Fix spelling mistakes (#1436)

Signed-off-by: Julian Amann <julian.amann@tum.de>

docs/TechnicalIntroduction.rst

@@ -827,13 +827,13 @@ Supported compression schemes:
    * - DWAA (lossy)
 
      - Lossy compression of RGB data by quantizing discrete cosine
-       transorm (DCT) components, in blocks of 32 scanlines. More
+       transform (DCT) components, in blocks of 32 scanlines. More
        efficient for partial buffer access.
 
    * - DWAB (lossy)
 
      - Lossy compression of RGB data by quantizing discrete cosine
-       transorm (DCT) components, in blocks of 256 scanlines. More
+       transform (DCT) components, in blocks of 256 scanlines. More
        efficient space wise and faster to decode full frames than DWAA
        access.
 

src/lib/OpenEXR/ImfIDManifest.cpp

@@ -247,7 +247,7 @@ writeStringList (char*& outPtr, const T& stringList, int entries = 0)
         //
         // variable length encoding:
         // values between 0 and 127 inclusive are stored in a single byte
-        // values betwwen 128 and 16384 are encoded with two bytes: 1LLLLLLL 0MMMMMMMM where L and M are the least and most significant bits of the value
+        // values between 128 and 16384 are encoded with two bytes: 1LLLLLLL 0MMMMMMMM where L and M are the least and most significant bits of the value
         // in general, values are stored least significant values first, with the top bit of each byte indicating more values follow
         // the top bit is clear in the last byte of the value
         // (this scheme requires two bytes to store values above 1<<7, and five bytes to store values above 1<<28)

src/lib/OpenEXRCore/internal_dwa_simd.h

@@ -362,7 +362,7 @@ convertFloatToHalf64_f16c (uint16_t* dst, float* src)
     //
     // Now, it's quite likely that we'll find ourselves in situations
     // where we want to build *without* VEX, in order to maintain
-    // maximum compatability. But to get there with intrinsics,
+    // maximum compatibility. But to get there with intrinsics,
     // we'd need to break out code into a separate file. Bleh.
     // I'll take the asm.
     //
@@ -542,7 +542,7 @@ fromHalfZigZag_scalar (uint16_t* src, float* dst)
 // the even rows are in descending order.
 //
 // If we 'fold' the bottom half up into the top, we can preserve ordered
-// runs accross rows, and still keep all the correct values in columns.
+// runs across rows, and still keep all the correct values in columns.
 // After transposing, we'll need to rotate things back into place.
 // This gives us:
 //
@@ -628,7 +628,7 @@ fromHalfZigZag_f16c (uint16_t* src, float* dst)
 
          /* Reverse the even rows. We're not using PSHUFB as
             * that requires loading an extra constant all the time,
-            * and we're alreadly pretty memory bound.
+            * and we're already pretty memory bound.
             */
 
          "vpshuflw $0x1b, %%xmm0, %%xmm0          \n"