Git Inbox Mirror of the ffmpeg-devel mailing list - see https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
 help / color / mirror / Atom feed
From: Allan Cady via ffmpeg-devel <ffmpeg-devel@ffmpeg.org>
To: FFmpeg development discussions and patches <ffmpeg-devel@ffmpeg.org>
Cc: Allan Cady <allancady@yahoo.com>
Subject: Re: [FFmpeg-devel] [PATCH] libavutil/timestamp.h: Fix loss of precision in timestamps for silencedetect on long files
Date: Fri, 23 Feb 2024 21:48:26 +0000 (UTC)
Message-ID: <1504257982.42894.1708724906320@mail.yahoo.com> (raw)
In-Reply-To: <0430c2a6-2e6a-ec15-24b3-409c55055dd5@passwd.hu>

[Apologies for the awful mess in the previous email. Trying again with straight text.]

On Thursday, February 22, 2024 at 01:16:19 AM PST, Marton Balint <cus@passwd.hu> wrote:

>> For starters, I'm curious why there are two functions & macros:
>>
>> av_ts2str/av_ts_make_string (which used "%" format specifier)
> 
> That takes a 64-bit integer timestamp and is actually using "%"PRId64 
> because that is the correct (portable) format string for an int64_t 
> variable.
> 
>> av_ts2timestr/av_ts_make_time_string (which used "%6g")
> 
> That takes an integer timestamp and a rational time base. Float timestamps 
> (in seconds) is calculated by multiplying the two, that is what is 
> printed.
> 
>>
>> Do you know the rationale for that? I see that only av_ts2timestr is used in silencedetect.c.
>>
>> And are you suggesting we should fold those two functions into one?
> 
> No, they have different purpose. The first prints out a timestamps which 
> can be in any time base. The second prints out a timestamp in seconds.


Understood, I think I'm up to speed now. av_ts2str prints a 64-bit integer timestamp;
av_ts2timestr prints a floating point timestamp in seconds with the timebase applied.


In your previous email, you said:


> I'd rather just to fix av_ts_make_string to not limit the number of 
> significant digits. Something like:
>
> 1) Print the number in decimal notation with at most 6 fractional digits. 
> 2) Use less fractional digits if the first format would not fit into 
> AV_TS_MAX_STRING_SIZE.
> 3) Use scientific notation if the second format would not fit into 
> AV_TS_MAX_STRINT_SIZE.


I think you probably meant to say "I'd rather just to fix
av_ts_make_time_string" (not av_ts_make_string)?
Since it's av_ts_make_time_string() that's formatting floating point.


So it makes sense to me to make the change to av_ts_make_time_string()
for all timestamps, as you suggest. As for how specifically to format them,
I'm fine with whatever you think is best, and I'm happy to work on this, but the
implementation has me a bit stumped for the moment, and I may need some
help with it. My C language skills are rusty to say the least.

It also occurs to me to wonder, would this warrant a formal problem ticket?
I haven't looked into how that works for ffmpeg.

Thanks for your help.




_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".

  parent reply	other threads:[~2024-02-23 21:48 UTC|newest]

Thread overview: 37+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <20240221073915.26055-1-allancady.ref@yahoo.com>
2024-02-21  7:39 ` Allan Cady via ffmpeg-devel
2024-02-21 20:25   ` Marton Balint
2024-02-22  1:21     ` Allan Cady via ffmpeg-devel
2024-02-22  9:16       ` Marton Balint
2024-02-23 21:44         ` Allan Cady via ffmpeg-devel
2024-02-23 21:48         ` Allan Cady via ffmpeg-devel [this message]
2024-02-23 22:47           ` Marton Balint
2024-03-11  2:56             ` [FFmpeg-devel] [PATCH] When the silencedetect filter is run against long files, the output timestamps gradually lose precision as the scan proceeds further into the file. This is because the output is formatted (in libavutil/timestamp.h) as "%.6g", which limits the total field length. Eventually, for offsets greater than 100000 seconds (about 28 hours), fractions of a second disappear altogether, and the timestamps are logged as whole seconds Allan Cady via ffmpeg-devel
2024-03-11 13:43               ` Andreas Rheinhardt
2024-03-11 14:26                 ` Andreas Rheinhardt
2024-03-11 19:49                   ` epirat07
2024-03-11 23:46                     ` [FFmpeg-devel] [libavutil/timestamp.h: Fix loss of precision in timestamps for silencedetect on long files] Allan Cady via ffmpeg-devel
2024-03-11 19:11                 ` [FFmpeg-devel] [PATCH] change av_ts_make_time_string precision Marton Balint
2024-03-12  2:57                   ` Allan Cady via ffmpeg-devel
2024-03-12 21:24                     ` Marton Balint
2024-03-13  6:03                       ` Allan Cady via ffmpeg-devel
2024-03-17 23:40                         ` [FFmpeg-devel] [PATCH] avutil/timestamp: keep microsecond precision in av_ts_make_time_string Marton Balint
2024-03-19 21:14                           ` Marton Balint
2024-03-19 23:38                           ` Andreas Rheinhardt
2024-03-20 12:44                             ` Andreas Rheinhardt
2024-03-20 19:51                               ` Marton Balint
2024-03-22 21:50                                 ` Marton Balint
2024-03-22 21:58                                 ` Andreas Rheinhardt
2024-03-23 11:14                                   ` [FFmpeg-devel] [PATCH v2 1/2] avutil/timestamp: introduce av_ts_make_time_string2 for better precision Marton Balint
2024-03-23 11:14                                     ` [FFmpeg-devel] [PATCH v2 2/2] avutil/timestamp: change precision of av_ts_make_time_string() Marton Balint
2024-03-25 18:27                                     ` [FFmpeg-devel] [PATCH v2 1/2] avutil/timestamp: introduce av_ts_make_time_string2 for better precision Marton Balint
2024-03-25 22:13                                       ` Allan Cady via ffmpeg-devel
2024-06-17 20:08                                     ` Hendrik Leppkes
2024-06-17 20:33                                       ` [FFmpeg-devel] [PATCH] avutil/timestamp: avoid possible FPE when 0 is passed to av_ts_make_time_string2() Marton Balint
2024-06-18  7:55                                         ` Rémi Denis-Courmont
2024-06-18 19:07                                           ` Marton Balint
2024-06-29  8:24                                             ` Marton Balint
2024-03-17 23:43                         ` [FFmpeg-devel] [PATCH] change av_ts_make_time_string precision Marton Balint
2024-03-19 22:46                           ` Allan Cady via ffmpeg-devel
2024-03-13  6:09                       ` Allan Cady via ffmpeg-devel
2024-03-11  3:57             ` [FFmpeg-devel] [PATCH] libavutil/timestamp.h: Fix loss of precision in timestamps for silencedetect on long files Allan Cady via ffmpeg-devel
2024-03-11  4:13             ` Allan Cady via ffmpeg-devel

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1504257982.42894.1708724906320@mail.yahoo.com \
    --to=ffmpeg-devel@ffmpeg.org \
    --cc=allancady@yahoo.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

Git Inbox Mirror of the ffmpeg-devel mailing list - see https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

This inbox may be cloned and mirrored by anyone:

	git clone --mirror https://master.gitmailbox.com/ffmpegdev/0 ffmpegdev/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 ffmpegdev ffmpegdev/ https://master.gitmailbox.com/ffmpegdev \
		ffmpegdev@gitmailbox.com
	public-inbox-index ffmpegdev

Example config snippet for mirrors.


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git