diff --git a/libraries/muxer/src/main/java/androidx/media3/muxer/Boxes.java b/libraries/muxer/src/main/java/androidx/media3/muxer/Boxes.java
index 10851ca00d..6db416fb23 100644
--- a/libraries/muxer/src/main/java/androidx/media3/muxer/Boxes.java
+++ b/libraries/muxer/src/main/java/androidx/media3/muxer/Boxes.java
@@ -42,7 +42,6 @@ import java.util.ArrayList;
 import java.util.Arrays;
 import java.util.Collections;
 import java.util.List;
-import java.util.Locale;
 
 /**
  * Writes out various types of boxes as per MP4 (ISO/IEC 14496-12) standards.
@@ -636,14 +635,13 @@ import java.util.Locale;
    * @return A list of all the sample durations.
    */
   // TODO: b/280084657 - Add support for setting last sample duration.
-  // TODO: b/317373578 - Consider changing return type to List<Integer>.
-  public static List<Long> convertPresentationTimestampsToDurationsVu(
+  public static List<Integer> convertPresentationTimestampsToDurationsVu(
       List<BufferInfo> samplesInfo,
       long firstSamplePresentationTimeUs,
       int videoUnitTimescale,
       @Mp4Muxer.LastFrameDurationBehavior int lastDurationBehavior) {
     List<Long> presentationTimestampsUs = new ArrayList<>(samplesInfo.size());
-    List<Long> durationsVu = new ArrayList<>(samplesInfo.size());
+    List<Integer> durationsVu = new ArrayList<>(samplesInfo.size());
 
     if (samplesInfo.isEmpty()) {
       return durationsVu;
@@ -672,23 +670,20 @@ import java.util.Locale;
       long currentSampleDurationVu =
           vuFromUs(nextSampleTimeUs, videoUnitTimescale)
               - vuFromUs(currentSampleTimeUs, videoUnitTimescale);
-      if (currentSampleDurationVu > Integer.MAX_VALUE) {
-        throw new IllegalArgumentException(
-            String.format(
-                Locale.US, "Timestamp delta %d doesn't fit into an int", currentSampleDurationVu));
-      }
-      durationsVu.add(currentSampleDurationVu);
+      checkState(
+          currentSampleDurationVu <= Integer.MAX_VALUE, "Only 32-bit sample duration is allowed");
+      durationsVu.add((int) currentSampleDurationVu);
       currentSampleTimeUs = nextSampleTimeUs;
     }
     // Default duration for the last sample.
-    durationsVu.add(0L);
+    durationsVu.add(0);
 
     adjustLastSampleDuration(durationsVu, lastDurationBehavior);
     return durationsVu;
   }
 
   /** Generates the stts (decoding time to sample) box. */
-  public static ByteBuffer stts(List<Long> durationsVu) {
+  public static ByteBuffer stts(List<Integer> durationsVu) {
     ByteBuffer contents = ByteBuffer.allocate(durationsVu.size() * 8 + MAX_FIXED_LEAF_BOX_SIZE);
 
     contents.putInt(0x0); // version and flags.
@@ -705,7 +700,7 @@ import java.util.Locale;
     // Note that the framework MediaMuxer adjust time deltas within plus-minus 100 us, so that
     // samples have repeating duration values. It saves few entries in the table.
     for (int i = 0; i < durationsVu.size(); i++) {
-      long durationVu = durationsVu.get(i);
+      int durationVu = durationsVu.get(i);
       if (lastDurationVu != durationVu) {
         lastDurationVu = durationVu;
         lastSampleCountIndex = contents.position();
@@ -713,7 +708,7 @@ import java.util.Locale;
         // sample_count; this will be updated instead of adding a new entry if the next sample has
         // the same duration.
         contents.putInt(1);
-        contents.putInt((int) durationVu); // sample_delta.
+        contents.putInt(durationVu); // sample_delta
         totalEntryCount++;
       } else {
         contents.putInt(lastSampleCountIndex, contents.getInt(lastSampleCountIndex) + 1);
@@ -728,7 +723,7 @@ import java.util.Locale;
 
   /** Returns the ctts (composition time to sample) box. */
   public static ByteBuffer ctts(
-      List<BufferInfo> samplesInfo, List<Long> durationVu, int videoUnitTimescale) {
+      List<BufferInfo> samplesInfo, List<Integer> durationVu, int videoUnitTimescale) {
     // Generate the sample composition offsets list to create ctts box.
     List<Integer> compositionOffsets =
         Boxes.calculateSampleCompositionTimeOffsets(samplesInfo, durationVu, videoUnitTimescale);
@@ -786,7 +781,7 @@ import java.util.Locale;
    * @return A list of all the sample composition time offsets.
    */
   public static List<Integer> calculateSampleCompositionTimeOffsets(
-      List<BufferInfo> samplesInfo, List<Long> durationVu, int videoUnitTimescale) {
+      List<BufferInfo> samplesInfo, List<Integer> durationVu, int videoUnitTimescale) {
     List<Integer> compositionOffsets = new ArrayList<>(samplesInfo.size());
     if (samplesInfo.isEmpty()) {
       return compositionOffsets;
@@ -802,7 +797,9 @@ import java.util.Locale;
           samplesInfo.get(sampleId).presentationTimeUs - firstSamplePresentationTimeUs;
       long currentCompositionOffsetVu =
           vuFromUs(currentSampleCompositionTimeUs, videoUnitTimescale) - currentSampleDecodeTime;
-      checkState(currentCompositionOffsetVu <= Integer.MAX_VALUE, "Only 32-bit offset is allowed");
+      checkState(
+          currentCompositionOffsetVu <= Integer.MAX_VALUE,
+          "Only 32-bit composition offset is allowed");
       currentSampleDecodeTime += durationVu.get(sampleId); // DT(n+1) = DT(n) + STTS(n)
       compositionOffsets.add((int) currentCompositionOffsetVu);
 
@@ -1022,7 +1019,7 @@ import java.util.Locale;
     contents.putInt(dataOffset); // A signed int(32)
     for (int i = 0; i < samplesMetadata.size(); i++) {
       SampleMetadata currentSampleMetadata = samplesMetadata.get(i);
-      contents.putInt((int) currentSampleMetadata.durationVu); // An unsigned int(32)
+      contents.putInt(currentSampleMetadata.durationVu); // An unsigned int(32)
       contents.putInt(currentSampleMetadata.size); // An unsigned int(32)
       contents.putInt(
           (currentSampleMetadata.flags & MediaCodec.BUFFER_FLAG_KEY_FRAME) != 0
@@ -1064,7 +1061,7 @@ import java.util.Locale;
   // TODO: b/317117431 - Change this method to getLastSampleDuration().
   /** Adjusts the duration of the very last sample if needed. */
   private static void adjustLastSampleDuration(
-      List<Long> durationsToBeAdjustedVu, @Mp4Muxer.LastFrameDurationBehavior int behavior) {
+      List<Integer> durationsToBeAdjustedVu, @Mp4Muxer.LastFrameDurationBehavior int behavior) {
     // Technically, MP4 file stores frame durations, not timestamps. If a frame starts at a
     // given timestamp then the duration of the last frame is not obvious. If samples follow each
     // other in roughly regular intervals (e.g. in a normal, 30 fps video), it can be safely assumed
diff --git a/libraries/muxer/src/main/java/androidx/media3/muxer/FragmentedMp4Writer.java b/libraries/muxer/src/main/java/androidx/media3/muxer/FragmentedMp4Writer.java
index 00b4a90adf..474133b2f9 100644
--- a/libraries/muxer/src/main/java/androidx/media3/muxer/FragmentedMp4Writer.java
+++ b/libraries/muxer/src/main/java/androidx/media3/muxer/FragmentedMp4Writer.java
@@ -49,12 +49,12 @@ import org.checkerframework.checker.nullness.qual.MonotonicNonNull;
 /* package */ final class FragmentedMp4Writer {
   /** Provides a limited set of sample metadata. */
   public static class SampleMetadata {
-    public final long durationVu;
+    public final int durationVu;
     public final int size;
     public final int flags;
     public final int compositionTimeOffsetVu;
 
-    public SampleMetadata(long durationsVu, int size, int flags, int compositionTimeOffsetVu) {
+    public SampleMetadata(int durationsVu, int size, int flags, int compositionTimeOffsetVu) {
       this.durationVu = durationsVu;
       this.size = size;
       this.flags = flags;
@@ -320,7 +320,7 @@ import org.checkerframework.checker.nullness.qual.MonotonicNonNull;
 
     boolean hasBFrame = false;
     ImmutableList<BufferInfo> pendingSamplesBufferInfo = pendingSamplesBufferInfoBuilder.build();
-    List<Long> sampleDurations =
+    List<Integer> sampleDurations =
         Boxes.convertPresentationTimestampsToDurationsVu(
             pendingSamplesBufferInfo,
             /* firstSamplePresentationTimeUs= */ currentFragmentSequenceNumber == 1
diff --git a/libraries/muxer/src/main/java/androidx/media3/muxer/Mp4MoovStructure.java b/libraries/muxer/src/main/java/androidx/media3/muxer/Mp4MoovStructure.java
index 8565430ee7..6eb348866f 100644
--- a/libraries/muxer/src/main/java/androidx/media3/muxer/Mp4MoovStructure.java
+++ b/libraries/muxer/src/main/java/androidx/media3/muxer/Mp4MoovStructure.java
@@ -79,7 +79,7 @@ import org.checkerframework.checker.nullness.qual.PolyNull;
       String languageCode = bcp47LanguageTagToIso3(format.language);
 
       // Generate the sample durations to calculate the total duration for tkhd box.
-      List<Long> sampleDurationsVu =
+      List<Integer> sampleDurationsVu =
           Boxes.convertPresentationTimestampsToDurationsVu(
               track.writtenSamples(),
               minInputPtsUs,
diff --git a/libraries/muxer/src/test/java/androidx/media3/muxer/BoxesTest.java b/libraries/muxer/src/test/java/androidx/media3/muxer/BoxesTest.java
index a6710fb37b..44e74aac17 100644
--- a/libraries/muxer/src/test/java/androidx/media3/muxer/BoxesTest.java
+++ b/libraries/muxer/src/test/java/androidx/media3/muxer/BoxesTest.java
@@ -403,14 +403,14 @@ public class BoxesTest {
     List<MediaCodec.BufferInfo> sampleBufferInfos =
         createBufferInfoListWithSamplePresentationTimestamps(0L);
 
-    List<Long> durationsVu =
+    List<Integer> durationsVu =
         Boxes.convertPresentationTimestampsToDurationsVu(
             sampleBufferInfos,
             /* firstSamplePresentationTimeUs= */ 0L,
             VU_TIMEBASE,
             LAST_FRAME_DURATION_BEHAVIOR_INSERT_SHORT_FRAME);
 
-    assertThat(durationsVu).containsExactly(0L);
+    assertThat(durationsVu).containsExactly(0);
   }
 
   @Test
@@ -419,14 +419,14 @@ public class BoxesTest {
     List<MediaCodec.BufferInfo> sampleBufferInfos =
         createBufferInfoListWithSamplePresentationTimestamps(5_000L);
 
-    List<Long> durationsVu =
+    List<Integer> durationsVu =
         Boxes.convertPresentationTimestampsToDurationsVu(
             sampleBufferInfos,
             /* firstSamplePresentationTimeUs= */ 0L,
             VU_TIMEBASE,
             LAST_FRAME_DURATION_BEHAVIOR_INSERT_SHORT_FRAME);
 
-    assertThat(durationsVu).containsExactly(0L);
+    assertThat(durationsVu).containsExactly(0);
   }
 
   @Test
@@ -435,14 +435,14 @@ public class BoxesTest {
     List<MediaCodec.BufferInfo> sampleBufferInfos =
         createBufferInfoListWithSamplePresentationTimestamps(0L, 30_000L, 80_000L);
 
-    List<Long> durationsVu =
+    List<Integer> durationsVu =
         Boxes.convertPresentationTimestampsToDurationsVu(
             sampleBufferInfos,
             /* firstSamplePresentationTimeUs= */ 0L,
             VU_TIMEBASE,
             LAST_FRAME_DURATION_BEHAVIOR_INSERT_SHORT_FRAME);
 
-    assertThat(durationsVu).containsExactly(3_000L, 5_000L, 0L);
+    assertThat(durationsVu).containsExactly(3_000, 5_000, 0);
   }
 
   @Test
@@ -451,14 +451,14 @@ public class BoxesTest {
     List<MediaCodec.BufferInfo> sampleBufferInfos =
         createBufferInfoListWithSamplePresentationTimestamps(0L, 30_000L, 80_000L);
 
-    List<Long> durationsVu =
+    List<Integer> durationsVu =
         Boxes.convertPresentationTimestampsToDurationsVu(
             sampleBufferInfos,
             /* firstSamplePresentationTimeUs= */ 0L,
             VU_TIMEBASE,
             LAST_FRAME_DURATION_BEHAVIOR_DUPLICATE_PREV_DURATION);
 
-    assertThat(durationsVu).containsExactly(3_000L, 5_000L, 5_000L);
+    assertThat(durationsVu).containsExactly(3_000, 5_000, 5_000);
   }
 
   @Test
@@ -467,19 +467,19 @@ public class BoxesTest {
     List<MediaCodec.BufferInfo> sampleBufferInfos =
         createBufferInfoListWithSamplePresentationTimestamps(0L, 10_000L, 1_000L, 2_000L, 11_000L);
 
-    List<Long> durationsVu =
+    List<Integer> durationsVu =
         Boxes.convertPresentationTimestampsToDurationsVu(
             sampleBufferInfos,
             /* firstSamplePresentationTimeUs= */ 0L,
             VU_TIMEBASE,
             LAST_FRAME_DURATION_BEHAVIOR_INSERT_SHORT_FRAME);
 
-    assertThat(durationsVu).containsExactly(100L, 100L, 800L, 100L, 0L);
+    assertThat(durationsVu).containsExactly(100, 100, 800, 100, 0);
   }
 
   @Test
   public void createSttsBox_withSingleSampleDuration_matchesExpected() throws IOException {
-    ImmutableList<Long> sampleDurations = ImmutableList.of(500L);
+    ImmutableList<Integer> sampleDurations = ImmutableList.of(500);
 
     ByteBuffer sttsBox = Boxes.stts(sampleDurations);
 
@@ -492,7 +492,7 @@ public class BoxesTest {
 
   @Test
   public void createSttsBox_withAllDifferentSampleDurations_matchesExpected() throws IOException {
-    ImmutableList<Long> sampleDurations = ImmutableList.of(1_000L, 2_000L, 3_000L, 5_000L);
+    ImmutableList<Integer> sampleDurations = ImmutableList.of(1_000, 2_000, 3_000, 5_000);
 
     ByteBuffer sttsBox = Boxes.stts(sampleDurations);
 
@@ -506,7 +506,7 @@ public class BoxesTest {
   @Test
   public void createSttsBox_withFewConsecutiveSameSampleDurations_matchesExpected()
       throws IOException {
-    ImmutableList<Long> sampleDurations = ImmutableList.of(1_000L, 2_000L, 2_000L, 2_000L);
+    ImmutableList<Integer> sampleDurations = ImmutableList.of(1_000, 2_000, 2_000, 2_000);
 
     ByteBuffer sttsBox = Boxes.stts(sampleDurations);
 
@@ -521,7 +521,7 @@ public class BoxesTest {
   public void createCttsBox_withSingleSampleTimestamp_returnsEmptyBox() {
     List<MediaCodec.BufferInfo> sampleBufferInfos =
         createBufferInfoListWithSamplePresentationTimestamps(400);
-    List<Long> durationsVu =
+    List<Integer> durationsVu =
         Boxes.convertPresentationTimestampsToDurationsVu(
             sampleBufferInfos,
             /* firstSamplePresentationTimeUs= */ 0L,
@@ -538,7 +538,7 @@ public class BoxesTest {
   public void createCttsBox_withNoBframesSampleTimestamps_returnsEmptyBox() throws IOException {
     List<MediaCodec.BufferInfo> sampleBufferInfos =
         createBufferInfoListWithSamplePresentationTimestamps(0L, 1000L, 2000L);
-    List<Long> durationsVu =
+    List<Integer> durationsVu =
         Boxes.convertPresentationTimestampsToDurationsVu(
             sampleBufferInfos,
             /* firstSamplePresentationTimeUs= */ 0L,
@@ -557,7 +557,7 @@ public class BoxesTest {
         createBufferInfoListWithSamplePresentationTimestamps(
             0, 400, 200, 100, 300, 800, 600, 500, 700);
 
-    List<Long> durationsVu =
+    List<Integer> durationsVu =
         Boxes.convertPresentationTimestampsToDurationsVu(
             sampleBufferInfos,
             /* firstSamplePresentationTimeUs= */ 0L,
@@ -577,7 +577,7 @@ public class BoxesTest {
         createBufferInfoListWithSamplePresentationTimestamps(
             23698215060L, 23698248252L, 23698347988L, 23698488968L, 23698547416L);
 
-    List<Long> durationsVu =
+    List<Integer> durationsVu =
         Boxes.convertPresentationTimestampsToDurationsVu(
             sampleBufferInfos,
             /* firstSamplePresentationTimeUs= */ 23698215060L,
@@ -679,7 +679,7 @@ public class BoxesTest {
     for (int i = 0; i < sampleCount; i++) {
       samplesMetadata.add(
           new SampleMetadata(
-              /* durationsVu= */ 2_000L,
+              /* durationsVu= */ 2_000,
               /* size= */ 5_000,
               /* flags= */ i == 0 ? MediaCodec.BUFFER_FLAG_KEY_FRAME : 0,
               /* compositionTimeOffsetVu= */ 0));
@@ -700,7 +700,7 @@ public class BoxesTest {
     for (int i = 0; i < sampleCount; i++) {
       samplesMetadata.add(
           new SampleMetadata(
-              /* durationsVu= */ 2_000L,
+              /* durationsVu= */ 2_000,
               /* size= */ 5_000,
               /* flags= */ i == 0 ? MediaCodec.BUFFER_FLAG_KEY_FRAME : 0,
               /* compositionTimeOffsetVu= */ 100));