Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: A Ripper that will provide "Level Input" (Read 7777 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

A Ripper that will provide "Level Input"

I did not know how to word what I am looking for in a one line Topic.  I am wondering if there is a Ripping device that will rip all songs from various CDs at a similar level.  It is frustrating when you use shuffle mode and you go from one song to the next and get your head blown off by the varying levels of output. 
I hope that there is some sort of software to get control of this phenomenon.  I would hate to blow out my speakers not to mention my ears!

A Ripper that will provide "Level Input"

Reply #1
Please read the Replay Gain article on the HA.org wiki. Perhaps that's what you are searching for.
Additionally, Replay Gain on wikipedia.org has even more information.

A Ripper that will provide "Level Input"

Reply #2
Thanks.  But that is only for replay, I want to possibly encode all of my songs at perhaps 89db as was mentioned in the article that you attach.  That way all of your devices will recieve the benefit.

A Ripper that will provide "Level Input"

Reply #3
MP3Gain?

A Ripper that will provide "Level Input"

Reply #4
Thanks.  But that is only for replay, I want to possibly encode all of my songs at perhaps 89db as was mentioned in the article that you attach.  That way all of your devices will recieve the benefit.

As you can read in the links I provided there are several ways to apply the Replaygain concept.
Many of them are meta-data based, this means that the audio is analyzed and some kind of tag is added to tell the (mostly software) player to adjust the loudness. Most hardware players ignore these kind of replaygain tags.
Another option is more or less a hard coding. for example WaveGain adjusts the loudness of a wavefile to match the desired level. If you encode your lossy files from this wave file, the resulting files will also be levelled.
An intermediate application is mp3gain. This one analyzes the audio and changes the 'global gain' field inside the mp3 files. Any compliant player (either software or hardware) will playback at the correct loudness.

Which specific implementation you want to choose depends on your needs. A possibility to automate the process is to use EAC in combination with  REACT and have it calculate replaygain values right after the CDDA extraction.

A Ripper that will provide "Level Input"

Reply #5
As you can read in the links I provided there are several ways to apply the Replaygain concept.
Many of them are meta-data based, this means that the audio is analyzed and some kind of tag is added to tell the (mostly software) player to adjust the loudness. Most hardware players ignore these kind of replaygain tags.
Another option is more or less a hard coding. for example WaveGain adjusts the loudness of a wavefile to match the desired level. If you encode your lossy files from this wave file, the resulting files will also be levelled.
An intermediate application is mp3gain. This one analyzes the audio and changes the 'global gain' field inside the mp3 files. Any compliant player (either software or hardware) will playback at the correct loudness.

Which specific implementation you want to choose depends on your needs. A possibility to automate the process is to use EAC in combination with  REACT and have it calculate replaygain values right after the CDDA extraction.
[/
Thanks, but that seems confuding

A Ripper that will provide "Level Input"

Reply #6
Well here's what I do:

1) Rip to hard drive as lossless FLAC via EAC
2) Apply replaygain scanning/tagging in Foobar2000
3) batch convert to MP3/AAC/OGG/whatever in Foobar2000, enable replaygain processing during conversion to normalize audio levels

This method obviously works best if you're keeping two copies of your collection, a lossless archive for backup/PC playback and a separate lossy collection for portable use.

A Ripper that will provide "Level Input"

Reply #7
Or just use iTunes with Sound Check turned on. Or like I sad before use mp3gain.

It's not a hard problem to solve nowadays, but you may still need to jump through some hoops. The solutions are very easy once you know how to do them. It helps if we know what your current ripping process is, to figure out what is going to be the least complicated solution.

A Ripper that will provide "Level Input"

Reply #8
I have always used WMP to convert to WMA.  However, I am open to using EAC along with the latest lame encoder and whatever else is necessary since  I am ripping my entire collection of CDs (about 500).
I just need a little help and guidance to get started.
Thanks, Dan

A Ripper that will provide "Level Input"

Reply #9
This is can be difficult problem.  Replay gain (and/or it's variations) is usually the best solution.  (But, some players and some formats don't support Replay Gain.)


Some "background" information...
-------------------------------------------------------------------------------------------------------------------
Human perception of loudness is very complicated.  Replay Gain takes human perception into account, and it works fairly well if you don't mix radically different styles of music, but there is no perfect solution.

And, since Replay Gain works at playback-time, it doesn't muck with the audio-data in your file, so it doesn't require re-coding, which could affect audio quality.

-------------------------------------------------------------------------------------------------------------------
Most CDs are already normalized.  That is, their peak levels are at (or near) the digital maximum (normally called 0dBFS = Zero Decibels Full Scale).  Some songs are intended to be quieter than others, so some "quiet" tracks on the CD might not reach 0dB.  And of course, some styles of music are intended to be louder than others...  A heavy metal band is usually louder than a folk group. 

It's a simple matter to normalize all of your files with an audio editor, and I think EAC has an option to normalize during ripping.  But, making the peaks levels equal will not make your files sound equally loud.

A slightly better solution is to measure the average (or RMS) of your songs, and then adjust your songs so that they all have the same average level.  This does not work as well as the analysis used by Replay Gain.  And, you cannot increase the levels to make them match.  You'd have to reduce the level of most tracks, since the peak levels are already at,  or near, the maximum.  (Replay Gain does the same thing...  It adjusts-down most tracks.)

And of course, you could adjust the level manually (with an audio editor), but  this would "take forever".


------------------------------------------------------------------------------------------------------------------
Also, most modern "Pop" CDs are highly dynamically-compressed to make the "quiet parts" of a song just as loud as the "loud parts".  And therefore, most pop CDs already have fairly-equal loudness.  With dynamic compression, you can increase the average level without increasing the peak level, and this makes CD sound louder.  (You can look-up "loudness wars".) 

In fact, you could add (dynamic) compression (with an audio editor) to make all of your songs sound louder and more equal.  Radio stations add compression for this reason.  But, it's a non-linear process and it can make music sound boring.  This is why most of us are annoyed about the "loudness wars"....  Most modern music (and re-masters) sounds bad and it's boring!


-------------------------------------------------------------------------------------------------------------------
Back to perception...

Our ears are more sensitive to mid-frequencies than to low or high frequencies.  If you play a 1kHz tone and a 50 Hz tone at equal levels, the 1kHz tone will sound louder.  And, it gets more complicated....  The difference is greater at low volume levels...  This is why the "loudness" switch on your receiver boosts bass at low listening levels, but has little or no effect at high levels.

Classical music has very-quiet and very-loud parts.  So, matching the loudness of a classical piece to a pop-song is practically impossible.  If you make the peaks equal, the pop song will sound much louder.  If you make average-levels equal, the loud parts in the classical piece will blow-out your ears!  And, if you asked two (or more) people to set the volume control for equal volume, each person would have a different opinion.

Imagine a recorded conversation between two people.  If you edit-out one of the speakers, the average volume level will be cut in half (approximately).  However, the perceived loudness will not change.

-------------------------------------------------------------------------------------------------------------------
Radio stations use several techniques to keep "constant volume".

- First, they stick to one style/genre of music.

- They use limiting & compression to make "everything loud".

- They "manually" adjust the volume for each song. (In the old days, the DJ had his hand on the volume control. Now, I assume it's done with something similar to a Replay Gain setting in their computer.)

- They program the music so that they don't play a slow-quiet song immediately after a loud-fast song. They will play a medium-song in the middle (they call this a "music curve"), or there might be some talking or a commercial in-between.

A Ripper that will provide "Level Input"

Reply #10
I listen to a variety of Rock (classic, progressive,Metal), and what I find is that older albums seem to be recorded at a lower loudness.  As an example; my Rush Fly By Night album is much softer than their newer stuff such as Snakes and Arrows.
This is not a problem when I listen to an album one at a time, but when I use the shuffle mode for all tracks on my Mp3 player it  can be annoying.
If there is no way to encode it some way without loss of quality, I will then just suck it up.

A Ripper that will provide "Level Input"

Reply #11
If you're listening to your music in a lossy format then you're already suffering from a loss of quality.

I'm at a loss as to what you really want.  All the posters to this discussion have done a remarkable job and have left little (if nothing) to add.

A Ripper that will provide "Level Input"

Reply #12
Duh....... I'm looking for the best sound Period....I'm willing to sacrifice some loss to lossy, but if there is a continued loss than I'm not for it.  Simple as that!


A Ripper that will provide "Level Input"

Reply #14
What do you think?


A Ripper that will provide "Level Input"

Reply #16
Eeeeeasy there, pardner.

moodyda: If I understand you correctly, you're worried that either you'll need to reencode all your WMAs to some other lossy format that supports ReplayGain, or rerip. I think you're basically right; if you've got WMAs, reripping is probably going to be your best bet to take advantage of ReplayGain. I'm not aware of any other loudness eq for WMA.

Again, if you have MP3s, there are a bazillion ways to get them ReplayGained in a manner compatible with your player of choice.

A Ripper that will provide "Level Input"

Reply #17
Thanks.  From what I understand replay gain is a tool that would be only compatible with certain players and Mp3 gain is a process attached to each file that will work with any player.  I suppose some prefer one over the other, I was just attempting to get some input from those of you who have used these in the past.

A Ripper that will provide "Level Input"

Reply #18
Have a look at Foobar2000. It allows you to permanently alter the internal data of the MP3 file in accordance with the value recommended by ReplayGain.

I wouldn't worry about any additional loss of quality personally as any further losses incurred are going to be miniscule to the point of insignificance compared to using a lossy format in the first place.

Cheers, Slipstreem. 

 

A Ripper that will provide "Level Input"

Reply #20
Is this the reason why people are hesitant to say there is no quality loss when using mp3gain?
http://www.hydrogenaudio.org/forums/index....st&p=578702


Hmm, maybe. Or maybe there's the technical possibility that some signal may sink into the noise, even though it's almost certainly inaudible.

Putting scalefactors out of bounds doesn't seem to be a problem with mp3, and I think mp3gain wraps around if it happens (the tags it puts in record the max and min scalefactors show the there's a huge range downward left on all the files I've looked at), and you wouldn't wish to increase gain of loud-mastered albums, as they're already at the limit of clipping.

Trying to find an album with modern loud mastering that I had applied mp3gain to, here's an example:

EAC/REACT2 // LAME 3.97 -V2 --vbr-new --noreplaygain --nohist
Scissor Sisters / Ta-Dah (with Bonus Track)
MP3GAIN_ALBUM_MINMAX = 135,249
MP3GAIN_UNDO = +006,+006,N

Basically, it has had -9.0 dB of Album Gain applied by mp3gain, so originally had max scalefactors of 255 on the album and could not be made louder by mp3gain (target volume would have to be 98.1 dB to try this).

However, the lowest scalefactor was originally 141, and can go as low as 0 before it's no longer allowed and would wrap to 255 instead.

Think about it: 256 x 1.505 dB steps is about over 385.3 dB of range, which is just huge!

I made a copy and applied Constant Gain of +21 dB then reverted (peak was over 4.0). Comparing the original to the copy, fb2k bit compare found no differences

As for losing quality through the psymodel, things like the ATH model tend to keep more information in louder music, so loud music turned down after encoding should have more quiet parts encoded than a file that is pre-scaled before encoding by LAME. Most other things are relative, like signal to mask ratios, so should remain unaffected.

I have no qualms about using mp3gain, or using pre-scaling (usually through fb2k's converter, via 24-bit input to LAME, but sometimes through LAME --scale 0.xx)
Dynamic – the artist formerly known as DickD

A Ripper that will provide "Level Input"

Reply #21
FWIW, I've been playing around with pre-scaling via batch with Lame and later with NeroAacEnc via Sox.  I wanted to see how my method compared with fb2k's and ran into something quite interesting.

It appears that using fb2k to apply replay gain when decoding flac to wave and then encoding the wave to mp3 (again using fb2k) resulted in a different file than using fb2k to apply replay gain when when transcoding the flac to mp3 without creating the interim wave.

When comparing the wave created by fb2k and one created by my batch, the only difference was samples with an amplitude no greater than 1.  I would imagine they were only rounding differences (EDIT: I don't think it's dither, but I could be wrong).  I'm still using 0.9.4.2, BTW.

A Ripper that will provide "Level Input"

Reply #22
Yikes! I'm seeing the same thing. Converting to LAME -V6 (3.97 IIRC), between letting foobar do the whole conversion vs converting to WAV first and then converting to MP3, with RG track gain applied, I see differences in some cases of about 0.1 (-20dbFS). However, these are only transients; the majority of samples are well below -40dbFS in difference.

I don't quite see how such a difference could be quantization-related. Maybe it's an ATH at work?

A Ripper that will provide "Level Input"

Reply #23
What about "Keep lossless sources at original bit depth" option?

Added: I mean, when I turn this option off, both mp3 files (flac->mp3 and flac->wav->mp3) became the same.