more questions on volume leveling

Get answers about using the current release of MediaMonkey for Windows.

Moderator: Gurus

cjcrawford
Posts: 60
Joined: Mon Jan 22, 2007 7:33 pm

more questions on volume leveling

Post by cjcrawford » Sun Mar 15, 2009 6:38 pm

Hi - I read the most recent, and informative, thread on volume leveling and still have a few more questions:

I assume that when synching to a device and "level volume' is checked under the player device properties (conversion) that the mp3 volume is permanantly changed only on the copy loaded to the remote device and not in the permanent library, correct?

I assume that all settings for leveling under options-playback are only for the player and don't touch the original file, correct?

Where does the default setting I see all over the place of 89db come from and what does this mean? Some threads have suggested setting to 95db? When does a file start to clip and is there any way of knowing it?

thanks all.

Chris

nohitter151
Posts: 23640
Joined: Wed Aug 09, 2006 10:20 am
Location: NJ, USA
Contact:

Re: more questions on volume leveling

Post by nohitter151 » Sun Mar 15, 2009 9:10 pm

cjcrawford wrote:Hi - I read the most recent, and informative, thread on volume leveling and still have a few more questions:

I assume that when synching to a device and "level volume' is checked under the player device properties (conversion) that the mp3 volume is permanantly changed only on the copy loaded to the remote device and not in the permanent library, correct?

I assume that all settings for leveling under options-playback are only for the player and don't touch the original file, correct?

Where does the default setting I see all over the place of 89db come from and what does this mean? Some threads have suggested setting to 95db? When does a file start to clip and is there any way of knowing it?

thanks all.

Chris
Yes.
Yes.
Tools | Options | Volume leveling
MediaMonkey user since 2006
Need help? Got a suggestion? Can't find something?

Please no PMs in reply to a post. Just reply in the thread.

MMan
Posts: 233
Joined: Wed Nov 22, 2006 7:20 pm
Location: Montclair, NJ

Re: more questions on volume leveling

Post by MMan » Mon Mar 16, 2009 8:45 am

cjcrawford wrote: Where does the default setting I see all over the place of 89db come from and what does this mean? Some threads have suggested setting to 95db? When does a file start to clip and is there any way of knowing it?
Your question was something that I was always curious about and caused me to do a little searching and between the Hydrogen Audio site and the Replay Gain site I found some interesting stuff. First, virtually all pop music today is already clipped in the mastering process see “Loudness Wars” Wikipedia page:

http://en.wikipedia.org/wiki/Loudness_war

Having said that, your question I believe relates to clipping as a result of volume leveling or analysis relative to what is on the original CD. It appears that 83db was the original Replay Gain target level and this was chosen based upon work done in the movie industry which I guess has more of a standard for masters than the audio industry, from the Replay Gain (http://replaygain.hydrogenaudio.org/contents.html) site:

“Having calculated a representative RMS energy value for the audio file, we now need to reference this to a real world sound pressure level. The audio industry doesn't have any standard for listening level, but the movie industry has worked to an 83dB standard for years.
What the standard actually states is that a single channel pink noise signal, with an RMS energy level of -20 dB relative to a full scale sinusoid should be reproduced at 83 dB SPL (measured using a C-weighted, slow averaging SPL meter). In simple terms, this means that everyone can set their volume control to the same (known, calibrated) gain.
ASIDE: This number (83dB SPL) wasn't picked at random. It represents a comfortable average listening level, determined by professionals from years of listening. That reference level of -20dB pink noise isn't random either. It causes the calibrated average level to be 20dB less than the peak level. In other words, it leaves 20dB of headroom for louder than average signals. So, if CDs were mastered this way, the average level would be around -20dB FS, leaving lots of room for the dramatic peaks which make music exciting. “


So the logical question is why is 89db what appears to be recommended level if 83db was the original standard based upon the work in the movie industry. That has to do with the “Loudness Wars”. Because today’s Pop CDs are mastered soooo loud, if targeted at 83db people wouldn’t like the playback relative to what they are used to. As a result, Replay Gain recommended a +6db preamp adjustment on playback to make Pop music sound “right”.

“Essentially the energy of the encoded signal is calibrated against a reference level, and the difference is stored is the Replay Gain adjustment value. The reference level set forth in the proposed standard is the SMPTE-sanctioned 83 dB SPL, representing a comfortable average listening level.

When Replay Gain is enabled, players are encouraged to supply a default pre-amp gain of at least +6 dB in addition to the Replay Gain adjustment. This is so pop music for example does not sound too quiet, but the pre-amp might be reduced or disabled for other types of music. madplay supports this with the existing -a (--amplify, --attenuate) option, which now has a default value of +6 dB when Replay Gain is enabled.”


83 + 6 = 89db. I think that’s how it got there! To the question of “when does clipping happen?” the answer seems to be very dependent upon the file being analyzed and the pre-amp setting or the player. I think that it is safe to say that there is probably little if any clipping using a target of 83db. At 89db it is likely that a “soft/quiet” recording (likely classical) with a wide dynamic range is getting clipped to some extent. However, this could be player dependent, see my questions below.

The research did raise a couple of questions on Monkey’s implementation of Replay Gain in volume leveling and analysis. The first is that the Replay Gain standard seems to recommend encoding the header with a Peak Level. It then recommends that players employ a check to see if the implementation of the Replay Gain adjustment would cause the peak level to be clipped. If it would cause clipping it recommends that the player impose a “hard limit” so that the Replay Gain adjustment gets reduced to avoid the clipping. My question is does Monkey do this or does it just clip the files on playback?

The next question relates to the player and the use of Volume Analysis. If all the files are analyzed to say 83db and the target playback level is set to 89db does that effectively cause a +6db “pre-amp” adjustment? If so, would make sense to analyze to 83db and then use the target playback level to adjust for the type of music. For example, use 83db for classical and 89db for pop so that there isn’t clipping but you still get the right feel of the music?

Any thoughts?

cjcrawford
Posts: 60
Joined: Mon Jan 22, 2007 7:33 pm

Re: more questions on volume leveling

Post by cjcrawford » Mon Mar 16, 2009 10:17 pm

Hi MMan and thanks for the research. Yest this topic begs many questions like yours and I'd love to hear from the MM designers (or the Wimamp designers? isn't that the core of MM?). I'm curious how volume is implemented in MP3 (or other) coding. If I remember my complex analysis from decades ago, once the waveform is sampled at a given frequency and given DAC resolution, the waveform can be reconstructed with the same precision at any volume by introducing a constant during the reconstruction. This implies that some kind of a replay gain tag should be inherent (and modifiable) in the MP3 coding and that the "original waveform" never needs to be touched now matter what the volume. But then... this is just speculation.

It would be nice to know in advance when a leveling or replay gain causes clipping.

Chris

MMan
Posts: 233
Joined: Wed Nov 22, 2006 7:20 pm
Location: Montclair, NJ

Re: more questions on volume leveling

Post by MMan » Tue Mar 17, 2009 8:00 am

First, a disclaimer. I am not a sound engineer and I never played one on TV, just a layperson with the internet. I think that you are correct that the Replay Gain algorithm doesn't touch the original Wave Form if being used with Volume Analyzing to get the constant. However, I do thinking that when the constant is applied during playback the wave form is changed, not just shifted. It has to do with the fact that the apparent loudness is different for different frequencies and various curves are used to adjust for this (see Replay Gain website for more detail). Said another way, to get the average or preceived "loudness" up or down, certain frequencies need to be adjusted up or down more than others. In and of itself, I don't think that this causes clipping. The clipping is caused when applying the constant (adjusted for the frequency issue) portions of the wave form go beyond 0db and therefore get clipped.

I agree that it would be cool if there were some kind of flag after the analysis which would indicate that the Replay Gain adjustment would result in clipping if played at the target 89db playback. That is why I think the concept in the Replay Gain original proposal to record the peak level during the analysis and use that to create a hard limit on the adjustment on playback to insure enough headroom to avoid clipping is pretty elegant. If there were a toggle for this feature it would be even better. Does anyone know if Monkey does this? Also, to cjcrawford's point on how volume leveling and analysis works for different file types, does Monkey apply the same algorithms to all file types? I was operating on the assumption it does.

Post Reply