Sorry, the information that the picture has acquired blue shades is false!

вт, 19 мая 2020 г. в 11:53, Виктор Мулин <[email protected]>:

> I tried the following before opening the encoder:
> codec_context-> intra_matrix = (uint16_t *) std_luminance_quant_tbl;
> codec_context-> croma_intra_matrix = (uint16_t *) std_crominance_quant_tbl;
> The picture has acquired a blue tint. When packet was written to a file,
> DQT tables were not added to the file.
> Questions remain open:
> 1) How to force a codec to use two standard DQT tables (JPEG ISO / IEC
> 10918-1 ITU-T recomendation T.81 Annex K.1 and K.2)
> 2) How to make the codec add both DQT tables to paccket->data, then save
> packet->data to a jpg file?
>
> пн, 18 мая 2020 г. в 23:16, Richard Hussong <[email protected]>:
>
>> Sorry, but I've never modified those values myself and I don't know how
>> they work. I think you'll have to read the source code for the JPEG encoder
>> to find out.
>>
>> - Richard
>>
>> On Mon, May 18, 2020 at 1:37 PM Виктор Мулин <[email protected]> wrote:
>>
>>> Thanks for the answer.
>>> intra_matrix and inter_matrix are of type uint16_t, and tables from
>>> jpegtables.c are of type uint8_t. What to do? Make a cast of uint8_t * to
>>> uint16_t * or define a new array of type uint16_t by filling it with values
>>> from standard tables?
>>> Will these tables be added to the output file automatically after I
>>> specify array references in the AVCodecContext structure?
>>>
>>> пн, 18 мая 2020 г. в 19:54, Richard Hussong <[email protected]>:
>>>
>>>>
>>>>
>>>> On Mon, May 18, 2020 at 7:38 AM Виктор Мулин <[email protected]> wrote:
>>>>
>>>>> Hello! When using the mjpeg codec to convert an image from bmp to jpg
>>>>> format, a picture with one quantization table is created. I need two
>>>>> standard tables described in the jpeg iso / iec 10918-1 T.81 standard in
>>>>> Appendix K.1 and K.2 to be added to the output image file. I also found
>>>>> them in the ffmpeg source codes (jpegtables.c -> std_luminance_quant_tbl
>>>>> and jpegtables.c -> std_chrominance_quant_tbl [url:
>>>>> https://ffmpeg.org/doxygen/trunk/jpegtables_8c_source.html]. I looked
>>>>> through the mailing list archive and found old solutions (
>>>>> https://ffmpeg.org/pipermail/libav-user/2015-July/008318.html), but
>>>>> now the av_codec_set_chroma_intra_matrix () function is declared
>>>>> depricated. How do I get what I want? How to set intra_matrix,
>>>>> chroma_intra_matrix and inter_matrix (in the description they are almost
>>>>> the same)?
>>>>>
>>>>
>>>> The intra_matrix and inter_matrix can now be set directly in the
>>>> AVCodecContext structure. See
>>>> https://ffmpeg.org/doxygen/trunk/structAVCodecContext.html, and search
>>>> for "intra_matrix" and "inter_matrix".
>>>>
>>>> P.S .: Sorry, I don't speak English well. I use google translator.
>>>>>
>>>>
>>>> Don't worry - what you wrote is clear enough.
>>>>
>>>>> _______________________________________________
>>>>> Libav-user mailing list
>>>>> [email protected]
>>>>> https://ffmpeg.org/mailman/listinfo/libav-user
>>>>>
>>>>> To unsubscribe, visit link above, or email
>>>>> [email protected] with subject "unsubscribe".
>>>>
>>>> _______________________________________________
>>>> Libav-user mailing list
>>>> [email protected]
>>>> https://ffmpeg.org/mailman/listinfo/libav-user
>>>>
>>>> To unsubscribe, visit link above, or email
>>>> [email protected] with subject "unsubscribe".
>>>
>>> _______________________________________________
>>> Libav-user mailing list
>>> [email protected]
>>> https://ffmpeg.org/mailman/listinfo/libav-user
>>>
>>> To unsubscribe, visit link above, or email
>>> [email protected] with subject "unsubscribe".
>>
>> _______________________________________________
>> Libav-user mailing list
>> [email protected]
>> https://ffmpeg.org/mailman/listinfo/libav-user
>>
>> To unsubscribe, visit link above, or email
>> [email protected] with subject "unsubscribe".
>
>
_______________________________________________
Libav-user mailing list
[email protected]
https://ffmpeg.org/mailman/listinfo/libav-user

To unsubscribe, visit link above, or email
[email protected] with subject "unsubscribe".

Reply via email to