How to set peak recording level at -6 db automatically?

I installed Audacity 2.1.2 on windows 10 64bit.
I would like to record from my analogue sources such as tape, MD using line-in.

FAQ page(*1) tells adjusting the level, aim for a maximum peak of around –6 dB on the Recording Meter.
So I try to search the peak, the loudest part, in the tapes and MDs, but I do not know where the loudest part is.

Are there any techniques or a features of Audacity to find loudest part in a analogue source?
I will record from 100 tapes and MDs, so I would like to find it automatically.

My image like this.

  1. Play a analogue source and do something of Audacity feature which detect the loudest part in the stream. This process needs realtime, for instance it requires 60 min on 60 min tape.
  2. Play the loudest part and set recording level at -6 db at the part.
  3. Record the analogue source from the beggining.

*1:FAQ page

The abbreviated version of that for live performance recording is “occasional peaks around -6dB.”

It’s up to you to adjust the recording system (or your performance) so you achieve that. That value was chosen so you avoid sound that’s so loud it becomes distorted and so quiet it gets noisy. It’s generally the sweet spot.

Audacity doesn’t do anything in real time. It only applies effects and filters in post production—after the fact. Once you’re in post production, it’s already too late. You can’t recover from loud distortion and it’s difficult to clean up a noisy recording.

You personally are replacing the recording engineer.

Sometimes you can play through a tape, find the loudest part and set Audacity that way. Generally old tapes have a narrower quality range than Audacity does, so if you hit it reasonably close, you should be good to go. Digital formats are harder. Can you transfer the Mini Disk digitally? Does your player have a digital connection?


Thanks for your reply.

I should adjust the value but it’s a bother, because I think that I must stay listening to the music stream in 60 min or longer time to feel the position of the loud parts.

I think that it is feasible to detect the loudest part in a music stream by software when I set Recording Volume slider at the center. The center means that it is not too high to distort and too low to be noisy.

In this condition, the loudest part from line-in input is less than 0 db.
So all of the music stream can be digital data like wave file.
Thus a software can detect the loudest part to search the max value in the digital data.

This is the reason why I think it’s feasible by software.

If you know such software, please tell me.

I can’t transfer the Mini Disk digitally, because I don’t have such devices, MZ-RH1. And MZ-RH1 is just expensive in Japan.

It does not need to be exact - you just need to be in the right ball-park.
The important thing is that the recording level remains below 0 dB. (0 dB is full track height. If you exceed 0 dB it will cause irreparable distortion in the recording).

So what you need to do is find “a loud bit” from “a loud tape”, set your recording level to that, and allow a bit of “head room”. Don’t set the level to exactly 0 dB because you probably don’t have the absolute loudest part - set the recording level to around half the track height (about -6 dB), or if your tapes vary a lot, allow a bit more head room by setting the recording level a bit lower.

This is recorded too loud (irreparably damaged by distortion):
This is recorded too quiet (and will be unnecessarily noisy when amplified):
This shows a good recording level:
This is OK, but is dangerously close to distorting on the highest peaks:

1 Like

My idea is …

One point of recording music from line-in, tapes and MDs, at high quality is setting recording level near 0 db at the loudest part in the source. It enables to minimize the data loss by the digitization.

Doing this, it needs search the loudest part in the every tapes and MDs. But it’s a bother. So I’ve been searching the automated feature.

The recent replies tell me there are not such automated feature in Audacity, so I’ll record by kozikowski’s method, record with digital connection to buy the device, or steve’s method, record as high level as I can by finding “a loud bit” from “a loud tape”.


There is a an unfinished feature in the code that tries to set the recording level automatically during pre-record monitoring:

It would probably still result in clipping if you monitored a quiet part of a track that had very loud parts. A tool that is to take action about that would have to work during recording and dynamically compress the recording if needs be.

If you start a new topic in Adding Features to Audacity describing what you want then the topic will remain open for a month and anyone who wants to vote for your idea can do so in that topic. After the month, the votes will be transferred to our “feature requests” system.


You can already do something like that. Here’s how you could do it:

  1. Record the source with the recording level set low - low enough that you can be sure that it will not clip. The recording will be too quiet, but that does not matter.
  2. Open the Amplify effect and see how much Amplify says it needs to amplify by to bring the level up to 0 dB (This is the “Amplification (dB)” number).
    No need to apply the effect, we only need to see what that number is.
  3. Undo the recording (Ctrl+Z)
  4. Work out how much the input level needs to be increased by.
    Example: If the “Amplification (dB)” number read in step 2 said “20.0 dB”, and we want our peak level to be -6 dB, then we need to record 20.0 - 6.0 = 14.0 dB higher than our first recording.
  5. Increase the recording level by the amount calculated in step 4 (“14 dB” in the example), and re-record.

Problems with doing this:

  1. The recording process for a 1 hour tape will take at least 2 hours because you have to play the tape twice.
  2. Your recording level settings are probably not calibrated in dB, so neither you or Audacity will know how much to turn up the recording level in order to achieve a 14 dB increase.

Additional reason why Audacity cannot do this automatically:
Control of the recording level can occur in many places, depending on the hardware used. If you are recording through a mixing desk, then the level at the mixing desk needs to be set, but this is usually a mechanical slider control which cannot be adjusted by the computer. The same thing goes for most USB audio interfaces. If recording an analog source such as a non-USB cassette player or radio, then the playback volume needs to be set, and obviously that is something that the computer cannot do. The only control that the computer has is the ability to adjust the gain after it has gone through the A/D converter, but the gain needs to also be set correctly before the A/D converter.

I use a procedure similar to what Steve suggested. But, I shoot-for -3 to -6dB and I ONLY re-record if I get clipping.

After recording I run the Amplify effect, and if the Amplify defaults to 0dB (no gain) I assume the recording is clipped and I reduce the levels and start-over. (I have to guess how much to reduce the levels for re-recording but I don’t try to guess perfectly, I just want to avoid clipping the 2nd time.)

It enables to minimize the data loss by the digitization.

Digital recording (at 44.1kHz/16-bits or better) is so good (better than human hearing) that getting the levels ‘just-right’ is not that important.

ADCs & DACs are hard-limited to 0dB, so it is important that you don’t “try” to go over 0dB… It’s OK if the levels are a little too low, it’s not OK for the levels to be too high…

Pros typically record at -12 to -18dB (at 24-bits).

You may lose some theoretical resolution with MD at lower levels (when recording at 16-bits), but it’s a long way from theoretical quality loss to quality loss that you can hear (and MD/ATRAC is lossy) and with analog sources the analog noise limits the resolution so even at lower levels, digital is superior.

If your soundcard is noisy the noise may be more-audible at low levels and in that case it would be better to get a better-quality audio interface.

Thank you for pleiotropic answers.

I’ll try to record by steve’s 2nd method in this case as I try to find the relationship between the recording level and Amplification (dB) number.
If it’s hard time, I’ll study about what recording Engineer do and try to record with DVDdoug’s method.

Also I’m interested in Automatic Volume Feature, but sadly there’s no time.

Many thanks!