diff --git a/src/Riff.c b/src/Riff.c
index 5181ba1..e86aa2c 100644
--- a/src/Riff.c
+++ b/src/Riff.c
@@ -137,13 +137,13 @@ char const* DmRiff_readString(DmRiff* slf) {
 	return str;
 }
 
-char16_t* DmRiff_readStringUtf(DmRiff* slf) {
+uint16_t* DmRiff_readStringUtf(DmRiff* slf) {
 	if (slf == NULL) {
-		return u"";
+		return NULL;
 	}
 
-	char16_t* mem = (char16_t*) slf->mem;
-	char16_t* str = (char16_t*) slf->mem + slf->pos;
+	uint16_t* mem = (uint16_t*) slf->mem;
+	uint16_t* str = (uint16_t*) slf->mem + slf->pos;
 
 	while (mem[slf->pos] != 0) {
 		slf->pos += 1;
diff --git a/src/_Internal.h b/src/_Internal.h
index ca39d09..717781f 100644
--- a/src/_Internal.h
+++ b/src/_Internal.h
@@ -12,7 +12,6 @@
 
 #include <stdatomic.h>
 #include <stdbool.h>
-#include <uchar.h>
 
 typedef enum DmResolveFlags {
 	DmResolve_AFTER_PREPARE_TIME = 1 << 10,
diff --git a/src/_Riff.h b/src/_Riff.h
index 10287d9..5d67092 100644
--- a/src/_Riff.h
+++ b/src/_Riff.h
@@ -2,7 +2,6 @@
 // SPDX-License-Identifier: MIT-Modern-Variant
 #pragma once
 #include <stdbool.h>
-#include <uchar.h>
 
 #define DM_FOURCC(a, b, c, d) (((d) << 24U) | ((c) << 16U) | ((b) << 8U) | (a))
 #define DM_FOURCC_RIFF DM_FOURCC('R', 'I', 'F', 'F')
@@ -118,7 +117,7 @@ DMINT void DmInfo_parse(DmInfo* slf, DmRiff* rif);
 DMINT void DmVersion_parse(DmVersion* slf, DmRiff* rif);
 DMINT void DmReference_parse(DmReference* slf, DmRiff* rif);
 
-DMINT char* Dm_utf16ToUtf8Inline(char* out, char16_t const* u16);
+DMINT char* Dm_utf16ToUtf8Inline(char* out, uint16_t const* u16);
 
 DMINT bool DmRiff_init(DmRiff* slf, void const* buf, size_t len);
 DMINT bool DmRiff_is(DmRiff const* slf, uint32_t id, uint32_t typ);
@@ -132,5 +131,5 @@ DMINT void DmRiff_readInt(DmRiff* slf, int32_t* buf);
 DMINT void DmRiff_readDword(DmRiff* slf, uint32_t* buf);
 DMINT void DmRiff_readDouble(DmRiff* slf, double* buf);
 DMINT char const* DmRiff_readString(DmRiff* slf);
-DMINT char16_t* DmRiff_readStringUtf(DmRiff* slf);
+DMINT uint16_t* DmRiff_readStringUtf(DmRiff* slf);
 DMINT void DmRiff_reportDone(DmRiff* slf);
diff --git a/src/io/Common.c b/src/io/Common.c
index 77ad34a..13ec3f3 100644
--- a/src/io/Common.c
+++ b/src/io/Common.c
@@ -2,14 +2,13 @@
 // SPDX-License-Identifier: MIT-Modern-Variant
 #include "_Internal.h"
 
-#include <uchar.h>
 #include <wchar.h>
 
 void DmGuid_parse(DmGuid* slf, DmRiff* rif) {
 	DmRiff_read(rif, slf->data, sizeof slf->data);
 }
 
-char* Dm_utf16ToUtf8Inline(char* out, char16_t const* u16) {
+char* Dm_utf16ToUtf8Inline(char* out, uint16_t const* u16) {
 	size_t len = 0;
 	while (u16[len] != 0) {
 		len += 1;
@@ -21,13 +20,7 @@ char* Dm_utf16ToUtf8Inline(char* out, char16_t const* u16) {
 	size_t i = 0;
 	size_t j = 0;
 	for (; i < len; ++i) {
-#ifndef _WIN32
-		j += c16rtomb(out + j, u16[i], &state);
-#else
-		// NOTE: MinGW on Windows does not correctly implement `c16romb`, thus we just use `wcrtomb`.
-		// According to MinGW, `char16_t` should be equivalent to `wchar_t` on Windows.
-		j += wcrtomb(out + j, (wchar_t) u16[i], &state);
-#endif
+		j += wcrtomb(out + j, u16[i], &state);
 	}
 
 	out[j] = '\0';
@@ -38,7 +31,7 @@ void DmUnfo_parse(DmUnfo* slf, DmRiff* rif) {
 	DmRiff cnk;
 	while (DmRiff_readChunk(rif, &cnk)) {
 		if (DmRiff_is(&cnk, DM_FOURCC_UNAM, 0)) {
-			char16_t* raw = DmRiff_readStringUtf(&cnk);
+			uint16_t* raw = DmRiff_readStringUtf(&cnk);
 			slf->unam = Dm_utf16ToUtf8Inline((char*) raw, raw);
 		} else {
 			DmRiff_reportDone(&cnk);
@@ -83,11 +76,11 @@ void DmReference_parse(DmReference* slf, DmRiff* rif) {
 		} else if (DmRiff_is(&cnk, DM_FOURCC_GUID, 0)) {
 			DmGuid_parse(&slf->guid, &cnk);
 		} else if (DmRiff_is(&cnk, DM_FOURCC_NAME, 0)) {
-			char16_t* raw = DmRiff_readStringUtf(&cnk);
+			uint16_t* raw = DmRiff_readStringUtf(&cnk);
 			slf->name = Dm_utf16ToUtf8Inline((char*) raw, raw);
 			continue; // Ignore following bytes
 		} else if (DmRiff_is(&cnk, DM_FOURCC_FILE, 0)) {
-			char16_t* raw = DmRiff_readStringUtf(&cnk);
+			uint16_t* raw = DmRiff_readStringUtf(&cnk);
 			slf->file = Dm_utf16ToUtf8Inline((char*) raw, raw);
 			continue; // Ignore following bytes
 		} else if (DmRiff_is(&cnk, DM_FOURCC_VERS, 0)) {
diff --git a/src/io/Segment.c b/src/io/Segment.c
index 73d6b6c..9704138 100644
--- a/src/io/Segment.c
+++ b/src/io/Segment.c
@@ -101,8 +101,8 @@ static void DmSegment_parseChordItem(DmMessage_Chord* slf, DmRiff* rif) {
 	{
 		uint32_t end_position = rif->pos + item_size;
 
-		char16_t name[16];
-		DmRiff_read(rif, name, sizeof name);
+		uint16_t name[16];
+		DmRiff_read(rif, name, 16 * sizeof(int16_t));
 		(void) Dm_utf16ToUtf8Inline(slf->name, name);
 
 		DmRiff_readDword(rif, &slf->time);