ffmpeg: store divided frames in memory cache

Since users will most likely be working with a divider most of the time,
there's no reason to waste memory space with undivided frames considering
the dividing process is quite fast.

app/render/videoparams.cpp

@@ -145,9 +145,8 @@ bool VideoParams::operator!=(const VideoParams &rhs) const
 
 void VideoParams::calculate_effective_size()
 {
-  // Fast rounding up to an even number
-  effective_width_ = qCeil(width() / divider_ * 0.5) * 2;
-  effective_height_ = qCeil(height() / divider_ * 0.5) * 2;
+  effective_width_ = GetScaledDimension(width(), divider_);
+  effective_height_ = GetScaledDimension(height(), divider_);
 }
 
 void VideoParams::validate_pixel_aspect_ratio()
@@ -195,4 +194,9 @@ QString VideoParams::FormatPixelAspectRatioString(const QString &format, const r
   return format.arg(QString::number(ratio.toDouble(), 'f', 4));
 }
 
+int VideoParams::GetScaledDimension(int dim, int divider)
+{
+  return qCeil(dim / divider * 0.5) * 2;
+}
+
 OLIVE_NAMESPACE_EXIT

app/render/videoparams.h

@@ -114,6 +114,8 @@ public:
   static QStringList GetStandardPixelAspectRatioNames();
   static QString FormatPixelAspectRatioString(const QString& format, const rational& ratio);
 
+  static int GetScaledDimension(int dim, int divider);
+
 private:
   void calculate_effective_size();