Slow Conversion Speed

[Gouranga Das]

Hi! ,

Every day, I need to convert videos in various formats. There are so many converters available to do this job. Some of them are also available at free of cost.

I need some facilities like auto crop area, volume normalization, splitting into multi-parts, scrolling text with shadow and transparency and multi output profile (i.e. One input video, more than one output type and settings) and all should be under one roof.

But I still haven’t found any tool having all of these facilities. So I have to give a try with VB.net and ffMPEG to make my own. But I am facing a very low speed conversion. Here is the procedure....

1. After adding a video file, it (i.e. my tool) generates a command to find all related information like crop-area, video codec, audio codec, bitrates etc.


ffMPEG -i “input.vob”

For Auto Crop Area ...........

ffMPEG -threads 4 -i “input.vob” -ss 0 -t 10 -vf "cropdetect=24:16:0" -an -f rawvideo -y nul

To extract a thumbnail image for further purpose like resizing crop-area...

ffMPEG -threads 4 -i “input.vob” -vcodec mjpeg -ss 200 -vframes 1 -an -f rawvideo “extracted.jpg”

2. It generates a command to extract audio as MP3.

ffMPEG -threads 4 -i "input.vob" -xerror -vn -acodec libmp3lame -ab 128K -ac 2 "extracted.mp3"

3. To normalize audio using MP3Gain.....

mp3Gain /r /c "extracted.mp3"

4. At last, to convert the video as MP4 also splitting into multi parts.....

ffMPEG -threads 4 -i "input.vob" -i "extracted.mp3" -map 0:v -map 1:a -force_key_frames 600,1200 -f segment-segment_times 600,1200 -segment_time_delta 0.05 -vf "crop=352:272:0:10 [tmp];[tmp] scale=320x240 [tem];[tem] drawtext=fontfile=TahomaBD.ttf:textfile=Text.txt:f ontsize=17:fontcolor=White@0.6:x=320-20*t:y=210: shadowx=1:shadowy=1" -vcodec mpeg4 -b:v 512K -r 15 -s 320x240 -aspect 4:3 -acodec aac -strict experimental -ab 96K -ar 32000 -ac 2 "output.mp4"

To make 3gp......

ffMPEG -threads 4 -i "input.vob" -i "extracted.mp3" -map 0:v -map 1:a -force_key_frames 600,1200 -f segment -segment_times 600,1200 -segment_time_delta 0.05 -vf "crop=352:272:0:10 [tmp];[tmp] scale=176x144 [tem];[tem] drawtext=fontfile=TahomaBD.ttf:textfile=Text.txt:f ontsize=13:fontcolor=White@0.6:x=176-20*t:y=124:shadowx=1:shadowy=1" -vcodec h263 -b:v 256K -r 15 -s 176x144 -aspect 4:3 -acodec aac -strict experimental -ab 96K -ar 32000 -ac 2 "output.3gp"


I have used “scale=320x240” and “scale=176x144 “because ffMPEG draws text on input video with given font size and results a small font sized text on output depending on input video resolution. But I want the text on given font size on output. Any other procedure to have the same result?

From 1 to 3 of above processes are usual. But at last process when actual conversion is going on, it takes 45 minutes per process for 30 minutes video.

How can I speed up the job?

Please help.

Thanks in advance.


Gouranga

[Baldrick]

Have you tried another ffmpeg frontend and just test encode the vob to 3gp to compare the encoding speed? Like winff. Use similar settings except the text/font drawing.

Or test encode without the drawtext and compare the speed.

[Gouranga Das]

Have you tried another ffmpeg frontend and just test encode the vob to 3gp to compare the encoding speed? Like winff. Use similar settings except the text/font drawing.

Or test encode without the drawtext and compare the speed.

If I exclude the draw text option from the command line parameter through winff or just typing in command prompt as below, it takes less than 3 minutes.

ffMPEG -threads 4 -i "input.vob" -force_key_frames 600,1200 -f segment -segment_times 600,1200 -segment_time_delta 0.05 -vf "crop=352:272:0:10" -vcodec h263 -b:v 256K -r 15 -s 176x144 -aspect 4:3 -acodec aac -strict experimental -ab 96K -ar 32000 -ac 2 "output.3gp"

or

ffMPEG -threads 4 -i "input.vob" -vf "crop=352:272:0:10" -vcodec h263 -b:v 256K -r 15 -s 176x144 -aspect 4:3 -acodec aac -strict experimental -ab 96K -ar 32000 -ac 2 "output.3gp"

I think there may be some errors in parameters as stated in my first post.
I need all of the options along-with speed.

What should I do now?

Thanks.

[hello_hello]

I need some facilities like auto crop area, volume normalization, splitting into multi-parts, scrolling text with shadow and transparency and multi output profile (i.e. One input video, more than one output type and settings) and all should be under one roof.

MeGUI?
The only part I'm not 100% sure of is the scrolling text (never done it myself) but if it can be done using an AVISynth script (and no doubt it can), MeGUI's OneClick encoder should do the rest.
It'll batch encode and split, and it'll also use any AVISynth script templates you care to create. The OneClick encoder itself can be configured to use a particular encoder preset for both audio and video, which script template it uses, and the cropping and resizing can be done automatically or manually added to a script template etc. The whole setup can then be saved as a OneClick encoder preset. Once you have OneClick configured with various presets, you can (batch) encode in a variety of ways simply by selecting the appropriate OneClick preset and opening a file or group of files.

3g isn't an output option, but AVI, M2TS, MKV and MP4 are.

[Gouranga Das]

Thanks hello_hello.

MeGUI ?

Its too hard to digest for a guy like me.
Simply can't understand what to do, how and why.

3g isn't an output option

It is 3gp, Sir.

Below, the screen shots are of my Simple GUI just as per my need.

[ozok]

^This is indeed a simple GUI.