staxrip
Being able to set detect volume in general options
Staxrip includes 3 types of normalizations volumedetect dynaudnorm loudnorm and you can choose one of them. But in my view volumedetct is complementary to the other two since it is the one that analyzes the volume level in order to set it at a correct level. I use the other two to normalize the dialogues of the surround formats. What I think would be best is to first analyze the audio with volumedetect and establish the db that the user considers correct. Then if you want to apply one of the other normalizations. In fact, I would put volumedetect in general options since I don't know who might be interested in having a video with audio with less than -13 db, for example. As has already been mentioned here on some occasions, I do not know if the way in which this application was first programmed allows this type of modifications to be possible, and if they are, they carry a significant workload. I leave a link where this matter is more or less explained. https://trac.ffmpeg.org/wiki/AudioVolume
I don't think that I understand your message. 😕
You can select the normalization method used with ffmpeg, which then will be used to do the normalization in case you enable this option.
In fact, I would put volumedetect in general options since I don't know who might be interested in having a video with audio with less than -13 db, for example.
Therefore you use normalization to raise the volume to 0dB. 🤔
Basically what I'm explaining is the possibility of being able to use two normalizers. Let's imagine an audio with -15db If I apply only dynaudnorm it balances the dialogues but the general volume level will still be low If I only apply volume, detect the level if I left it at the correct level, but it is a linear increase in volume, the effect of the 5.1 type audios will remain the same, very low voices and very loud music. The correct thing is to apply volumedetect to set audio to 0db for example and then apply dynaudnorm to level the voices and other effects.
It seems one of us has an understanding problem. 🤔
Normalization by definition means, that you make the track as loud as possible without clipping, meaning bringing the peak volume to 0dB. There are certain ways to do that, but none would make some channels louder while keeping the overall peak volume at -15dB. Whatever the effect is, it is not Normalization anymore.
And from what I see and know is, that each of those methods brings the result to 0dB. 🤔
When audio normalization is applied in Staxrip with the volumedetect option, an analysis of the audio is first performed to check where the maximum dB peaks are, and then the dB reduction or increase is applied based on what has been specified in the parameters panel, which by default is set to 0dB.
However, if the dynaudnorm option is chosen, this preliminary analysis is not done, so it is not known whether I need to increase or decrease the dB in the options panel. If an audio is at -15dB, even though this normalization is configured to set the audio to 0dB, it does not do so and remains at -15dB. The same happens with loudnorm.
Therefore, after applying either of these two options to the resulting audio, I have to apply the volumedetect option to find out the dB level of the audio and adjust it up or down to obtain, in my case, 0dB.
And excuse me if I am not making myself clear or if I am completely wrong.
Ok, I got you.
The problem is not, that a second run is needed - when using dynaudnorm or loudnorm the workflow is broken, means that they are not applied correctly. So we have a bug here.