# Allow volume to be used

There are many occasions I wanted normalizing to amplitude. For example I want to normalize to 0.5. How are novices supposed to know that -6 means halving the volume? Also, converting to 8-bit PCM requires 0.00390625, but -48 dB is 0.003981071, not exactly the same, so some volumes may map wrong. I have to use calculator to get 20*log(1/256) which is -48.16479931. Very annoying.

Also, amplifying with amplitude. Peak could be amplitude. There can be a factor for amplifying, like 0.5 for half volume.

Here is a “Quick Conversion Calculator” that you may find useful: https://forum.audacityteam.org/t/quick-conversion-calculator/15589/1

Quick Conversion Calculator isn’t built in Audacity, so novices won’t know about it.

If new users join Audacity, they won’t know how to use dB. I’m worried about these.

Personally I think it was a mistake for Audacity to use linear amplitude scales at all. Outside of Audacity, dB scale is used everywhere in audio, because in general, the dB scale is much better suited for audio than a linear scale. “dB” is not difficult once you get the idea that half or double is about 6 dB. In my opinion, the thing that is confusing, is having a mix of linear (+/- 1) scale in some places and dB scale in others. I think it would be far easier for new users if “dB” were used consistently throughout.

Cool Edit split the difference. Their waveforms are in dB, but they use a distorted scale to reflect the most important parts of the show being in the top third. The straight dB range in Audacity reduces that range to a tiny, hard-to-see graphic element.

I noticed in graphic samples posted by someone in dB, their -65dB noise (which passes ACX conformance), is the same size as the show. That’s not useful to a new user, either.

Koz

And while we’re here. The out-of-the-box Cool Edit had the sound meters the whole length of the bottom of the work window. I don’t remember the range they defaulted to, but I wouldn’t be shocked to learn it’s the same expanded range I use.

Koz

I did keep a picture.

Koz

I think dB is bad. The worst places of dB that could be used is waveforms and spectrograms. Waveforms and spectrograms are much better with linear volume scale. dB is made for relative changes. Even if functions use relative changes (equalization, amplification) there should be option to use factor, like in change speed.

and yet “ACX” specify that for audiobooks to be accepted for sale by Amazon, the peak level must be no more than -3 dBFS. I’m very glad that I’m not having to constantly tell people that means “+/- 0.707946”.

Why would they limit it to 70% volume File System (FS means File System)? It doesn’t make sense.

“dBFS” means dB with reference to “Full Scale” (https://en.wikipedia.org/wiki/DBFS)
In the case of Audacity, full scale can be seen as the vertical height of the track, which we label as +/- 1.0. So +1 and -1 are equivalent to 0 dB.

Interestingly, for 32-bit float samples, 0 dB really is +1 or -1, but for 16-bit, 0 dB is actually +32767 or -32768, and for 24-bit 0 dB is +8,388,607 or - 8,388,608.

70% volume … It doesn’t make sense.

It also makes zero sense that they insist on posting master works as MP3, particularly because they’re going to resample down to other products and services. If you’re going to complain about quality/standards problems, that would be right at the top of my list. Most people can readily hear problems if they do that wrong.

Koz

Don’t rearrange words to make it completely different.

It doesn’t matter if it’s 32-bit float, 64-bit float, 16-bit float, 6-bit float, or 4000-bit float. It’s still -1 to 1. Pointless to say 32-bit float if it doesn’t matter AND it’s the only float in Audacity.

For PCM, I don’t think so. I see 8-bit PCM as -1 to 127/128 in intervals of 1/128 (if only this existed in Audacity), 16-bit PCM as -1 to 32767/32768 in intervals of 1/32768, 3-bit PCM as -1 to 3/4 in intervals of 1/4 and so on. I think they’re fixed-point, but not integer. Audacity even shows so, and if you zoom in the max is not exactly 1.

Volume effects, like Normalize or Equalizer use dB. This is very bad. I don’t like it! Even when we do operations which don’t rely on exact integers or fractions, we use simple numbers, so please do the same for volume effects. I really prefer normalizing to 0.5 volume than -6 dB. For relative volume changes, there is such a thing called factor.

Bump; this post shows the reality and what–it–should–be brilliantly: https://forum.audacityteam.org/t/6-features/46147/24

Users could do everything with their ears, still. Why mess with -6dB or -4.8dB? And why should beginners convert to 8 bit audio? I still have an ISA Soundblaster AWE 32 that is ready for a museum, but even that can play 16 bit audio…

According to https://de.wikipedia.org/wiki/DBFS#Effektivwert -3 dbFS means the tip of the sine just touches 0dB, so you would just normalize to 0dB if normalization is correct. Otherwise -3dB is about 0.5 (1/1.99526314969).