diff --git a/Assets/dll/bsnes.wbx.zst b/Assets/dll/bsnes.wbx.zst
index b7f9a2da65..a43f9bba32 100644
Binary files a/Assets/dll/bsnes.wbx.zst and b/Assets/dll/bsnes.wbx.zst differ
diff --git a/waterbox/bsnescore/bsnes/emulator/emulator.hpp b/waterbox/bsnescore/bsnes/emulator/emulator.hpp
index 71cd809a5f..1ea50f351b 100644
--- a/waterbox/bsnescore/bsnes/emulator/emulator.hpp
+++ b/waterbox/bsnescore/bsnes/emulator/emulator.hpp
@@ -35,7 +35,7 @@ namespace Emulator {
   static const string Website   = "https://bsnes.dev";
 
   //incremented only when serialization format changes
-  static const string SerializerVersion = "115";
+  static const string SerializerVersion = "115.1";
 
   namespace Constants {
     namespace Colorburst {
diff --git a/waterbox/bsnescore/bsnes/sfc/cpu/cpu.hpp b/waterbox/bsnescore/bsnes/sfc/cpu/cpu.hpp
index e636f8e8e5..878a094173 100644
--- a/waterbox/bsnescore/bsnes/sfc/cpu/cpu.hpp
+++ b/waterbox/bsnescore/bsnes/sfc/cpu/cpu.hpp
@@ -121,6 +121,12 @@ private:
     bool hdmaMode = 0;  //0 = init, 1 = run
 
     uint autoJoypadCounter = 33;  //state machine; 4224 / 128 = 33 (inactive)
+
+    uint2 autoJoypadPort1 = 0;
+    uint2 autoJoypadPort2 = 0;
+
+    bool cpuLatch = false;
+    bool autoJoypadLatch = false;
   } status;
 
   struct IO {
diff --git a/waterbox/bsnescore/bsnes/sfc/cpu/io.cpp b/waterbox/bsnescore/bsnes/sfc/cpu/io.cpp
index 4fd0901cbe..2b168c9357 100644
--- a/waterbox/bsnescore/bsnes/sfc/cpu/io.cpp
+++ b/waterbox/bsnescore/bsnes/sfc/cpu/io.cpp
@@ -132,13 +132,22 @@ auto CPU::writeCPU(uint addr, uint8 data) -> void {
     //bit 0 is shared between JOYSER0 and JOYSER1:
     //strobing $4016.d0 affects both controller port latches.
     //$4017 bit 0 writes are ignored.
-    controllerPort1.device->latch(data & 1);
-    controllerPort2.device->latch(data & 1);
+    status.cpuLatch = data & 1;
+    controllerPort1.device->latch(status.autoJoypadLatch | status.cpuLatch);
+    controllerPort2.device->latch(status.autoJoypadLatch | status.cpuLatch);
     return;
 
   case 0x4200:  //NMITIMEN
     io.autoJoypadPoll = data & 1;
-    if(!io.autoJoypadPoll) status.autoJoypadCounter = 33; // Disable auto-joypad read
+    if(status.autoJoypadCounter == 0) {
+      // allow controller latches during this time
+      status.autoJoypadLatch = io.autoJoypadPoll;
+      controllerPort1.device->latch(status.autoJoypadLatch | status.cpuLatch);
+      controllerPort2.device->latch(status.autoJoypadLatch | status.cpuLatch);
+    } else if (!io.autoJoypadPoll && status.autoJoypadCounter >= 2) {
+      status.autoJoypadCounter = 33;
+    }
+
     nmitimenUpdate(data);
     return;
 
diff --git a/waterbox/bsnescore/bsnes/sfc/cpu/serialization.cpp b/waterbox/bsnescore/bsnes/sfc/cpu/serialization.cpp
index 7a1c240100..ec6f2e399c 100644
--- a/waterbox/bsnescore/bsnes/sfc/cpu/serialization.cpp
+++ b/waterbox/bsnescore/bsnes/sfc/cpu/serialization.cpp
@@ -45,6 +45,12 @@ auto CPU::serialize(serializer& s) -> void {
 
   s.integer(status.autoJoypadCounter);
 
+  s.integer(status.autoJoypadPort1);
+  s.integer(status.autoJoypadPort2);
+
+  s.boolean(status.cpuLatch);
+  s.boolean(status.autoJoypadLatch);
+
   s.integer(io.wramAddress);
 
   s.boolean(io.hirqEnable);
diff --git a/waterbox/bsnescore/bsnes/sfc/cpu/timing.cpp b/waterbox/bsnescore/bsnes/sfc/cpu/timing.cpp
index d0e64022f6..333f427803 100644
--- a/waterbox/bsnescore/bsnes/sfc/cpu/timing.cpp
+++ b/waterbox/bsnescore/bsnes/sfc/cpu/timing.cpp
@@ -211,45 +211,54 @@ auto CPU::dmaEdge() -> void {
 
 //called every 128 clocks from inside the CPU::stepOnce() function
 auto CPU::joypadEdge() -> void {
-  //it is not yet confirmed if polling can be stopped early and/or (re)started later
-  if(!io.autoJoypadPoll) return;
-
-  if(vcounter() == ppu.vdisp() && hcounter() >= 130 && hcounter() <= 256) {
+  if(vcounter() == ppu.vdisp() && (counter.cpu & 255) == 0 && hcounter() >= 130 && hcounter() <= 384) {
     //begin new polling sequence
     status.autoJoypadCounter = 0;
-  }
+  } else {
+    //stop after polling has been completed for this frame
+    if(status.autoJoypadCounter >= 33) return;
 
-  //stop after polling has been completed for this frame
-  if(status.autoJoypadCounter >= 33) return;
+    status.autoJoypadCounter++;
+  }
 
   if(status.autoJoypadCounter == 0) {
     //latch controller states on the first polling cycle
-    controllerPort1.device->latch(1);
-    controllerPort2.device->latch(1);
+    status.autoJoypadLatch = io.autoJoypadPoll;
+    controllerPort1.device->latch(status.autoJoypadLatch | status.cpuLatch);
+    controllerPort2.device->latch(status.autoJoypadLatch | status.cpuLatch);
+    if(io.autoJoypadPoll) {
+      //shift registers are cleared to zero at start of auto-joypad polling
+      io.joy1 = 0;
+      io.joy2 = 0;
+      io.joy3 = 0;
+      io.joy4 = 0;
+    }
   }
 
   if(status.autoJoypadCounter == 1) {
     //release latch and begin reading on the second cycle
-    controllerPort1.device->latch(0);
-    controllerPort2.device->latch(0);
-
-    //shift registers are cleared to zero at start of auto-joypad polling
-    io.joy1 = 0;
-    io.joy2 = 0;
-    io.joy3 = 0;
-    io.joy4 = 0;
+    status.autoJoypadLatch = 0;
+    controllerPort1.device->latch(status.autoJoypadLatch | status.cpuLatch);
+    controllerPort2.device->latch(status.autoJoypadLatch | status.cpuLatch);
   }
 
-  if(status.autoJoypadCounter >= 2 && !(status.autoJoypadCounter & 1)) {
+  if(status.autoJoypadCounter != 1 && !io.autoJoypadPoll) {
+    // if auto-joypad polling is disabled at this point skip the rest of the polling
+    status.autoJoypadCounter = 33;
+    return;
+  }
+
+  if(status.autoJoypadCounter >= 2) {
     //sixteen bits are shifted into joy{1-4}, one bit per 256 clocks
-    uint2 port0 = controllerPort1.device->data();
-    uint2 port1 = controllerPort2.device->data();
-
-    io.joy1 = io.joy1 << 1 | port0.bit(0);
-    io.joy2 = io.joy2 << 1 | port1.bit(0);
-    io.joy3 = io.joy3 << 1 | port0.bit(1);
-    io.joy4 = io.joy4 << 1 | port1.bit(1);
+    //the bits are read on one 128-clock cycle and written on the next
+    if ((status.autoJoypadCounter & 1) == 0) {
+      status.autoJoypadPort1 = controllerPort1.device->data();
+      status.autoJoypadPort2 = controllerPort2.device->data();
+    } else {
+      io.joy1 = io.joy1 << 1 | status.autoJoypadPort1.bit(0);
+      io.joy2 = io.joy2 << 1 | status.autoJoypadPort2.bit(0);
+      io.joy3 = io.joy3 << 1 | status.autoJoypadPort1.bit(1);
+      io.joy4 = io.joy4 << 1 | status.autoJoypadPort2.bit(1);
+    }
   }
-
-  status.autoJoypadCounter++;
 }
diff --git a/waterbox/bsnescore/bsnes/sfc/ppu/background.cpp b/waterbox/bsnescore/bsnes/sfc/ppu/background.cpp
index e9d3dcb3b9..7f04726c34 100644
--- a/waterbox/bsnescore/bsnes/sfc/ppu/background.cpp
+++ b/waterbox/bsnescore/bsnes/sfc/ppu/background.cpp
@@ -61,7 +61,7 @@ auto PPU::Background::fetchNameTable() -> void {
         if(!(hlookup & 0x8000)) {
           hoffset = hpixel + (hlookup & ~7) + (hscroll & 7);
         } else {
-          voffset = vpixel + (vlookup);
+          voffset = vpixel + (hlookup);
         }
       }
     } else {