Re: [PATCH 1/2] OMAPDSS: DISPC: Update Fir Coefficients

["Mahapatra, Chandrabhanu" <cmahapatra@ti.com>]

On Thu, Dec 15, 2011 at 2:25 PM, Tomi Valkeinen <tomi.valkeinen@ti.com> wrote:
> On Wed, 2011-12-14 at 10:21 +0530, Chandrabhanu Mahapatra wrote:
>
>> +const struct dispc_coef *dispc_ovl_get_scale_coef(int inc, int five_taps)
>> +{
>> +     int i;
>> +     static const struct {
>> +             int Mmin;
>> +             int Mmax;
>> +             const struct dispc_coef *coef_3;
>> +             const struct dispc_coef *coef_5;
>> +     } coefs[] = {
>> +             { 26, 32, coef3_M32, coef5_M32 },
>> +             { 22, 26, coef3_M26, coef5_M26 },
>> +             { 19, 22, coef3_M22, coef5_M22 },
>> +             { 16, 19, coef3_M19, coef5_M19 },
>> +             { 14, 16, coef3_M16, coef5_M16 },
>> +             { 13, 14, coef3_M14, coef5_M14 },
>> +             { 12, 13, coef3_M13, coef5_M13 },
>> +             { 11, 12, coef3_M12, coef5_M12 },
>> +             { 10, 11, coef3_M11, coef5_M11 },
>> +             {  9, 10, coef3_M10, coef5_M10 },
>> +             {  8,  9,  coef3_M9,  coef5_M9 },
>> +             {  3,  8,  coef3_M8,  coef5_M8 },
>> +             /*
>> +              * When upscaling more than two times, blockiness and outlines
>> +              * around the image are observed when M8 tables are used. M11,
>> +              * M16 and M19 tables are used to prevent this.
>> +              */
>> +             {  2,  3, coef3_M11, coef5_M11 },
>> +             {  1,  2, coef3_M16, coef5_M16 },
>> +     };
>> +
>> +     inc /= 128;
>> +     for (i = 0; i < ARRAY_LEN(coefs); ++i)
>> +             if (inc > coefs[i].Mmin && inc <= coefs[i].Mmax)
>> +                     return five_taps ? coefs[i].coef_5 : coefs[i].coef_3;
>> +     if (inc = 1)
>> +             return five_taps ? coef3_M19 : coef5_M19;
>> +     return NULL;
>> +}
>
> Why don't you handle the inc = 1 case the same as others? Just have an
> entry in the table for Mmin=0, Mmax = 1.
>
For inc=1 i.e. M=1 , scaling ratio is maximum as L/M=8. DISPC scaler
doesnot support upscaling more than 8 itmes. Having an (Mmin,Mmax] of
(0,1] will allow such cases.
> Also, I think it's a bit confusing that Mmin is exclusive and Mmax is
> inclusive in the comparison. It makes the table a bit hard to read, when
> looking at which entry is used for which inc. I'd recommend using
> inclusive comparison for both.
>
>  Tomi
>
Having both inclusive will allow us to delete the extra comparison for
inc=1 but in my opinion having Mmin exclusive and Mmax inclusive
actually gives an clear idea of comparison. The tables mostly go by
the Mmax value.
For example, for inc& coef3/5_M26 table is selected, for inc"
coef3/5_M22 is selected etc.
If we have both Mmin and Mmax as inclusive above case becomes slightly
incoherent. Say for M& instead of coef3/5_M26 which seems more
obvious choice coef3/5_M32 is selected.

For both inclusive cases to work and avoid confusion and delete extra
comparison for inc=1 , I have to reverse the order of table entries
in "coef" table. But for that I will have to put the "When upscaling
more than two times, blockiness and outlines" comment at the beginning
of the table and then start with  {  1,  2, coef3_M16, coef5_M16 }.
This will create even more confusion.


-- 
Chandrabhanu Mahapatra
Texas Instruments India Pvt. Ltd.