diff --git a/ChangeLog b/ChangeLog
index 36217b0e..aec48727 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -7,6 +7,8 @@
 - support tiff XMP metadata
 - support @density arg to magickload [Lovell]
 - support python3.4 and python2.7 in new python binding
+- vips_gaussmat() and vips_logmat() are now int by default, to match
+  vips_conv(), and use @precision, not @integer
 
 25/7/14 started 7.41.0
 - start working on --disable-deprecated
diff --git a/TODO b/TODO
index 797bfee0..5dce5c3f 100644
--- a/TODO
+++ b/TODO
@@ -1,15 +1,3 @@
-- try:
-
-	vips gaussmat m2.v 2 0.1 --separable
-	vips convsep k2.jpg x2.jpg m2.v
-	eog x2.jpg
-
-  it's too bright ... because gaussmat makes a float mask by default, but
-  conv does an int convolution by default
-
-  rather than an --integer switch on gaussmat, should it take --precision
-  instead, and default to INT, like conv?
-
 - configure should check for pygobject too
 
 	PKG_CHECK_MODULES(PYTHON, [pygobject-3.0 >= $PYGOBJECT_REQUIRED])
diff --git a/libvips/convolution/gaussblur.c b/libvips/convolution/gaussblur.c
index d1bf2ac9..322b656c 100644
--- a/libvips/convolution/gaussblur.c
+++ b/libvips/convolution/gaussblur.c
@@ -75,7 +75,7 @@ vips_gaussblur_build( VipsObject *object )
 
 	if( vips_gaussmat( &t[0], gaussblur->sigma, gaussblur->min_ampl, 
 		"separable", TRUE,
-		"integer", gaussblur->precision != VIPS_PRECISION_FLOAT,
+		"precision", gaussblur->precision,
 		NULL ) )
 		return( -1 ); 
 
diff --git a/libvips/convolution/sharpen.c b/libvips/convolution/sharpen.c
index 0a2fc3fb..c4424c73 100644
--- a/libvips/convolution/sharpen.c
+++ b/libvips/convolution/sharpen.c
@@ -236,7 +236,7 @@ vips_sharpen_build( VipsObject *object )
 	 */
 	if( vips_gaussmat( &t[1], 1 + sharpen->radius / 2, 0.2, 
 		"separable", TRUE,
-		"integer", TRUE,
+		"precision", VIPS_PRECISION_INTEGER,
 		NULL ) )
 		return( -1 ); 
 
diff --git a/libvips/create/gaussmat.c b/libvips/create/gaussmat.c
index 1b83f836..e0de1a9a 100644
--- a/libvips/create/gaussmat.c
+++ b/libvips/create/gaussmat.c
@@ -15,6 +15,9 @@
  * 	- gtkdoc
  * 20/10/13
  * 	- redone as a class 
+ * 16/12/14
+ * 	- default to int output to match vips_conv()
+ * 	- use @precision, not @integer
  */
 
 /*
@@ -69,7 +72,8 @@ typedef struct _VipsGaussmat {
 	double min_ampl;
 
 	gboolean separable;
-	gboolean integer;
+	gboolean integer;		/* Deprecated */
+	VipsPrecision precision; 
 
 } VipsGaussmat;
 
@@ -96,6 +100,13 @@ vips_gaussmat_build( VipsObject *object )
 	if( VIPS_OBJECT_CLASS( vips_gaussmat_parent_class )->build( object ) )
 		return( -1 );
 
+	if( vips_object_argument_isset( object, "integer" ) ) 
+		vips_warn( class->nickname, 
+			"'integer' is deprecated, use 'precision' instead" );
+	if( vips_check_precision_intfloat( class->nickname, 
+		gaussmat->precision ) )
+		return( -1 ); 
+
 	/* Find the size of the mask. Limit the mask size to 10k x 10k for 
 	 * sanity. 
 	 */
@@ -129,7 +140,7 @@ vips_gaussmat_build( VipsObject *object )
 			double distance = xo * xo + yo * yo;
 			double v = exp( -distance / sig2 );
 
-			if( gaussmat->integer )
+			if( gaussmat->precision == VIPS_PRECISION_INTEGER )
 				v = VIPS_RINT( 20 * v );
 
 			*VIPS_MATRIX( create->out, x, y ) = v;
@@ -180,10 +191,17 @@ vips_gaussmat_class_init( VipsGaussmatClass *class )
 	VIPS_ARG_BOOL( class, "integer", 5, 
 		_( "Integer" ), 
 		_( "Generate integer Gaussian" ),
-		VIPS_ARGUMENT_OPTIONAL_INPUT,
+		VIPS_ARGUMENT_OPTIONAL_INPUT | VIPS_ARGUMENT_DEPRECATED,
 		G_STRUCT_OFFSET( VipsGaussmat, integer ),
 		FALSE );
 
+	VIPS_ARG_ENUM( class, "precision", 6, 
+		_( "Precision" ), 
+		_( "Generate with this precision" ),
+		VIPS_ARGUMENT_OPTIONAL_INPUT, 
+		G_STRUCT_OFFSET( VipsGaussmat, precision ), 
+		VIPS_TYPE_PRECISION, VIPS_PRECISION_INTEGER ); 
+
 }
 
 static void
@@ -191,6 +209,7 @@ vips_gaussmat_init( VipsGaussmat *gaussmat )
 {
 	gaussmat->sigma = 1;
 	gaussmat->min_ampl = 0.1;
+	gaussmat->precision = VIPS_PRECISION_INTEGER;
 }
 
 /**
@@ -203,7 +222,7 @@ vips_gaussmat_init( VipsGaussmat *gaussmat )
  * Optional arguments:
  *
  * @separable: generate a separable gaussian
- * @integer: generate an integer gaussian
+ * @precision: #VipsPrecision for @out
  *
  * Creates a circularly symmetric Gaussian image of radius 
  * @sigma.  The size of the mask is determined by the variable @min_ampl; 
@@ -215,13 +234,13 @@ vips_gaussmat_init( VipsGaussmat *gaussmat )
  *   H(r) = exp( -(r * r) / (2 * @sigma * @sigma) )
  *
  * The generated image has odd size and its maximum value is normalised to
- * 1.0, unless @integer is set.
+ * 1.0, unless @precision is #VIPS_PRECISION_INTEGER.
  *
  * If @separable is set, only the centre horizontal is generated. This is
  * useful for separable convolutions. 
  *
- * If @integer is set, an integer gaussian is generated. This is useful for
- * integer convolutions. 
+ * If @precision is #VIPS_PRECISION_INTEGER, an integer gaussian is generated. 
+ * This is useful for integer convolutions. 
  *
  * "scale" is set to the sum of all the mask elements.
  *
diff --git a/libvips/create/logmat.c b/libvips/create/logmat.c
index 09bcb6ef..087060c8 100644
--- a/libvips/create/logmat.c
+++ b/libvips/create/logmat.c
@@ -14,6 +14,9 @@
  * 	- gtkdoc
  * 20/10/13
  * 	- redone as a class from logmat.c
+ * 16/12/14
+ * 	- default to int output to match vips_conv()
+ * 	- use @precision, not @integer
  */
 
 /*
@@ -68,7 +71,8 @@ typedef struct _VipsLogmat {
 	double min_ampl;
 
 	gboolean separable;
-	gboolean integer;
+	gboolean integer;		/* Deprecated */
+	VipsPrecision precision; 
 
 } VipsLogmat;
 
@@ -95,6 +99,13 @@ vips_logmat_build( VipsObject *object )
 	if( VIPS_OBJECT_CLASS( vips_logmat_parent_class )->build( object ) )
 		return( -1 );
 
+	if( vips_object_argument_isset( object, "integer" ) ) 
+		vips_warn( class->nickname, 
+			"'integer' is deprecated, use 'precision' instead" );
+	if( vips_check_precision_intfloat( class->nickname, 
+		logmat->precision ) )
+		return( -1 ); 
+
 	/* Find the size of the mask. We want to eval the mask out to the 
 	 * flat zero part, ie. beyond the minimum and to the point where it 
 	 * comes back up towards zero.
@@ -153,7 +164,7 @@ vips_logmat_build( VipsObject *object )
 				(2.0 - (distance / sig2)) *
 				exp( -distance / (2.0 * sig2) );
 
-			if( logmat->integer )
+			if( logmat->precision == VIPS_PRECISION_INTEGER )
 				v = VIPS_RINT( 20 * v );
 
 			*VIPS_MATRIX( create->out, x, y ) = v;
@@ -204,10 +215,17 @@ vips_logmat_class_init( VipsLogmatClass *class )
 	VIPS_ARG_BOOL( class, "integer", 5, 
 		_( "Integer" ), 
 		_( "Generate integer Logmatian" ),
-		VIPS_ARGUMENT_OPTIONAL_INPUT,
+		VIPS_ARGUMENT_OPTIONAL_INPUT | VIPS_ARGUMENT_DEPRECATED,
 		G_STRUCT_OFFSET( VipsLogmat, integer ),
 		FALSE );
 
+	VIPS_ARG_ENUM( class, "precision", 6, 
+		_( "Precision" ), 
+		_( "Generate with this precision" ),
+		VIPS_ARGUMENT_OPTIONAL_INPUT, 
+		G_STRUCT_OFFSET( VipsLogmat, precision ), 
+		VIPS_TYPE_PRECISION, VIPS_PRECISION_INTEGER ); 
+
 }
 
 static void
@@ -215,6 +233,7 @@ vips_logmat_init( VipsLogmat *logmat )
 {
 	logmat->sigma = 1;
 	logmat->min_ampl = 0.1;
+	logmat->precision = VIPS_PRECISION_INTEGER;
 }
 
 /**
@@ -226,8 +245,8 @@ vips_logmat_init( VipsLogmat *logmat )
  *
  * Optional arguments:
  *
- * @separable: generate a separable logmatian
- * @integer: generate an integer logmatian
+ * @separable: generate a separable mask
+ * @precision: #VipsPrecision for @out
  *
  * Creates a circularly symmetric Laplacian of Gaussian mask 
  * of radius 
@@ -246,13 +265,13 @@ vips_logmat_init( VipsLogmat *logmat )
  * where s2 = @sigma * @sigma, s4 = s2 * s2, r2 = r * r.  
  *
  * The generated mask has odd size and its maximum value is normalised to 
- * 1.0, unless @integer is set.
+ * 1.0, unless @precision is #VIPS_PRECISION_INTEGER.
  *
  * If @separable is set, only the centre horizontal is generated. This is
  * useful for separable convolutions. 
  *
- * If @integer is set, an integer logmatian is generated. This is useful for
- * integer convolutions. 
+ * If @precision is #VIPS_PRECISION_INTEGER, an integer mask is generated. 
+ * This is useful for integer convolutions. 
  *
  * "scale" is set to the sum of all the mask elements.
  *
diff --git a/libvips/deprecated/vips7compat.c b/libvips/deprecated/vips7compat.c
index 4e432db0..75ec4ad8 100644
--- a/libvips/deprecated/vips7compat.c
+++ b/libvips/deprecated/vips7compat.c
@@ -2169,6 +2169,7 @@ im_gauss_dmask( const char *filename, double sigma, double min_ampl )
 	DOUBLEMASK *msk;
 
 	if( vips_gaussmat( &t, sigma, min_ampl,
+		"precision", VIPS_PRECISION_FLOAT,
 		NULL ) )
 		return( NULL );
 	if( !(msk = im_vips2mask( t, filename )) ) {
@@ -2187,6 +2188,7 @@ im_gauss_dmask_sep( const char *filename, double sigma, double min_ampl )
 	DOUBLEMASK *msk;
 
 	if( vips_gaussmat( &t, sigma, min_ampl,
+		"precision", VIPS_PRECISION_FLOAT,
 		"separable", TRUE,
 		NULL ) )
 		return( NULL );
@@ -2205,9 +2207,7 @@ im_gauss_imask( const char *filename, double sigma, double min_ampl )
 	VipsImage *t;
 	INTMASK *msk;
 
-	if( vips_gaussmat( &t, sigma, min_ampl,
-		"integer", TRUE,
-		NULL ) )
+	if( vips_gaussmat( &t, sigma, min_ampl, NULL ) )
 		return( NULL );
 	if( !(msk = im_vips2imask( t, filename )) ) {
 		g_object_unref( t );
@@ -2225,7 +2225,6 @@ im_gauss_imask_sep( const char *filename, double sigma, double min_ampl )
 	INTMASK *msk;
 
 	if( vips_gaussmat( &t, sigma, min_ampl,
-		"integer", TRUE,
 		"separable", TRUE,
 		NULL ) )
 		return( NULL );
@@ -2244,9 +2243,7 @@ im_log_imask( const char *filename, double sigma, double min_ampl )
 	VipsImage *t;
 	INTMASK *msk;
 
-	if( vips_logmat( &t, sigma, min_ampl,
-		"integer", TRUE,
-		NULL ) )
+	if( vips_logmat( &t, sigma, min_ampl, NULL ) )
 		return( NULL );
 	if( !(msk = im_vips2imask( t, filename )) ) {
 		g_object_unref( t );
@@ -2264,6 +2261,7 @@ im_log_dmask( const char *filename, double sigma, double min_ampl )
 	DOUBLEMASK *msk;
 
 	if( vips_logmat( &t, sigma, min_ampl,
+		"precision", VIPS_PRECISION_FLOAT,
 		NULL ) )
 		return( NULL );
 	if( !(msk = im_vips2mask( t, filename )) ) {
diff --git a/libvips/include/vips/Makefile.am b/libvips/include/vips/Makefile.am
index 0f264c6a..ddaf6cc2 100644
--- a/libvips/include/vips/Makefile.am
+++ b/libvips/include/vips/Makefile.am
@@ -69,6 +69,7 @@ vips_scan_headers = \
 	${top_srcdir}/libvips/include/vips/convolution.h \
 	${top_srcdir}/libvips/include/vips/morphology.h \
 	${top_srcdir}/libvips/include/vips/draw.h \
+	${top_srcdir}/libvips/include/vips/basic.h \
 	${top_srcdir}/libvips/include/vips/object.h
 
 enumtypes.h: $(vips_scan_headers) Makefile
diff --git a/libvips/include/vips/basic.h b/libvips/include/vips/basic.h
index 633f4f31..8f846531 100644
--- a/libvips/include/vips/basic.h
+++ b/libvips/include/vips/basic.h
@@ -59,6 +59,13 @@ typedef void *(*VipsSListMap4Fn)( void *item,
 typedef void *(*VipsSListFold2Fn)( void *item, 
 	void *a, void *b, void *c );
 
+typedef enum {
+	VIPS_PRECISION_INTEGER,
+	VIPS_PRECISION_FLOAT,
+	VIPS_PRECISION_APPROXIMATE,
+	VIPS_PRECISION_LAST
+} VipsPrecision;
+
 #ifdef __cplusplus
 }
 #endif /*__cplusplus*/
diff --git a/libvips/include/vips/convolution.h b/libvips/include/vips/convolution.h
index f2685542..1240301e 100644
--- a/libvips/include/vips/convolution.h
+++ b/libvips/include/vips/convolution.h
@@ -38,13 +38,6 @@
 extern "C" {
 #endif /*__cplusplus*/
 
-typedef enum {
-	VIPS_PRECISION_INTEGER,
-	VIPS_PRECISION_FLOAT,
-	VIPS_PRECISION_APPROXIMATE,
-	VIPS_PRECISION_LAST
-} VipsPrecision;
-
 typedef enum {
 	VIPS_COMBINE_MAX,
 	VIPS_COMBINE_SUM,
diff --git a/libvips/include/vips/enumtypes.h b/libvips/include/vips/enumtypes.h
index 251d387f..37380972 100644
--- a/libvips/include/vips/enumtypes.h
+++ b/libvips/include/vips/enumtypes.h
@@ -78,8 +78,6 @@ GType vips_pcs_get_type (void) G_GNUC_CONST;
 GType vips_operation_flags_get_type (void) G_GNUC_CONST;
 #define VIPS_TYPE_OPERATION_FLAGS (vips_operation_flags_get_type())
 /* enumerations from "../../../libvips/include/vips/convolution.h" */
-GType vips_precision_get_type (void) G_GNUC_CONST;
-#define VIPS_TYPE_PRECISION (vips_precision_get_type())
 GType vips_combine_get_type (void) G_GNUC_CONST;
 #define VIPS_TYPE_COMBINE (vips_combine_get_type())
 /* enumerations from "../../../libvips/include/vips/morphology.h" */
@@ -88,6 +86,9 @@ GType vips_operation_morphology_get_type (void) G_GNUC_CONST;
 /* enumerations from "../../../libvips/include/vips/draw.h" */
 GType vips_combine_mode_get_type (void) G_GNUC_CONST;
 #define VIPS_TYPE_COMBINE_MODE (vips_combine_mode_get_type())
+/* enumerations from "../../../libvips/include/vips/basic.h" */
+GType vips_precision_get_type (void) G_GNUC_CONST;
+#define VIPS_TYPE_PRECISION (vips_precision_get_type())
 /* enumerations from "../../../libvips/include/vips/object.h" */
 GType vips_argument_flags_get_type (void) G_GNUC_CONST;
 #define VIPS_TYPE_ARGUMENT_FLAGS (vips_argument_flags_get_type())
diff --git a/libvips/include/vips/error.h b/libvips/include/vips/error.h
index d445eb0c..d7699e9d 100644
--- a/libvips/include/vips/error.h
+++ b/libvips/include/vips/error.h
@@ -73,6 +73,7 @@ int vips_check_bands_1orn( const char *domain, VipsImage *im1, VipsImage *im2 );
 int vips_check_bands_1orn_unary( const char *domain, VipsImage *im, int n );
 int vips_check_bands_same( const char *domain, VipsImage *im1, VipsImage *im2 );
 int vips_check_bandno( const char *domain, VipsImage *im, int bandno );
+
 int vips_check_int( const char *domain, VipsImage *im );
 int vips_check_uint( const char *domain, VipsImage *im );
 int vips_check_uintorf( const char *domain, VipsImage *im );
@@ -91,6 +92,9 @@ int vips_check_hist( const char *domain, VipsImage *im );
 int vips_check_matrix( const char *domain, VipsImage *im, VipsImage **out );
 int vips_check_separable( const char *domain, VipsImage *im );
 
+int vips_check_precision_intfloat( const char *domain, 
+	VipsPrecision precision );
+
 #ifdef __cplusplus
 }
 #endif /*__cplusplus*/
diff --git a/libvips/iofuncs/Makefile.am b/libvips/iofuncs/Makefile.am
index 9cc0c7bf..d0d32998 100644
--- a/libvips/iofuncs/Makefile.am
+++ b/libvips/iofuncs/Makefile.am
@@ -66,6 +66,7 @@ vips_scan_headers = \
 	${top_srcdir}/libvips/include/vips/convolution.h \
 	${top_srcdir}/libvips/include/vips/morphology.h \
 	${top_srcdir}/libvips/include/vips/draw.h \
+	${top_srcdir}/libvips/include/vips/basic.h \
 	${top_srcdir}/libvips/include/vips/object.h
 
 enumtypes.c: $(vips_scan_headers) Makefile
diff --git a/libvips/iofuncs/enumtypes.c b/libvips/iofuncs/enumtypes.c
index e182bfc8..540f49c4 100644
--- a/libvips/iofuncs/enumtypes.c
+++ b/libvips/iofuncs/enumtypes.c
@@ -674,25 +674,6 @@ vips_operation_flags_get_type( void )
 }
 /* enumerations from "../../libvips/include/vips/convolution.h" */
 GType
-vips_precision_get_type( void )
-{
-	static GType etype = 0;
-
-	if( etype == 0 ) {
-		static const GEnumValue values[] = {
-			{VIPS_PRECISION_INTEGER, "VIPS_PRECISION_INTEGER", "integer"},
-			{VIPS_PRECISION_FLOAT, "VIPS_PRECISION_FLOAT", "float"},
-			{VIPS_PRECISION_APPROXIMATE, "VIPS_PRECISION_APPROXIMATE", "approximate"},
-			{VIPS_PRECISION_LAST, "VIPS_PRECISION_LAST", "last"},
-			{0, NULL, NULL}
-		};
-		
-		etype = g_enum_register_static( "VipsPrecision", values );
-	}
-
-	return( etype );
-}
-GType
 vips_combine_get_type( void )
 {
 	static GType etype = 0;
@@ -748,6 +729,26 @@ vips_combine_mode_get_type( void )
 
 	return( etype );
 }
+/* enumerations from "../../libvips/include/vips/basic.h" */
+GType
+vips_precision_get_type( void )
+{
+	static GType etype = 0;
+
+	if( etype == 0 ) {
+		static const GEnumValue values[] = {
+			{VIPS_PRECISION_INTEGER, "VIPS_PRECISION_INTEGER", "integer"},
+			{VIPS_PRECISION_FLOAT, "VIPS_PRECISION_FLOAT", "float"},
+			{VIPS_PRECISION_APPROXIMATE, "VIPS_PRECISION_APPROXIMATE", "approximate"},
+			{VIPS_PRECISION_LAST, "VIPS_PRECISION_LAST", "last"},
+			{0, NULL, NULL}
+		};
+		
+		etype = g_enum_register_static( "VipsPrecision", values );
+	}
+
+	return( etype );
+}
 /* enumerations from "../../libvips/include/vips/object.h" */
 GType
 vips_argument_flags_get_type( void )
diff --git a/libvips/iofuncs/error.c b/libvips/iofuncs/error.c
index 6df5f743..15e879ef 100644
--- a/libvips/iofuncs/error.c
+++ b/libvips/iofuncs/error.c
@@ -1317,3 +1317,29 @@ vips_check_separable( const char *domain, VipsImage *im )
 
 	return( 0 );
 }
+
+/**
+ * vips_check_precision_intfloat:
+ * @domain: the originating domain for the error message
+ * @precision: precision to check
+ *
+ * Check that @prec image is either float or int.
+ * If not, set an error message
+ * and return non-zero.
+ *
+ * See also: vips_error().
+ *
+ * Returns: 0 on OK, or -1 on error.
+ */
+int
+vips_check_precision_intfloat( const char *domain, VipsPrecision precision )
+{
+	if( precision != VIPS_PRECISION_INTEGER &&
+		precision != VIPS_PRECISION_FLOAT ) {
+		vips_error( domain,
+			"%s", _( "precision must be int or float" ) );
+		return( -1 );
+	}
+
+	return( 0 );
+}