[FFmpeg-user] ffprobe bug in 'interlaced_frame'?

[Mark Filipak]

On 01/26/2020 04:04 PM, Carl Eugen Hoyos wrote:
> Am So., 26. Jan. 2020 um 22:00 Uhr schrieb Mark Filipak
> <markfilipak.windows+ffmpeg at gmail.com>:
>>
>> On 01/26/2020 03:50 PM, Carl Eugen Hoyos wrote:
>>> Am So., 26. Jan. 2020 um 21:19 Uhr schrieb Mark Filipak
>>> <markfilipak.windows+ffmpeg at gmail.com>:
>>>
>>>> On 01/26/2020 03:03 PM, Carl Eugen Hoyos wrote:
>>>>> Am So., 26. Jan. 2020 um 20:51 Uhr schrieb Mark Filipak
>>>>> <markfilipak.windows+ffmpeg at gmail.com>:
>>>>>>
>>>>>> For soft telecined videos, all frames have
>>>>>> 'progressive_frame' = 1
>>>>>
>>>>> I may miss something but since FFmpeg does not "support" soft-telecine
>>>>> why should there be an interlaced frame?
>>>>
>>>> Yes, ffmpeg does not make soft telecined streams. However, the soft
>>>> telecined videos are inputs, not outputs.
>>>
>>> I don't think FFmpeg "supports" soft-telecined input streams.
>>> At least not in the way once opon a time defined in an ancient NTSC
>>> standard...
>>
>> You are confusing soft telecine with hard telecine.
> 
> I thought you are...
> (hard-telecine encoding and decoding is - of course - supported
> by FFmpeg)

No, it's not. ffmpeg makes only frame pictures, not field pictures. Hard 
telecine is an interlaced format and ffmpeg doesn't make interlaced output.

It's really moot. I'm not writing about ffmpeg. I'm writing about ffprobe.

>> Since autumn 1999, nearly 100% of all region 1 DVDs are
>> soft telecined
> 
> And such dvds are progressive if you don't watch on an
> (American) crt tv ...

Yes, the DVD VOBs all have the MPEG2 'progressive-frame' metadata bit set.

>> so, of course, ffmpeg supports soft-telecined inputs.
> 
> ... which you cannot use with FFmpeg - FFmpeg
> therefore will ignore soft-telecine (not "support" it).

I am analyzing videos using ffprobe. I'm not making videos. That 'said', 
of course ffmpeg supports soft telecine. Otherwise, it couldn't 
decode/transcode any DVDs mastered since 1999.

> Carl Eugen