Hello,
(I'm not sure this is where to ask, if not, could you tell me where should I do it please ?)

I'm here to ask about this post : https://e621.net/posts/5634956
Parented to this one : https://e621.net/posts/5636212?q=sugene

Mine was an audio edit of the original Gif, it has been deleted for "Inferior version/duplicate"

I would like to know which of those two option has caused deletion ?

(Because the already webm version without sound is still there with also a loss of quality compared to the original gif, And, that it was previously already uploaded on my account with less good audio, and bigger filesize, But, were approved and kept)

Waiting for answers, Have a great day.

jdjckjctt.

For one thing, your version was a lower resolution - yours is 657x960, the parent post is 720x1052. Notably, the GIF is also 657x960 - the video is not just a conversion of the GIF (and not just resolution, it's missing many of the artifacts that GIF adds).
I'm guessing your version was a direct conversion of the GIF with added audio, which is just inferior to a WEBM created from the original non-GIF source.

jdjckjctt said:
Hello,
(I'm not sure this is where to ask, if not, could you tell me where should I do it please ?)

I'm here to ask about this post : https://e621.net/posts/5634956
Parented to this one : https://e621.net/posts/5636212?q=sugene

Mine was an audio edit of the original Gif, it has been deleted for "Inferior version/duplicate"

I would like to know which of those two option has caused deletion ?

(Because the already webm version without sound is still there with also a loss of quality compared to the original gif, And, that it was previously already uploaded on my account with less good audio, and bigger filesize, But, were approved and kept)

Waiting for answers, Have a great day.

jdjckjctt.

See the pinnee topic #23578 for a list of deletion reasons and what they mean.

In this case, I believe your sound edit was deleted for being "inferior" due to it having a lower resolution (i.e., 657x960) as compared to the original WEMB (i.e., 720x1052).
Edits must be at the same level of quality as the original content.

Your previous sound edit was approved as it had the same quality of the original GIF version at that time.
However, the original WEBM version has since been posted in a higher resolution. Any subsequent edits must be at the same quality as the newer version.

Okay So i'll have to re-export with the new sized webm, but what about the IPS in videos, does it have to match as well ? (Might be stupid to ask, but the new one have 30 IPS which looks laggy, when my edit was at 23,976 IPS and looks smoother)

Just in case it is that precise matching to not get deleted again.

jdjckjctt said:
Okay So i'll have to re-export with the new sized webm, but what about the IPS in videos, does it have to match as well ? (Might be stupid to ask, but the new one have 30 IPS which looks laggy, when my edit was at 23,976 IPS and looks smoother)

Just in case it is that precise matching to not get deleted again.

You should not be re-encoding the video - use the webm that's on the site, add a sound track, but do not re-encode the video. That may be difficult depending on the tool you use. The best bet is to use something like ffmpeg where you can explicitly specify that the video doesn't get re-encoded. Talking about video settings means you almost certainly are reencoding it.

Let's do it all over again then ... Hope this will work

About FFMPEG at which point of production is it used ? I heard the name but not sure about what it does exactly

jdjckjctt said:
About FFMPEG at which point of production is it used ? I heard the name but not sure about what it does exactly

FFMPEG is a general command-line tool for converting video and audio formats. It can be used for transcoding between formats, doing high-quality compression, mixing tracks (like taking video from one source and audio from another), along with quite a few other things. Notably, one of the formats it works with is individual frames as separate files, so it can be used for making a video file after exporting as frames as well.

For adding audio to a video, doing it properly depends on the source format of the audio, and its bitrate. This usually would work well enough though:
ffmpeg -i (input video file) -i (input audio file) -map 0:v -map 1:a -c:v copy -c:a libopus -b:a 160k (output file)
Breaking that down, the command is saying to map video tracks from the first input and audio tracks from the second input into a single output, copying the video without re-encoding, while re-encoding the audio track into the opus audio codec with a 160k bitrate. If the audio codec you're using is already opus or vorbis you can just use c:a copy instead of c:a libopus -b:a 160k, which would directly copy the audio as well. Other formats need to be transcoded to fit into a webm container.

160k is also a bit low, depending on the audio contents. For high-fidelity audio you could use something higher.

scth said:
FFMPEG is a general command-line tool for converting video and audio formats. It can be used for transcoding between formats, doing high-quality compression, mixing tracks (like taking video from one source and audio from another), along with quite a few other things. Notably, one of the formats it works with is individual frames as separate files, so it can be used for making a video file after exporting as frames as well.

For adding audio to a video, doing it properly depends on the source format of the audio, and its bitrate. This usually would work well enough though:
ffmpeg -i (input video file) -i (input audio file) -map 0:v -map 1:a -c:v copy -c:a libopus -b:a 160k (output file)
Breaking that down, the command is saying to map video tracks from the first input and audio tracks from the second input into a single output, copying the video without re-encoding, while re-encoding the audio track into the opus audio codec with a 160k bitrate. If the audio codec you're using is already opus or vorbis you can just use c:a copy instead of c:a libopus -b:a 160k, which would directly copy the audio as well. Other formats need to be transcoded to fit into a webm container.

160k is also a bit low, depending on the audio contents. For high-fidelity audio you could use something higher.

I'm working on AE for the editing part, isn't it okay to add sound through it, export it and then convert the end file to webm with ffmpeg ?

jdjckjctt said:
I'm working on AE for the editing part, isn't it okay to add sound through it, export it and then convert the end file to webm with ffmpeg ?

If you export a video file in after effects, it's going to re-encode. That loses quality. However, once you have the audio how you want it, you could export it and then combine it with the original video file using a different tool that doesn't re-encode.

So it'll be better if i do my sound design in AE, export the audio only, then add it to the original video through ffmpeg and convert it to webm if it isn't the case with also ffmpeg ?
For editing videos, not gif's since we have to re-encode them anyways

jdjckjctt said:
So it'll be better if i do my sound design in AE, export the audio only, then add it to the original video through ffmpeg and convert it to webm if it isn't the case with also ffmpeg ?
For editing videos, not gif's since we have to re-encode them anyways

Yes. The command I gave doesn't do conversion to webm if the video isn't already, though - that's a more involved process. When you start with a post on e621 it'll usually already be webm, though (or mp4 with AV1 video for more recent ones, which is a bit different as well).

And if i do a longer version of it (like when the loops are too short to really enjoy it, and the sounds is looping obviously without variation)
And so add more loops, is it accepted to re-encode it then ? Because it isn't matching anymore (which is the case for this one as well)

jdjckjctt said:
And if i do a longer version of it (like when the loops are too short to really enjoy it, and the sounds is looping obviously without variation)
And so add more loops, is it accepted to re-encode it then ? Because it isn't matching anymore (which is the case for this one as well)

FFMPEG can actually loop videos without re-encoding as well. For instance, if you want to make a video loop twice, you can do
ffmpeg -stream_loop 2 -i (input file) -c copy (output file)
Other effects generally would require re-encoding, yes, but at that point it's not just an audio edit.
(Whenever -c copy works, it means the actual video contents aren't being damaged at all; stream_loop basically just literally copies the video content, without otherwise effecting it)
Oh, and I misremembered a bit - stream_loop 2 actually means to copy the entire video two times, so the overall effect is playing the video three times. -stream_loop 1 would copy it only once, so the output contains the original video twice. In general stream_loop x means that the output has x + 1 copies of the original.

Okay Thank's, for just adding loops i'll try this out !!

Hope Edited content with added phases (like speeding up loops to then mix with an already animated climax scene) won't be deleted as well T-T if there's little loss since i can not avoid re-encoding

jdjckjctt said:
Okay Thank's, for just adding loops i'll try this out !!

Hope Edited content with added phases (like speeding up loops to then mix with an already animated climax scene) won't be deleted as well T-T if there's little loss since i can not avoid re-encoding

Unfortunately, changing part of a video's speed is especially lossy. Not only are you re-encoding, but you're also necessarily deleting frames, and likely interpolating between them as well. Doing it properly requires designing with the changing speed in mind upon initially creating the frames. At that point it's best to work with the artist, rather than trying to do it in post.