Happy Easter!

I need to combine many individual audio snippets into one big whole. Unfortunately, the audio files are levelled differently in terms of loudness. To save myself work, I wanted to bring all files to the same loudness. And as automated and reliable as possible.
I tried it with NUGEN Audio AMB. But the program is frighteningly slow on my brand new laptop. In the same time I could do it myself by ear.

Does anyone know another tool with batch processing?

Thank you for your suggestions and recommendations.

 

Which DAW do you use ?

Try Audacity: www.audacityteam.org
- There you can also specify the normalization module under File - Batch Setup.

Audiograbber: www.audiograbber.org
- If you drag'n drop Wav files into Audiograbber, you will be offered to normalize the files. Works also with sampling rates 44.1 kHz

 

Last edited: Apr 18, 2022

Wavelab!

 

didn't try it, but have a look at Zplane - ppmBatch

 

In Reaper you can batch export separate tracks/clips/whatever and normalize them automatically to values of your choice. You can set them all to the same LUFS-i too, which i think is what you're looking for. You can also normalize the clips in session.

 

I was looking for this the other day and here's the tutorial for it

its exactly what @mk_96 said - the lufs matching thing

 

Izotope RX

 

Thank you very much for your answers.

The important thing is that not all files are levelled to a uniform value (e.g. EBU R128). Instead, many files are to be levelled to the loudness of a "reference audio file".
The point is that I am currently working on the audio mix for a television series. Extended scenes are to be added for release on Blu-ray. These new scenes now need to be adjusted to match the volume of the existing scenes. Up until now, I've always done this mostly by ear, because that's how I get the best results. But now there are just too many scenes. (And the time until the deadline is just too short.) So I would like to automate this.

Sequoia. I have set Sequoia to automatically measure each loaded file to EBU R128 and then write this value to a text file.
I can also level each file (destructively) by loudness in Sequoia. But then all files have the same loudness. This would "destroy" the existing audio mix.

 

ffmpeg if you're comfortable with the command line.

1. Measure LUFS-I of reference track:

ffmpeg -i myreferencefile.wav -af loudnorm=print_format=summary -f null -
Input Integrated: -28.4 LUFS

2. Apply reference LUFS-I (-28.4 in this case) to other file(s):

ffmpeg -i track246.wav -af loudnorm=I=-28.4 track246_new.wav

Use whatever your OS comes with to automate.

 

@Edna_Uebel is this the kind of thing your after?

if it is I'm sure someone will have a idea if there are better applications than this.

 

Diagnosis: Skilled but lack of time

 

Last edited: Apr 18, 2022

The important part starts at 7:20, Reaper can do this natively now, realtime or when rendering.

 

Adobe Audition,and use match loudness settings. It will measure your ref track, give you the info, and any track you want at that level add them then run the processing, super easy.

 

For all audio batch operations I always use dbPowerAmp Batch converter. It's freaking fast (uses all the cores or the number you choose), easy to use and good quality.
I can't say enough good things about this program. When slimming down large Kontakt libraries (loads of samples from 24 to 16 bits) it's almost a pleasure seeing how well it performs. Hard battle tested

Found the effect in dbPowerAmp:
"
Volume Normalize

There are 6 different ways of adjusting the volume of an audio track. Note volume normalize is actually applied to the audio and is a non-reversible proces, as such it should be done on a copy of a lossless library, keeping the originals safe:

• Peak to Peak: simplest volume normalize, the peak (or maximum) value is found and the whole track volume is adjusted so that peak touches the desired volume,
  
  
• ReplayGain: a more advanced normalization, the average loudness is calculated and the volume is adjusted. This method does not use ID Tags to store ReplayGain values, rather the audio is adjusted. If a Replay Gained track is played next to a non-Replay Gained track, it will sound quieter, as Replay Gain requires head-room to prevent clipping.
  
  
• EBU R128: a newer standard of ReplayGain, the average loudness is calculated and the volume is adjusted. This method adjusts the audio. Internally the audio is calculated to a -18LUFS level, it can be adjusted with the dB scale on the right.
  
  
• Adaptive: the best way to describe adaptive, is like having your hand on the volume knob, constantly adjusting it up and down (a window is used to detect loudness) to compensate for quiet parts, ideal for creating a track to play in a noisy environment (such as car). Adaptive is offered for EBU and Peak to Peak volume calculation.
  
  
• Fixed Amplification: Amplifies the audio by a fixed amount, +6 dB is x2, where -6 dB is divide by 2, be careful not to clip audio (go over the maximum).

Maximum Amplification sets a limit on the amount of amplification applied (set too high, on adaptive and silence will become noise).

Desired Volume can be set higher than 0 dB (which is maximum), setting 12 dB is 4x the maximum, use with care otherwise the signal will clip.

This effect is non-live."

PS. About other options.
Reaper seems a good alternative because if memory serves the last versions can do batch processing with multicore.
ffmpeg is a beast if you're willing to go command line.
Never tested RX's batch conversion. But if the amazing @rollerball 's guide for slimming down Kontakt libs recommends it can't go wrong.

 

Last edited: Apr 19, 2022

Thanks for the interesting suggestions. I am normally using Nuendo. Working with Sequoia right now is just due to current circumstances. Getting into Reaper or Audition right now would make my brain explode. I didn't even know that Audition was still being developed. The last time I used the program was 8 or 10 years ago.
I have now found a way to do it in Sequoia. There I can solve the task via a plug-in in the object editor.