diff --git a/packages/voice/src/networking/Networking.ts b/packages/voice/src/networking/Networking.ts
index ded6c85a03ee..5004baed38d1 100644
--- a/packages/voice/src/networking/Networking.ts
+++ b/packages/voice/src/networking/Networking.ts
@@ -8,6 +8,7 @@ import type { VoiceReceivePayload, VoiceSpeakingFlags } from 'discord-api-types/
 import { VoiceEncryptionMode, VoiceOpcodes } from 'discord-api-types/voice/v8';
 import type { CloseEvent } from 'ws';
 import * as secretbox from '../util/Secretbox';
+import { RTP_OPUS_PAYLOAD_TYPE } from '../util/constants';
 import { noop } from '../util/util';
 import { DAVESession, getMaxProtocolVersion } from './DAVESession';
 import { VoiceUDPSocket } from './VoiceUDPSocket';
@@ -745,7 +746,7 @@ export class Networking extends EventEmitter {
 	private createAudioPacket(opusPacket: Buffer, connectionData: ConnectionData, daveSession?: DAVESession) {
 		const rtpHeader = Buffer.alloc(12);
 		rtpHeader[0] = 0x80;
-		rtpHeader[1] = 0x78;
+		rtpHeader[1] = RTP_OPUS_PAYLOAD_TYPE;
 
 		const { sequence, timestamp, ssrc } = connectionData;
 
diff --git a/packages/voice/src/receive/VoiceReceiver.ts b/packages/voice/src/receive/VoiceReceiver.ts
index 90e4015cb9b1..94fd95f683c0 100644
--- a/packages/voice/src/receive/VoiceReceiver.ts
+++ b/packages/voice/src/receive/VoiceReceiver.ts
@@ -7,6 +7,7 @@ import { VoiceOpcodes } from 'discord-api-types/voice/v8';
 import { VoiceConnectionStatus, type VoiceConnection } from '../VoiceConnection';
 import { NetworkingStatusCode, type ConnectionData } from '../networking/Networking';
 import { methods } from '../util/Secretbox';
+import { RTP_OPUS_PAYLOAD_TYPE } from '../util/constants';
 import {
 	AudioReceiveStream,
 	createDefaultAudioReceiveStreamOptions,
@@ -137,6 +138,15 @@ export class VoiceReceiver {
 		let packet: Buffer = this.decrypt(buffer, mode, nonce, secretKey);
 		if (!packet) throw new Error('Failed to parse packet');
 
+		// Strip padding (RFC3550 5.1)
+		const hasPadding = buffer[0] && Boolean(buffer[0] & 0b100000);
+		if (hasPadding) {
+			const paddingAmount = packet[packet.length - 1]!;
+			if (paddingAmount < packet.length) {
+				packet = packet.subarray(0, packet.length - paddingAmount);
+			}
+		}
+
 		// Strip decrypted RTP Header Extension if present
 		// The header is only indicated in the original data, so compare with buffer first
 		if (buffer.subarray(12, 14).compare(HEADER_EXTENSION_BYTE) === 0) {
@@ -176,6 +186,13 @@ export class VoiceReceiver {
 		if (!stream) return;
 
 		if (this.connectionData.encryptionMode && this.connectionData.nonceBuffer && this.connectionData.secretKey) {
+			// As a guard, we shouldn't parse packets that (1) aren't voice packets and (2) are not in the right RTP version
+			if ((msg[1]! & 0x7f) !== RTP_OPUS_PAYLOAD_TYPE) return;
+
+			// Ignore packets not in RTP version 2
+			const rtpVersion = msg[0]! >> 6;
+			if (rtpVersion !== 2) return;
+
 			try {
 				const packet = this.parsePacket(
 					msg,
diff --git a/packages/voice/src/util/constants.ts b/packages/voice/src/util/constants.ts
new file mode 100644
index 000000000000..4b3d2225c6fa
--- /dev/null
+++ b/packages/voice/src/util/constants.ts
@@ -0,0 +1 @@
+export const RTP_OPUS_PAYLOAD_TYPE = 0x78;
diff --git a/packages/voice/src/util/index.ts b/packages/voice/src/util/index.ts
index 265b33ec40cd..5856d2efcaed 100644
--- a/packages/voice/src/util/index.ts
+++ b/packages/voice/src/util/index.ts
@@ -1,3 +1,4 @@
+export * from './constants';
 export * from './generateDependencyReport';
 export * from './entersState';
 export type * from './adapter';