Audacity and Normalize

[alanclare]

What am I actually doing by selecting the whole song and then "Effect/Normalize", and accepting "Remove DC offset (center on 0.0 vertically)" and "Normalize maximum amplitude to -1.0 dB"?

It's just something that in another thread it seems as if people I respect are doing it - so I do it. But why?

Alan

[pjd]

Hi Alan --

Normalization consists of three steps: 1. Find the highest peak in the audio file (i.e., the loudest sample in the file). 2. Compute a scale factor (gain) that will boost the highest peak to a target level. 3. Apply the scale factor (gain) to all of the samples in the audio file.

The user specifies the target level such as +0dB or -1dB. -1dB is just a thin hair under +0dB. Choice of target level depends on the down-stream use of the audio. The closer to 0dB, the more likely the signal will be clipped (distorted) by a down-stream effect. If the file is used for playback, 0dB might be OK. In terms of perceived loudness, -1dB is not much different than 0dB.

The main goal is to increase the overall loudness of the audio during playback. Normalization is not the same as compression, BTW.

Hope this helps -- pj


https://en.wikipedia.org/wiki/Audio_normalization

[Roland]

Exactly !  Note that Audacity also has an "Amplify" feature, which is useful if two tracks have different peak levels. Amplify applies gain to both channels equally.

Cheers, Roland

[tyrosaurus]

A link about 'DC Offset'      https://manual.audacityteam.org/man/dc_offset.html

Regarding 'normalising maximum amplitude to -1 dB', I find that when normalising files in Audacity for playback on my Tyros4 hard disk player/recorder, I need to limit the maximum amplitude to -3 dB or even lower to avoid distortion.

Your own ears should tell you what the best value for playing back on your keyboard, but keep it as high as you can without it distorting.


Regards

Ian

[pjd]

Hi Ian --

Thanks for your comment.

Usually, I normalize to -3dB.

However, there are times when I've over-amplified the samples where a few samples are overdriven and clipped. Sometime these short duration clipped segments blend in with the music, sometimes they stick out like a sore thumb. Probably ought to be compressed instead...

Take care -- pj

[alanclare]

Pj, Roland & Ian

My sincere thanks for the explanations and for your time. I can follow the logic and will prove it by trying the effects of various settings on the resulting song.

Thanks Ian for the reference to the entry in the manual. I find the manual quite difficult to navigate. It's probably my fault. Impatience on my part.

Alan

[Toril S]

Hello Alan!
I got a tip from Gary that I am using when recording songs from my keyboards.
When I record with the audio recorder of the keyboard I have the master volume at 12 o'cklock.
Then I turn down record volume from 100 to 70, this is to avoid distortion.
Then I import the recording to Audacity, where I select Normalize. I set the value to -0.6. Done.
I believe the recordings are clear.
If i don't use Normalize, the recordings are too faint for my taste.

[panos]

Thank you for the info Ian and Pj. 
May I ask something?
Is it necessary to normalize always a recording that we have done on our keyboard?
I record as a midi to the psr and then export the midi as a .wav file with the on board audio recorder (and I don't change it to mp3 anymore because it is sounds more badly if I do  )

ok it is not the greatest of sound recordings with this method but I hope the sounds are   tolerant for an amateur recording.

The thing that puzzles me is the EQ settings when transforming from midi to wav and not so much a distortion sound or something that sounds suddenly too loud.
If something sounds louder is because usually I want it to sound that way (I guess).

How can I trust my ears for something that recorded to be heard from other people's ears?    
These programs maybe can correct things that we don't hear?

What is your opinion? We should always use Audacity or a DAW and do things like normalization,compression etc?

Also is it better to record to a normal pc than to the on board audio recorder?
I have bought some cables for that recently but I haven't used them yet.

[alanclare]

Thanks Toril. I'll try your settings and see how the results compare with what I've been doing up to now.

Alan

[Roland]

Hello Panos,

For the most part, the keyboards "write" (D/A) to WAV quite nicely. And if you set-up the sound on your keyboard the way you like it, this is pretty much the way it will sound in your .wav file - IF you play that file back on the same keyboard.

But as pj had already hinted to; It depends on what medium you want to play the .wav file and for what audience. A broadcaster has different sound and loudness requirements compared to a CD that you listen to in your car. And to change the original sound to these different usages is easier and more effective when done in a DAW or a very good audio mastering programme. For the normal home player, doing this on the keyboard however is totally fine.

Another time where mixing details is beneficial when done in a DAW are the drum / percussion sets. If you want to take every single one of the 16 plus drum kit items and mix them separately and put them in a perfect stereo picture - then doing that in a DAW is the only way to go about it in my view (by dissolving the drum kit to separate parts).

Hope that helps answering your question :-)  Cheers, Roland

[panos]

Thanks for talking the time to answer my question Roland! 
I can see...
Yes you 've got it right.I hear the same recording differently to different kind of devices so it is hard to find out whta exactly to correct.
Inserting the midi to a DAW I have done it before just to learn how a DAW works  but it is time concuming to adjust all parameters for each track.
Took me an afternoon to "compose" a little tune using two scales and almost a month to equalize the parts,change the voices,correct the lenghts and the timing of all notes in all trucks etc.(it was a midi from my old dgx 305)
Luckily afterwards I bought the s750 with the styles so the arrangments may not be unique but they do save a looooot of time,especially when the songs are just covers and it is just for fun.
I think I will better start by purchasing some more expencive headphones than 30 euros and I hope to hear better the details of the sounds.     

[alanclare]

Toril

I did some controlled tests, using my usual setting on Audacity and then using your figures. Your way is definitely better, even through my very poor hearing (hole in one eardrum, hearing aids in both). To get a better impression of the sound, after the Audacity treatment I save the recording as an MP3, copy it on to a memory stick and play it on my music centre which has two good speakers. That can be done in a few minutes. My best test is to play it through my car stereo which has the best sound. If I'm still doubtful I email it to my daughter and await her judgement.

I had never changed that Volume setting, so that was a first for me. Presumably, you have to change it every time you use the Audio Recorder?

Thank you for your help.

Alan

[Toril S]

Glad it worked Alan! Yes, you have to change the volume for every new recording. But that is fine. I also try my recordings on several devices. I suffer from tinnitus, so I know about hearing problems.