cb4960 Wrote:So audio normalization... I found a tool called "normalize" (https://neon1.net/prog/normalizer.html) that is quite fast, but I'm not really sure which option would be best.

Code:

...

Any thoughts? I was thinking maybe the -a option and presenting the user with a field where they can enter +/- decibels or 0 to forgo normalization.

Seems like it uses a lossy method to normalize the audio, that is by re-encoding the file.

I prefer to use a tool called mp3gain and adjust each file individually to the same loudness without re-encoding. Here is some info and usage of the tool.

Quote:MP3Gain does not just do peak normalization, as many normalizers do. Instead, it does some statistical analysis to determine how loud the file actually sounds to the human ear.
Also, the changes MP3Gain makes are completely lossless. There is no quality lost in the change because the program adjusts the mp3 file directly, without decoding and re-encoding.

Usage:

Code:

Normalize several files:
mp3gain -r -k *.mp3
-r - apply Track gain automatically (all files set to equal loudness)
-k - automatically lower Track/Album gain to not clip audio

Normalize an album:
mp3gain -a -k *.mp3
-a - apply Album gain automatically (files are all from the same
                  album: a single gain change is applied to all files, so
                  their loudness relative to each other remains unchanged,
                  but the average album loudness is normalized)
-k - automatically lower Track/Album gain to not clip audio

Thanks! That does seem to be the better tool. Plus it can operate directly on .mp3 files so I wouldn't have to extract audio to .wav first.

Hello,

I have just released version 26.0 of subs2srs.

Download subs2srs v26.0 via SourceForge

What Changed?

Added option to normalize the audio clips (Thanks juniperpansy and Pauline!). If you would like to change the way normalization is done, you may edit the "Normalize Audio Arguments" preference in the Preferences dialog. By default, subs2srs uses the method in Pauline's first example (see 2 posts up).

cb4960

cb4960 Wrote:I have just released version 26.0 of subs2srs.
cb4960

Hi cb4960! Only just got to using this software properly today (thanks to Nukemarine's video series) and just wanted to say thanks for putting in the effort to develop and maintain this software! It's come such a long way and is really great! I didn't have much to contribute to the thread other than that...

The only thing that really occurred to me was that on some screen captures, the image hadn't changed from one subtitle to the next. There was a lot of dialogue and the camera angle never changed so the images are all more or less the same. Not only that, but they were really quite boring images.

Obviously that's nothing to do with Subs2SRS, but maybe there could be some feature on the Preview window to select maybe... Preview images +/-2sec in each direction? So essentially you'd get the choice of 3 images; the original, auto-generated one, plus 2 that were from the video +2 seconds and -2 seconds from the auto-generated one.

Maybe no one else cares about this kind of thing but I just thought I'd share anyways, as a kind of first-reaction idea (Obviously generating extra shots would add a lot of processing time, so it could just be a check-box or something in the Generate Snapshot portion of the interface).

Just finished importing FLCL Episode 1. Going to do the rest of the series shortly. Thanks so much for this software - it works like a champ!

-J-

Thanks so much for this software. It's so useful.

I managed to make up cards from one DVD and it worked great.

However, on a different DVD I have trouble getting past the first 26 mins, i.e. the first VOB file.

The DVD is split over several VOB files with a single IFO. Tools like VSrip take the IFO and find the VOBs correctly, producing a single set of .idx/.sub files.

With subs2srs however, I would get no sound or images after about the 26 minute mark. At first I thought it was simply that subs2srs doesn't deal with multiple VOB files so I tried making a single VOB file. I used various tools to merge the VOBs and I also ripped the DVD again as a single VOB.

However, in all cases the result is that the audio is extracted correctly for the whole DVD, but the thumbnails are completely wrong (i.e. from a different part of the show) after about the 25 minute mark, i.e. just before the original boundary in VOB files.

I suspected maybe the subtitle file is to blame but I've been unable to find another tool to product the .idx/.sub files. The files produced by AviDemux can't be read by subs2srs (I get errors about there being no lines of dialogue).

Any ideas? If it's the VOB file to blame to I just convert to MPEG? Or is it more likely the subtitle files? And if so, is there another tool I should try for extracting them?

Thanks for any help!

brianskold Wrote:Thanks so much for this software. It's so useful.

I managed to make up cards from one DVD and it worked great.

However, on a different DVD I have trouble getting past the first 26 mins, i.e. the first VOB file.

The DVD is split over several VOB files with a single IFO. Tools like VSrip take the IFO and find the VOBs correctly, producing a single set of .idx/.sub files.

With subs2srs however, I would get no sound or images after about the 26 minute mark. At first I thought it was simply that subs2srs doesn't deal with multiple VOB files so I tried making a single VOB file. I used various tools to merge the VOBs and I also ripped the DVD again as a single VOB.

However, in all cases the result is that the audio is extracted correctly for the whole DVD, but the thumbnails are completely wrong (i.e. from a different part of the show) after about the 25 minute mark, i.e. just before the original boundary in VOB files.

I suspected maybe the subtitle file is to blame but I've been unable to find another tool to product the .idx/.sub files. The files produced by AviDemux can't be read by subs2srs (I get errors about there being no lines of dialogue).

Any ideas? If it's the VOB file to blame to I just convert to MPEG? Or is it more likely the subtitle files? And if so, is there another tool I should try for extracting them?

Thanks for any help!

For movies on DVD, I recommended converting the movie to a single mpeg4 file using something like Handbrake.

For examining VOBSUB (.idx/.sub) subtitles, Subtitle Creator is a good choice.

For extracting subtitles from DVD, the only tool I have experience with is VSrip.

Thanks cb4960! I managed to get it to work after converting to MPG (using VOB2MPG). It seems there's something amiss in the way subs2srs extracts snapshots from VOB files.

I didn't know about SubtitleCreator. Thanks again!

Support for .smi subtitles in the future? How about using 3 or more types of subs if their available?

arnaldosfjunior Wrote:Support for .smi subtitles in the future? How about using 3 or more types of subs if their available?

I'm not sure that I've ever seen a .smi file. How are they generated? Where would I be able to find some examples?

Does anyone have any recommendation of whats worth watching (and creating an anki deck on?) I watch most genres of shows.

dancc Wrote:Does anyone have any recommendation of whats worth watching (and creating an anki deck on?) I watch most genres of shows.

Currently, I'm greatly enjoying the drama Vision (殺しが見える女). There are downloadable 日本語字幕　available for all of the eps on d-addicts.com.

http://d-addicts.com/forum/viewtopic_113893.htm

The dialog is pretty damn fast, so the subtitles help. I've noticed later episodes are becoming easier to understand, however, as I drill some of the subs2srs cards in Anki, and I become more familiar with the pace of rapid Japanese speech.

I've also gotten a lot of cards out of the 実写版 of らんま (also on d-addicts). Plus, it's just a damn fun show.

http://d-addicts.com/forum/viewtopic_102340.htm

I recommend doing a lot of listening without subs, too. I'm currently watching 謎の彼女Ｘ, which doesn't have subs (that I can find), but contains pretty straightforward informal Japanese. Same goes for the corresponding 漫画, btw - very easy read.

cb4960 Wrote:

arnaldosfjunior Wrote:Support for .smi subtitles in the future? How about using 3 or more types of subs if their available?

I'm not sure that I've ever seen a .smi file. How are they generated? Where would I be able to find some examples?

http://d-addicts.com/forum/viewtopic_97298.htm
This torrent contains its example.

Just a minor pet peeve. It seems that files on a network drive don't get sorted properly when using the asterisk patterns * and ?. Moving the same files to a local folder gets them sorted correctly but it would be nice not having to do that. It might be linked to how the OS handles glob(). Specifically, I am using Windows 7 on VirtualBox under a linux host with mounted file shares.

you can still use smi files just use vobsub-> subresync and save it as srt or ass or whatever. The only thing is the fps. For me it's always been 23.99. you can just check if it's correct or not by playing it with the video once you save it. if it's not correct then try another fps

I've only seen smi files with korean subtitle files.

Hello,

I have just released version 26.1 of subs2srs.

Download subs2srs v26.1 via SourceForge

What Changed?

Fixed bug that prevented the correct context lines from being added to the import file if those lines were pruned out.

For example, say that I setup subs2srs to both prune lines > 2500 milliseconds and also provide 2 leading and 2 trailing context lines. Here is the result of the pruning for the example:

Line 1
Line 2
Line 3 (pruned out)
Line 4 (pruned out)
Line 5
Line 6 (pruned out)
Line 7
Line 8

When determining the context for Line 5 this was the result:

Leading: {Line 1, Line 2}. (It should be {Line 3, Line 4}).
Trailing: {Line 7, Line 8}. (It should be {Line 6, Line 7}).

cb4960

cb4960 Wrote:Hello,

I have just released version 26.1 of subs2srs.

Download subs2srs v26.1 via SourceForge

What Changed?

Fixed bug that prevented the correct context lines from being added to the import file if those lines were pruned out.

For example, say that I setup subs2srs to both prune lines > 2500 milliseconds and also provide 2 leading and 2 trailing context lines. Here is the result of the pruning for the example:

Line 1
Line 2
Line 3 (pruned out)
Line 4 (pruned out)
Line 5
Line 6 (pruned out)
Line 7
Line 8

When determining the context for Line 5 this was the result:

Leading: {Line 1, Line 2}. (It should be {Line 3, Line 4}).
Trailing: {Line 7, Line 8}. (It should be {Line 6, Line 7}).

cb4960

Thanks a ton!

Downloaded the new version and used it to import GTO 第三話. Seems to work great. Thanks!!

Hello,

I have just released version 26.2 of subs2srs.

Download subs2srs v26.2 via SourceForge

What Changed?

● Fixed issue where .trs parser ignored the last line of dialog.

● Fixed issue where .srt parser ignored the last line of dialog unless file ended with 2 newlines.

cb4960

Hello,

I have just released version 26.3 of subs2srs.

Download subs2srs v26.3 via SourceForge

What Changed?

● Upgraded ffmpeg to zeranoe git-8bdba0b (120914) build. It's been 11 months since the last time I upgraded ffmpeg. It should have support for all the latest and greatest video formats.

cb4960

Hello,
I'm trying to import a tsv file, but it works only for the first card. I get this message:

"Appeared twice in file: 2001_1"

("2001" is the name of the deck, but it's the same with all the decks)

I've had this issue since switching to Anki 2, Ubuntu edition.

The first post in this thread clearly states that this resource is for Windows only.
Are we Mac users out of luck, or is there information elsewhere (perhaps even buried in the 30+ pages of this thread!) as to how a similar learning method could be implemented on our machines?

jmignot Wrote:The first post in this thread clearly states that this resource is for Windows only.
Are we Mac users out of luck, or is there information elsewhere (perhaps even buried in the 30+ pages of this thread!) as to how a similar learning method could be implemented on our machines?

Use Parallels Desktop or VMware Fusion and run all the Windows programs you want till the cows come home.

Marzian Wrote:Hello,
I'm trying to import a tsv file, but it works only for the first card. I get this message:

"Appeared twice in file: 2001_1"

("2001" is the name of the deck, but it's the same with all the decks)

I've had this issue since switching to Anki 2, Ubuntu edition.

This question might be more appropriate for the Anki forums, but if you email me the tsv file, I can probably help you.

shinsen Wrote:

jmignot Wrote:The first post in this thread clearly states that this resource is for Windows only.
Are we Mac users out of luck, or is there information elsewhere (perhaps even buried in the 30+ pages of this thread!) as to how a similar learning method could be implemented on our machines?

Use Parallels Desktop or VMware Fusion and run all the Windows programs you want till the cows come home.

If you want to avoid buying a copy of Windows, subs2srs will also work under Linux with Wine installed. From the readme:

1) Install Wine 1.3 with the following command:
sudo apt-get install wine1.3

2) Install .Net 3.5 with the following command:
bash winetricks dotnet35

3) Unzip subs2srs.

4) In the unzipped directory, either double-click on subs2srs.exe or
type "wine subs2srs.exe".

I have tested this with Linux Mint 12.

Edit: For jmignot's case, step 0 would be to install Linux Mint (or whatever) in a Virtual Machine created with VMware/Virtual Box.