iparout
9th May 2003, 21:45
Hi.
Up to a few days ago, I had been using vob2audio to extract the WAV out of the vob files, then the WAV was normalized with normalize.exe and then encoded to mp3. However, vob2audio stopped working on my PC when I changed CPUs so I was "forced" to use BeSweet. The tool is very very good, much customizable and has the advantage of an all-in-one solution, however there is only one "problem" with it. The mp3 that is created by the BeSweet is a bit lower in volume than the one created with my original (vob2audio) method. I tried messing around with BeSweet's options to see if something can do the trick but I was not fully satissfied by the result. To be more precise these were the switches I used and my comments regarding the output mp3, compared to the mp3 that was created by the vob2audio method.
-azid( -z1 -b1 -s surround -c normal --maximize ) -lame( --alt-preset 128 )
The output mp3 was about 0,3 dB lower in the high-volume scenes (gunfires, explosions e.t.c) and about 4 dB lower in the low-volume scenes (dialogues e.t.c.) don’t care about the 0.3 dB difference since it’s not audible, but the 4 dB difference is very noticeable and gives me the impression of a low-volume mp3 overall.
Afterwards, I tried to “fix” the 4 dB difference in the low-volume parts by applying a PreGain of 4 dB
-azid( -z1 -b1 -s surround -c normal --maximize ) -ota( -g 4db ) -lame( --alt-preset 128 )
Now the volume was the same in the low-volume parts of the mp3, however in the high-volume parts SERIOUS clipping occurred.
I then tried -azid( -z1 -b1 -s surround -c normal ) -ota( -G 1 ) -lame( --alt-preset 128 ) which gave me the exact same results as the first command.
Finally I used hybrid gain but again the output was the same as the 1st and 3rd command.
My questions :
1) How is it possible to make the low-volume parts of the BeSweet mp3 louder, by leaving the high-volume parts as they are so as to avoid clipping ? Why can vob2audio achieve that ?
2) As I said, with the 2nd command SERIOUS clipping occurred, however I had to open the file with an audio editor to find that out. Simply by listening to the mp3, I couldn’t tell the difference (I don’t have very well-trained ears). So, since I cannot hear the clipping, should I leave it as it is or will it have any other impact on the speakers e.t.c
I have attached a picture with the waveforms of vob2audio soundfile and the ones that BeSweet produced. The one on the top is the 4dB Gain file, the one in the middle is the vob2audio one (used for reference) and the bottom one is the soundfile that was produced by either -azid( -z1 -b1 -s surround -c normal --maximize ) -lame( --alt-preset 128 ) , PostGain or HybridGain
http://users.panafonet.gr/iparout/Waveforms.jpg
Thanks in advance for any help.
Up to a few days ago, I had been using vob2audio to extract the WAV out of the vob files, then the WAV was normalized with normalize.exe and then encoded to mp3. However, vob2audio stopped working on my PC when I changed CPUs so I was "forced" to use BeSweet. The tool is very very good, much customizable and has the advantage of an all-in-one solution, however there is only one "problem" with it. The mp3 that is created by the BeSweet is a bit lower in volume than the one created with my original (vob2audio) method. I tried messing around with BeSweet's options to see if something can do the trick but I was not fully satissfied by the result. To be more precise these were the switches I used and my comments regarding the output mp3, compared to the mp3 that was created by the vob2audio method.
-azid( -z1 -b1 -s surround -c normal --maximize ) -lame( --alt-preset 128 )
The output mp3 was about 0,3 dB lower in the high-volume scenes (gunfires, explosions e.t.c) and about 4 dB lower in the low-volume scenes (dialogues e.t.c.) don’t care about the 0.3 dB difference since it’s not audible, but the 4 dB difference is very noticeable and gives me the impression of a low-volume mp3 overall.
Afterwards, I tried to “fix” the 4 dB difference in the low-volume parts by applying a PreGain of 4 dB
-azid( -z1 -b1 -s surround -c normal --maximize ) -ota( -g 4db ) -lame( --alt-preset 128 )
Now the volume was the same in the low-volume parts of the mp3, however in the high-volume parts SERIOUS clipping occurred.
I then tried -azid( -z1 -b1 -s surround -c normal ) -ota( -G 1 ) -lame( --alt-preset 128 ) which gave me the exact same results as the first command.
Finally I used hybrid gain but again the output was the same as the 1st and 3rd command.
My questions :
1) How is it possible to make the low-volume parts of the BeSweet mp3 louder, by leaving the high-volume parts as they are so as to avoid clipping ? Why can vob2audio achieve that ?
2) As I said, with the 2nd command SERIOUS clipping occurred, however I had to open the file with an audio editor to find that out. Simply by listening to the mp3, I couldn’t tell the difference (I don’t have very well-trained ears). So, since I cannot hear the clipping, should I leave it as it is or will it have any other impact on the speakers e.t.c
I have attached a picture with the waveforms of vob2audio soundfile and the ones that BeSweet produced. The one on the top is the 4dB Gain file, the one in the middle is the vob2audio one (used for reference) and the bottom one is the soundfile that was produced by either -azid( -z1 -b1 -s surround -c normal --maximize ) -lame( --alt-preset 128 ) , PostGain or HybridGain
http://users.panafonet.gr/iparout/Waveforms.jpg
Thanks in advance for any help.