diff --git a/tfhe/src/core_crypto/commons/math/decomposition/decomposer.rs b/tfhe/src/core_crypto/commons/math/decomposition/decomposer.rs
index 3862aa7463..0d765305ef 100644
--- a/tfhe/src/core_crypto/commons/math/decomposition/decomposer.rs
+++ b/tfhe/src/core_crypto/commons/math/decomposition/decomposer.rs
@@ -2,7 +2,7 @@ use crate::core_crypto::commons::ciphertext_modulus::CiphertextModulus;
 use crate::core_crypto::commons::math::decomposition::{
     SignedDecompositionIter, SignedDecompositionIterNonNative,
 };
-use crate::core_crypto::commons::numeric::{Numeric, UnsignedInteger};
+use crate::core_crypto::commons::numeric::UnsignedInteger;
 use crate::core_crypto::commons::parameters::{DecompositionBaseLog, DecompositionLevelCount};
 use crate::core_crypto::prelude::misc::divide_round_to_u128_custom_mod;
 use std::marker::PhantomData;
@@ -104,17 +104,20 @@ where
 
         // We compute the number of least significant bits which can not be represented by the
         // decomposition
-        let non_rep_bit_count: usize = <Scalar as Numeric>::BITS - self.level_count * self.base_log;
-        // We generate a mask which captures the non representable bits
-        let non_rep_mask = Scalar::ONE << (non_rep_bit_count - 1);
-        // We retrieve the non representable bits
-        let non_rep_bits = input & non_rep_mask;
-        // We extract the msb of the  non representable bits to perform the rounding
-        let non_rep_msb = non_rep_bits >> (non_rep_bit_count - 1);
-        // We remove the non-representable bits and perform the rounding
-        let res = input >> non_rep_bit_count;
-        let res = res + non_rep_msb;
-        res << non_rep_bit_count
+        // Example with level_count = 3, base_log = 4 and BITS == 64 -> 52
+        let non_rep_bit_count: usize = Scalar::BITS - self.level_count * self.base_log;
+        let shift = non_rep_bit_count - 1;
+        // Move the representable bits + 1 to the LSB, with our example :
+        //       |-----| 64 - (64 - 12 - 1) == 13 bits
+        // 0....0XX...XX
+        let mut res = input >> shift;
+        // Add one to do the rounding by adding the half interval
+        res += Scalar::ONE;
+        // Discard the LSB which was the one deciding in which direction we round
+        // -2 == 111...1110, i.e. all bits are 1 except the LSB which is 0 allowing to zero it
+        res &= Scalar::TWO.wrapping_neg();
+        // Shift back to the right position
+        res << shift
     }
 
     /// Generate an iterator over the terms of the decomposition of the input.
diff --git a/tfhe/src/core_crypto/fft_impl/common.rs b/tfhe/src/core_crypto/fft_impl/common.rs
index b1fee4b343..2297f9cc38 100644
--- a/tfhe/src/core_crypto/fft_impl/common.rs
+++ b/tfhe/src/core_crypto/fft_impl/common.rs
@@ -25,7 +25,7 @@ pub fn pbs_modulus_switch<Scalar: UnsignedTorus + CastInto<usize>>(
     // Start doing the right shift
     output >>= Scalar::BITS - poly_size.log2().0 - 2 + lut_count_log.0;
     // Do the rounding
-    output += output & Scalar::ONE;
+    output += Scalar::ONE;
     // Finish the right shift
     output >>= 1;
     // Apply the lsb padding