|
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Le 14-11-29 07:18, Thomas de Groot a écrit :
> On 28-11-2014 13:24, clipka wrote:
>> A) 16-bit png images with alpha channel.
>>
>
> Which reminds me: Using IC, how can one convert an 8-bit png into a
> 16-bit png? I suppose the image needs to be resampled but with what
> settings?
>
> Thomas
>
>
Using most image procesing tools supporting saving PNG with multiple bit
per channel:
Load the original image whitch can be in any format, not just PNG.
Save it as a 16 bit per channel PNG file.
The only thing been that the lower 8 bits will be patched with zeros.
This may be aleviated by enlarging the image and bluring the enlarged
version. This may force some interpolation between the original pixels.
Far from perfect, but beter than nothing.
If you can render the image in POV-Ray, add +fn16 to the command line to
create the image as a 16 bit per channel PNG.
You may try rendering the original image as an image_map of at least
twice the pixel size of the original and use one of the interpolation
methods.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Am 29.11.2014 13:18, schrieb Thomas de Groot:
> On 28-11-2014 13:24, clipka wrote:
>> A) 16-bit png images with alpha channel.
>>
>
> Which reminds me: Using IC, how can one convert an 8-bit png into a
> 16-bit png? I suppose the image needs to be resampled but with what
> settings?
>
> Thomas
>
>
image->mode->16bps
then save as PNG
this makes only sense if you are going to resize, tonemap, filter, do
*anything* with the image afterwards.
-Ive
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Am 30.11.2014 01:44, schrieb Alain:
> The only thing been that the lower 8 bits will be patched with zeros.
err, no, any decent image processing tool will multiply by 257 instead a
simple left-shift.
-Ive
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On 30-11-2014 11:42, Ive wrote:
> Am 29.11.2014 13:18, schrieb Thomas de Groot:
>> On 28-11-2014 13:24, clipka wrote:
>>> A) 16-bit png images with alpha channel.
>>>
>>
>> Which reminds me: Using IC, how can one convert an 8-bit png into a
>> 16-bit png? I suppose the image needs to be resampled but with what
>> settings?
>>
>> Thomas
>>
>>
>
> image->mode->16bps
>
> then save as PNG
>
> this makes only sense if you are going to resize, tonemap, filter, do
> *anything* with the image afterwards.
>
> -Ive
Yes! You confirm what I suspected already (but was not sure). With only
using mode->16bps, nothing happens. Thanks indeed.
Thomas
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Le 14-11-30 05:47, Ive a écrit :
> Am 30.11.2014 01:44, schrieb Alain:
>> The only thing been that the lower 8 bits will be patched with zeros.
>
> err, no, any decent image processing tool will multiply by 257 instead a
> simple left-shift.
>
> -Ive
>
That 8 bits left shift or multiply by 256 is exactly the same as left
padding with zeros.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Am 30.11.2014 19:09, schrieb Alain:
> Le 14-11-30 05:47, Ive a écrit :
>> Am 30.11.2014 01:44, schrieb Alain:
>>> The only thing been that the lower 8 bits will be patched with zeros.
>>
>> err, no, any decent image processing tool will multiply by 257 instead a
>> simple left-shift.
>>
>> -Ive
>>
> That 8 bits left shift or multiply by 256 is exactly the same as left
> padding with zeros.
257 != 256
-Ive
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |