|
|
Am 28.04.2010 19:00, schrieb Thorsten Froehlich:
> On 27.04.10 23:05, clipka wrote:
>> BTW, as a matter of fact, the wrong channel ordering is a 3.7 issue;
>> POV-Ray 3.6 at least did it the right way round.
>>
>> I consider the use of RGB instead of YCbCr a bug (already filed as
>> FS#103 ;-)).
>
> No, it isn't a bug. This *really* is intentional behavior! With the
> correct settings in other places (not sure if they are), you can get rid
> of a lot of color bleeding when using sharply ray-traced objects by
> using RGB.
From experiments, I really don't see any advantage of RGB over YCbCr
without chroma sub-sampling.
To the contrary: At the same output file size, YCbCr without chroma
sub-sampling appears to give slightly superior quality.
So given that the same (or even better) quality/size tradeoff as RGB
seems to be obtainable in a fully JFIF-compatible way, that's the way to
go if I'm asked.
Even if for some reason we would want to do non-JFIF-compatible JPEG
output, I strongly advocate doing so only when the user /explicitly/
chooses it via some option. The image file format commonly known as
"JPEG" /is/ actually JFIF, so by default POV-Ray should follow that
convention for the sake of compatibility.
Note that if quality is paramount, JPEG is a bad choice of output file
format anyway, so I don't see much of a point in sacrificing
compatibility for quality with this format.
Post a reply to this message
|
|
|
|
On 28.04.10 20:10, clipka wrote:
> Am 28.04.2010 19:00, schrieb Thorsten Froehlich:
>> On 27.04.10 23:05, clipka wrote:
>>> BTW, as a matter of fact, the wrong channel ordering is a 3.7 issue;
>>> POV-Ray 3.6 at least did it the right way round.
>>>
>>> I consider the use of RGB instead of YCbCr a bug (already filed as
>>> FS#103 ;-)).
>>
>> No, it isn't a bug. This *really* is intentional behavior! With the
>> correct settings in other places (not sure if they are), you can get rid
>> of a lot of color bleeding when using sharply ray-traced objects by
>> using RGB.
>
> From experiments, I really don't see any advantage of RGB over YCbCr
> without chroma sub-sampling.
The way to measure is using statistical tests designed for image comparison
... I still have the code somewhere.
> To the contrary: At the same output file size, YCbCr without chroma
> sub-sampling appears to give slightly superior quality.
Indeed, the file size of the RGB image will be larger because quantization
works differently. In essence RGB works like three grayscale images in JPEG.
Thorsten
Post a reply to this message
|
|