Audio Level Meter vs Audio Output Levels

Jamie C

That sound card is a cheap sound card.
Therefore I would target your output at a maximum of -3db, even lower would be better. This is to have a ceiling below 0db.
Going above 0db will diminish a quality output.
 
Using MP3gain to normalize your files sounds good in theory.
And can improve things in some music libraries in which recordings haven't been well managed whilst recording.

The downside is a loss of dynamics.
 
DJSTU said:
The downside is a loss of dynamics.
AFAIK MP3Gain applies the same level to the whole file - there won't be any loss in dynamics.
Loss of dynamics occurs when you useAutoAmp or other AGC which adjusts sound level on the fly.
 
MP2 encoders were made for commercial use and are highly editable.
MP3 encoders were made for domestic use and cannot be finely edited, as the need was not perceived to be needed.
 
As they (MP3Gain developers) say, MP3 frames have "level" field for frames, that's why it's easy to adjust MP3 volume and not lose quality. File is not re-encoded.
 
Dmitry, Stu,

We have been having zero luck with our ReplayGain tags fixing the audio levels.  Today I made a series of audio test files to try and identify the problem.  Here is what I fpund:

1. Audio file with ReplayGain tags - low level, not peaking anywhere near the peak of onscreen meter, not acceptable
2. Audio test file with 400Hz tone created with Audacity at a level of "zero" or 1.0 was played and worked correctly in terms of audio level and the RadioBOSS onscreen meter level.
3. Audio file processed with MP3trim to "zero" level - acceptable level, peaking towards peak of onscreen meter, more than acceptable

The file we tested showed a ReplayGain figure of -8.6dB on RadioBOSS screen, and in general the audio level sat pn the bottom 1/3 of the RadioBOSS onscreen meter.

Is there any chance that RadioBOSS has a "double error" in it's code ?  So +8.6dB to -8.6dB = 17.2dB error in how it has adjusted the audio level ?

As I said, processing the MP3 audio file with MP3trim set to "zero" level completely fixed the problem AND it was clearly a lot louder in the speakers compared to the MP3 file with the ReplayGain tags.

Any suggestions ?


Thanks, Jamie C.
 
Dmitry,

Just discovered that I am running version 4.8.1 so I will upgrade it tomorrow and re-run the tests ....

Sorry for overlooking this.


Thanks, Jamie C.
 
Yes, the latest version 4.8.3 has some fixes regarding ReplayGain handling for mp3 files.
I don't think there's an error calculating the resulting sound level. But if you experience problems with some files (too loud or too quiet) - please send one of those to support@djsoft.net
 
Dmitry,

Yep, version 4.8.3.0 fixed the problem so that is good ... thanks.

All music tracks above came out okay as you suggested.  However the 400 Hz tone tests were obscure and did not come up to a consistent normalised level - probably because ReplayGain and MP3trim are both designed for music and not steady state audio tones.

However, can you explain why you decided on the protocol of having to add +6dB of pre-amp gain rather than just leaving it at 0dB ?    (> Settings > Miscellaneous > ReplayGain)

I'll will now create an internal procedure for us on how we process our music files before using them on RadioBOSS and a list of the RadioBOSS settings to suit.


Thanks, Jamie C.
 
radiodungog said:
However, can you explain why you decided on the protocol of having to add +6dB of pre-amp gain rather than just leaving it at 0dB ?    (> Settings > Miscellaneous > ReplayGain)
By default ReplayGain normalizes tracks to -6dB level. When your library has tracks with and without ReplayGain - tracks without RG will play louder (as they are at 0dB level). We compensate for it by adding +6dB for ReplayGain tracks.

I just looked at Winamp, and they did a similar thing: they left RG tracks at 0dB, but change non-RG tracks level to -6dB.
 
Dmitry,

Okay, if those two figures are known why not do the matching in the software code and leave a simple setting of 0dB for the user to access ... so in essence the user adjustments end up being a personal "fine tune" rather than a system interface config ?


Here is some interesting stuff for ReplayGain in terms of a reference level made of pink noise.  This file is the official reference wav file, it's track gain should be +6dB for 89dB reference level.  It may be useful for people to play with as a reference level for an end to end audio line-up.

http://replaygain.hydrogenaudio.org/proposal/ref_pink.wav


Thanks,  Jamie C.
 
radiodungog said:
Okay, if those two figures are known why not do the matching in the software code and leave a simple setting of 0dB for the user to access ... so in essence the user adjustments end up being a personal "fine tune" rather than a system interface config ?
There's such a setting: Gain slider on the FX tab controls the final playback level.
ReplayGain levels are left for those who want to have more control. Eg. their ReplayGain scanner's target level could be other than -6dB.

Anyway, presence of those settings doesn't hurt anyone, and I'm absolutely sure that there's a percentage of RB users who may want to adjust those.
 
Back
Top