| Message ID | 20230210225622.24411-3-laurent.pinchart+renesas@ideasonboard.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | media: i2c: adv7604: Fix handling of video adjustments | expand |
Hello! On 2/11/23 1:56 AM, Laurent Pinchart wrote: > The ADV7604, ADV7611 and ADV7612 software manuals different the CP_HUE s/different/document/? > value differently: > > - For ADV7604 and ADV7611, the hue is specified as an unsigned 8-bit > value, and calculated as > > (CP_HUE[7:0] * 180) / 256 - 90 > > with the range set to [-90°, 90°]. Additionally, the values 0x00, 0x0f > and 0xff are documented as corresponding to -90°, 0° and 90° > respectively. > > - For ADV7612, the hue is specified as a signed 8-bit value in the range > [0°, 360°[ without any formula. Additionally, the value 0x00 is > documented as corresponding to 0°. > > The ADV7604 and ADV7611 documentation is clearly wrong as the example > values don't correspond to the formula. Furthermore, the [-90°, 90°] > range seems incorrect as it would cover only half of the hue space. > > The ADV7612 documentation is better, although the range should likely be > [-180°, 180°[ if the value is signed. Given that the values wrap around, > this makes no difference in practice. > > The hue values have been verified on ADV7612 to follow the > documentation. There is a high chance they do as well on ADV7604 and > ADV7611. > > In any case, all software manuals document the register as 8-bit, so the > current range of the V4L2 hue control [0, 128] is not correct. Expand it > to cover the full 8-bit space, using unsigned values to avoid breaking > any application that may rely on 128 being a valid value. > > Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com> [...] MBR, Sergey
Hi Sergei, On Sat, Feb 11, 2023 at 12:11:17PM +0300, Sergei Shtylyov wrote: > Hello! > > On 2/11/23 1:56 AM, Laurent Pinchart wrote: > > > The ADV7604, ADV7611 and ADV7612 software manuals different the CP_HUE > > s/different/document/? Indeed. I'll fix this in v2, but will wait for Hans to test the patches on ADV7604 and/or ADV7611 before sending the new version. > > value differently: > > > > - For ADV7604 and ADV7611, the hue is specified as an unsigned 8-bit > > value, and calculated as > > > > (CP_HUE[7:0] * 180) / 256 - 90 > > > > with the range set to [-90°, 90°]. Additionally, the values 0x00, 0x0f > > and 0xff are documented as corresponding to -90°, 0° and 90° > > respectively. > > > > - For ADV7612, the hue is specified as a signed 8-bit value in the range > > [0°, 360°[ without any formula. Additionally, the value 0x00 is > > documented as corresponding to 0°. > > > > The ADV7604 and ADV7611 documentation is clearly wrong as the example > > values don't correspond to the formula. Furthermore, the [-90°, 90°] > > range seems incorrect as it would cover only half of the hue space. > > > > The ADV7612 documentation is better, although the range should likely be > > [-180°, 180°[ if the value is signed. Given that the values wrap around, > > this makes no difference in practice. > > > > The hue values have been verified on ADV7612 to follow the > > documentation. There is a high chance they do as well on ADV7604 and > > ADV7611. > > > > In any case, all software manuals document the register as 8-bit, so the > > current range of the V4L2 hue control [0, 128] is not correct. Expand it > > to cover the full 8-bit space, using unsigned values to avoid breaking > > any application that may rely on 128 being a valid value. > > > > Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com> > [...] > > MBR, Sergey
On 10/02/2023 23:56, Laurent Pinchart wrote: > The ADV7604, ADV7611 and ADV7612 software manuals different the CP_HUE > value differently: > > - For ADV7604 and ADV7611, the hue is specified as an unsigned 8-bit > value, and calculated as > > (CP_HUE[7:0] * 180) / 256 - 90 > > with the range set to [-90°, 90°]. Additionally, the values 0x00, 0x0f > and 0xff are documented as corresponding to -90°, 0° and 90° > respectively. > > - For ADV7612, the hue is specified as a signed 8-bit value in the range > [0°, 360°[ without any formula. Additionally, the value 0x00 is > documented as corresponding to 0°. > > The ADV7604 and ADV7611 documentation is clearly wrong as the example > values don't correspond to the formula. Furthermore, the [-90°, 90°] > range seems incorrect as it would cover only half of the hue space. > > The ADV7612 documentation is better, although the range should likely be > [-180°, 180°[ if the value is signed. Given that the values wrap around, > this makes no difference in practice. > > The hue values have been verified on ADV7612 to follow the > documentation. There is a high chance they do as well on ADV7604 and > ADV7611. > > In any case, all software manuals document the register as 8-bit, so the > current range of the V4L2 hue control [0, 128] is not correct. Expand it > to cover the full 8-bit space, using unsigned values to avoid breaking > any application that may rely on 128 being a valid value. > > Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com> > --- > drivers/media/i2c/adv7604.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/media/i2c/adv7604.c b/drivers/media/i2c/adv7604.c > index 3af0e67f9edb..3d0898c4175e 100644 > --- a/drivers/media/i2c/adv7604.c > +++ b/drivers/media/i2c/adv7604.c > @@ -3548,7 +3548,7 @@ static int adv76xx_probe(struct i2c_client *client) > v4l2_ctrl_new_std(hdl, &adv76xx_ctrl_ops, > V4L2_CID_SATURATION, 0, 255, 1, 128); > v4l2_ctrl_new_std(hdl, &adv76xx_ctrl_ops, > - V4L2_CID_HUE, 0, 128, 1, 0); > + V4L2_CID_HUE, 0, 255, 1, 0); > ctrl = v4l2_ctrl_new_std_menu(hdl, &adv76xx_ctrl_ops, > V4L2_CID_DV_RX_IT_CONTENT_TYPE, V4L2_DV_IT_CONTENT_TYPE_NO_ITC, > 0, V4L2_DV_IT_CONTENT_TYPE_NO_ITC); Tested with the adv7604. Reviewed-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Tested-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Thank you for looking at this! Hans
diff --git a/drivers/media/i2c/adv7604.c b/drivers/media/i2c/adv7604.c index 3af0e67f9edb..3d0898c4175e 100644 --- a/drivers/media/i2c/adv7604.c +++ b/drivers/media/i2c/adv7604.c @@ -3548,7 +3548,7 @@ static int adv76xx_probe(struct i2c_client *client) v4l2_ctrl_new_std(hdl, &adv76xx_ctrl_ops, V4L2_CID_SATURATION, 0, 255, 1, 128); v4l2_ctrl_new_std(hdl, &adv76xx_ctrl_ops, - V4L2_CID_HUE, 0, 128, 1, 0); + V4L2_CID_HUE, 0, 255, 1, 0); ctrl = v4l2_ctrl_new_std_menu(hdl, &adv76xx_ctrl_ops, V4L2_CID_DV_RX_IT_CONTENT_TYPE, V4L2_DV_IT_CONTENT_TYPE_NO_ITC, 0, V4L2_DV_IT_CONTENT_TYPE_NO_ITC);
The ADV7604, ADV7611 and ADV7612 software manuals different the CP_HUE value differently: - For ADV7604 and ADV7611, the hue is specified as an unsigned 8-bit value, and calculated as (CP_HUE[7:0] * 180) / 256 - 90 with the range set to [-90°, 90°]. Additionally, the values 0x00, 0x0f and 0xff are documented as corresponding to -90°, 0° and 90° respectively. - For ADV7612, the hue is specified as a signed 8-bit value in the range [0°, 360°[ without any formula. Additionally, the value 0x00 is documented as corresponding to 0°. The ADV7604 and ADV7611 documentation is clearly wrong as the example values don't correspond to the formula. Furthermore, the [-90°, 90°] range seems incorrect as it would cover only half of the hue space. The ADV7612 documentation is better, although the range should likely be [-180°, 180°[ if the value is signed. Given that the values wrap around, this makes no difference in practice. The hue values have been verified on ADV7612 to follow the documentation. There is a high chance they do as well on ADV7604 and ADV7611. In any case, all software manuals document the register as 8-bit, so the current range of the V4L2 hue control [0, 128] is not correct. Expand it to cover the full 8-bit space, using unsigned values to avoid breaking any application that may rely on 128 being a valid value. Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com> --- drivers/media/i2c/adv7604.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)