_LOG_ONCE(subareaClut, G3D, "Texturing from framebuffer using CLUT with offset at %08x +%dx%d", fb_address, matchInfo->xOffset, matchInfo->yOffset);
+				WARN_LOG_ONCE(subareaClut, G3D, "Texturing from framebuffer (%s) using %s with offset at %08x +%dx%d", channel == RASTER_DEPTH ? "DEPTH" : "COLOR", GeTextureFormatToString(entry.format), fb_address, matchInfo->xOffset, matchInfo->yOffset);
+			} else {
+				WARN_LOG_ONCE(subareaClut, G3D, "Texturing from framebuffer (%s) using %s at %08x", channel == RASTER_DEPTH ? "DEPTH" : "COLOR", GeTextureFormatToString(entry.format), fb_address);
 			}
 			return true;
 		} else if (IsClutFormat((GETextureFormat)(entry.format)) || IsDXTFormat((GETextureFormat)(entry.format))) {
@@ -1823,12 +1831,33 @@ void TextureCacheCommon::ApplyTexture() {
 	gstate_c.SetTextureIs3D((entry->status & TexCacheEntry::STATUS_3D) != 0);
 }
 
+bool CanDepalettize(GETextureFormat texFormat, GEBufferFormat bufferFormat) {
+	if (IsClutFormat(texFormat)) {
+		switch (bufferFormat) {
+		case GE_FORMAT_4444:
+		case GE_FORMAT_565:
+		case GE_FORMAT_5551:
+		case GE_FORMAT_DEPTH16:
+			return texFormat == GE_TFMT_CLUT16;
+		case GE_FORMAT_8888:
+			return texFormat == GE_TFMT_CLUT32;
+		}
+		WARN_LOG(G3D, "Invalid CLUT/framebuffer combination: %s vs %s", GeTextureFormatToString(texFormat), GeBufferFormatToString(bufferFormat));
+		return false;
+	} else if (texFormat == GE_TFMT_5650 && bufferFormat == GE_FORMAT_DEPTH16) {
+		// We can also "depal" 565 format, this is used to read depth buffers as 565 on occasion (#15491).
+		return true;
+	} else {
+		return false;
+	}
+}
+
 void TextureCacheCommon::ApplyTextureFramebuffer(VirtualFramebuffer *framebuffer, GETextureFormat texFormat, RasterChannel channel) {
 	DepalShader *depalShader = nullptr;
 	uint32_t clutMode = gstate.clutformat & 0xFFFFFF;
 
-	bool need_depalettize = IsClutFormat(texFormat);
 	bool depth = channel == RASTER_DEPTH;
+	bool need_depalettize = CanDepalettize(texFormat, depth ? GE_FORMAT_DEPTH16 : framebuffer->drawnFormat);
 	bool useShaderDepal = framebufferManager_->GetCurrentRenderVFB() != framebuffer && !depth && !gstate_c.curTextureIs3D;
 
 	// TODO: Implement shader depal in the fragment shader generator for D3D11 at least.
@@ -1878,7 +1907,7 @@ void TextureCacheCommon::ApplyTextureFramebuffer(VirtualFramebuffer *framebuffer
 			return;
 		}
 
-		depalShader = depalShaderCache_->GetDepalettizeShader(clutMode, depth ? GE_FORMAT_DEPTH16 : framebuffer->drawnFormat);
+		depalShader = depalShaderCache_->GetDepalettizeShader(clutMode, texFormat, depth ? GE_FORMAT_DEPTH16 : framebuffer->drawnFormat);
 		gstate_c.SetUseShaderDepal(false);
 	}
 
diff --git a/unittest/TestShaderGenerators.cpp b/unittest/TestShaderGenerators.cpp
index 19b0b0d2e9..a96df6657c 100644
--- a/unittest/TestShaderGenerators.cpp
+++ b/unittest/TestShaderGenerators.cpp
@@ -309,7 +309,8 @@ bool TestDepalShaders() {
 		config.shift = 8;
 		config.startPos = 64;
 		config.mask = 0xFF;
-		config.pixelFormat = GE_FORMAT_8888;
+		config.bufferFormat = GE_FORMAT_8888;
+		config.textureFormat = GE_TFMT_CLUT32;
 
 		GenerateDepalFs(buffer, config, desc);
 		if (!TestCompileShader(buffer, languages[k], ShaderStage::Fragment, &errorMessage)) {