[2/2] ipa: rkisp1: mira220: Adjust relativeLuminanceTarget
diff mbox series

Message ID 20260911-mira220-tuning-file-v1-2-c3c588f4188a@ideasonboard.com
State New
Headers show
Series
  • ipa: rkisp1: mira220: Add tuning file
Related show

Commit Message

Jacopo Mondi Sept. 11, 2026, 7:49 a.m. UTC
Adjust relativeLuminanceTarget parameter for Mira220 as it provide
better performances under day-light environments.

Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
---
 src/ipa/rkisp1/data/mira220.yaml | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

Kieran Bingham Sept. 14, 2026, 10:21 a.m. UTC | #1
Quoting Jacopo Mondi (2026-09-11 08:49:23)
> Adjust relativeLuminanceTarget parameter for Mira220 as it provide
> better performances under day-light environments.

What are the impacts without this ? Is it overexposing?

The mira is known to have a high sensitivity, I wonder if that plays a
part here to needing to turn the AGC down ? But I'd have thought that
would balance out through the normal AGC operation...

Stefan - any thoughts ?

> 
> Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
> ---
>  src/ipa/rkisp1/data/mira220.yaml | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/src/ipa/rkisp1/data/mira220.yaml b/src/ipa/rkisp1/data/mira220.yaml
> index b45410923007..d8431966f82e 100644
> --- a/src/ipa/rkisp1/data/mira220.yaml
> +++ b/src/ipa/rkisp1/data/mira220.yaml
> @@ -29,7 +29,7 @@ algorithms:
>              qLo: 0.98
>              qHi: 1.0
>              yTarget: 0.8
> -      relativeLuminanceTarget: 0.5
> +      relativeLuminanceTarget: 0.2
>    - Lux:
>        referenceY: 0.1767075076869632
>        referenceExposureTime: 29980
> 
> -- 
> 2.55.0
>
Jacopo Mondi Sept. 25, 2026, 10:17 a.m. UTC | #2
Hi Kieran

On Mon, Sep 14, 2026 at 11:21:11AM +0100, Kieran Bingham wrote:
> Quoting Jacopo Mondi (2026-09-11 08:49:23)
> > Adjust relativeLuminanceTarget parameter for Mira220 as it provide
> > better performances under day-light environments.
>
> What are the impacts without this ? Is it overexposing?

yes, images are generally quite overexposed.

>
> The mira is known to have a high sensitivity, I wonder if that plays a
> part here to needing to turn the AGC down ? But I'd have thought that
> would balance out through the normal AGC operation...

My understanding was that the 0.5 value generated by config file is a
default. I'm not sure how this is expected to be better characterized
though..

>
> Stefan - any thoughts ?
>
> >
> > Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
> > ---
> >  src/ipa/rkisp1/data/mira220.yaml | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/src/ipa/rkisp1/data/mira220.yaml b/src/ipa/rkisp1/data/mira220.yaml
> > index b45410923007..d8431966f82e 100644
> > --- a/src/ipa/rkisp1/data/mira220.yaml
> > +++ b/src/ipa/rkisp1/data/mira220.yaml
> > @@ -29,7 +29,7 @@ algorithms:
> >              qLo: 0.98
> >              qHi: 1.0
> >              yTarget: 0.8
> > -      relativeLuminanceTarget: 0.5
> > +      relativeLuminanceTarget: 0.2
> >    - Lux:
> >        referenceY: 0.1767075076869632
> >        referenceExposureTime: 29980
> >
> > --
> > 2.55.0
> >
Barnabás Pőcze Sept. 25, 2026, 11:39 a.m. UTC | #3
2026. 09. 25. 12:17 keltezéssel, Jacopo Mondi írta:
> Hi Kieran
> 
> On Mon, Sep 14, 2026 at 11:21:11AM +0100, Kieran Bingham wrote:
>> Quoting Jacopo Mondi (2026-09-11 08:49:23)
>>> Adjust relativeLuminanceTarget parameter for Mira220 as it provide
>>> better performances under day-light environments.
>>
>> What are the impacts without this ? Is it overexposing?
> 
> yes, images are generally quite overexposed.
> 
>>
>> The mira is known to have a high sensitivity, I wonder if that plays a
>> part here to needing to turn the AGC down ? But I'd have thought that
>> would balance out through the normal AGC operation...
> 
> My understanding was that the 0.5 value generated by config file is a
> default. I'm not sure how this is expected to be better characterized
> though..

The default in the algorithm itself is 0.16, which in my limited experience
gives mostly acceptable results. So just omitting it might be an option
until tuning methodology for the parameter has been established.


> 
>>
>> Stefan - any thoughts ?
>>
>>>
>>> Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
>>> ---
>>>   src/ipa/rkisp1/data/mira220.yaml | 2 +-
>>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/src/ipa/rkisp1/data/mira220.yaml b/src/ipa/rkisp1/data/mira220.yaml
>>> index b45410923007..d8431966f82e 100644
>>> --- a/src/ipa/rkisp1/data/mira220.yaml
>>> +++ b/src/ipa/rkisp1/data/mira220.yaml
>>> @@ -29,7 +29,7 @@ algorithms:
>>>               qLo: 0.98
>>>               qHi: 1.0
>>>               yTarget: 0.8
>>> -      relativeLuminanceTarget: 0.5
>>> +      relativeLuminanceTarget: 0.2
>>>     - Lux:
>>>         referenceY: 0.1767075076869632
>>>         referenceExposureTime: 29980
>>>
>>> --
>>> 2.55.0
>>>

Patch
diff mbox series

diff --git a/src/ipa/rkisp1/data/mira220.yaml b/src/ipa/rkisp1/data/mira220.yaml
index b45410923007..d8431966f82e 100644
--- a/src/ipa/rkisp1/data/mira220.yaml
+++ b/src/ipa/rkisp1/data/mira220.yaml
@@ -29,7 +29,7 @@  algorithms:
             qLo: 0.98
             qHi: 1.0
             yTarget: 0.8
-      relativeLuminanceTarget: 0.5
+      relativeLuminanceTarget: 0.2
   - Lux:
       referenceY: 0.1767075076869632
       referenceExposureTime: 29980