Equalize Volume Across Many Tracks

Discussion forum for Amadeus users

Moderator: Martin Hairer

JonLawry
Posts: 7
Joined: Tue Dec 29, 2015 3:39 pm

Equalize Volume Across Many Tracks

Post by JonLawry »

I have been building an extensive library of recordings by Billie Holiday. These tracks came from many sources - CDs, vinyl LPs, YouTube videos. As you can imagine the variance in sound quality, EQ, volume level, etc. is all over the map. But I'd at least like to bring things to an "average" volume level so I don't have to constantly adjust the volume. I believe the Normalize function is intended for a single file, but I'm talking about a range of tracks using a batch processor. Is there any sort of way to automatically bring a range of files into some sort of RIAA-average? I know iTunes has its "Sound Check" playback feature, but is there any way to permanently apply this across a bunch of tracks? One thing that's good is that the dynamic range of these recordings made between 1933 and 1959 isn't that wide. Thanks!

CDJonah
Posts: 119
Joined: Sat Nov 11, 2006 12:10 am

Equalize Volume Across Many Tracks

Post by CDJonah »

I would look at normalizing to RMS average -- not perfect but I would
expect it to be pretty good as the music is probably somewhat similar in
dynamic range distribution (I had trouble doing something with Vivaldi,
which has little range with Mozart where some sections were quite loud
and others quite low). You have to decide what level the rms level
should be to avoid clipping.

Chuck
On 2/27/16, 11:34 AM, JonLawry wrote:
I have been building an extensive library of recordings by Billie Holiday. These tracks came from many sources - CDs, vinyl LPs, YouTube videos. As you can imagine the variance in sound quality, EQ, volume level, etc. is all over the map. But I'd at least like to bring things to an "average" volume level so I don't have to constantly adjust the volume. I believe the Normalize function is intended for a single file, but I'm talking about a range of tracks using a batch processor. Is there any sort of way to automatically bring a range of files into some sort of RIAA-average? I know iTunes has its "Sound Check" playback feature, but is there any way to permanently apply this across a bunch of tracks? One thing that's good is that the dynamic range of these recordings made between 1933 and 1959 isn't that wide. Thanks!





User avatar
Martin Hairer
Site Admin
Posts: 1975
Joined: Wed Nov 08, 2006 11:49 am
Contact:

Equalize Volume Across Many Tracks

Post by Martin Hairer »

Hi Jon, the Normalize function is available in the batch processor...
Regards,

Martin

--
HairerSoft
http://www.hairersoft.com/

JonLawry
Posts: 7
Joined: Tue Dec 29, 2015 3:39 pm

Post by JonLawry »

Hi Martin and Chuck,

Thanks for the replies. Yes, I see that "Normalize" is available in the batch processor, but wouldn't that just go in and normalize each track TO ITSELF? What I want is for the system to go in, look at each track, and adjust the volume so it's relative to a certain level - this one's too loud, so drop it down, this one's too soft, so bump it up, this one's OK, skip, etc. I think this is what iTunes' Sound Check does, right?

User avatar
Martin Hairer
Site Admin
Posts: 1975
Joined: Wed Nov 08, 2006 11:49 am
Contact:

Equalize Volume Across Many Tracks

Post by Martin Hairer »

What I want is for the system to go in, look at each track, and adjust the volume so it's relative to a certain level
Yes, that's precisely what normalisation does. Of course to get similar-sounding results this would assume that the different tracks have similar dynamic ranges, but seems to be the case from what you're saying. Regards,

Martin

--
HairerSoft
http://www.hairersoft.com/

JonLawry
Posts: 7
Joined: Tue Dec 29, 2015 3:39 pm

Post by JonLawry »

OK, thanks, I'll give it a try.

User avatar
Jim Edgar
Posts: 118
Joined: Mon Mar 01, 2010 7:20 am
Contact:

Equalize Volume Across Many Tracks

Post by Jim Edgar »

Martin - I was just noticing that with the Normalization command in AP, you can either Normalize using the Maximal Sample Value (dB PEAK) or by using Maximal RMS Power.  In other Normalization utilities I've used, the RMS Normalization function uses the RMS Average - which I've found to be a much more useful in terms of matching recordings (audiobook mastering, for example). 

I see I could probably end up in the same place using the Maximal RMS and using a negative percentage value, but is there a way to directly input a desired RMS Average for Normalization?
Jim Edgar

jimedgarvoices@gmail.com (jimedgarvoices@gmail.com) - @jimedgarvoices
http://www.jimedgarvoices.com  -  http://about.me/jimedgar




On Sat, Feb 27, 2016 at 1:20 PM, JonLawry <forum2mail@hairersoft.com (forum2mail@hairersoft.com)> wrote:
OK, thanks, I'll give it a try.




JonLawry
Posts: 7
Joined: Tue Dec 29, 2015 3:39 pm

Post by JonLawry »

In case anyone's interested I did a little further research on this. Apparently I'm not the only one wanting to arrive at some sort of "average" standard for loudness levels. The problem is that what's "average" can depend on different listeners, at different ages, different styles of music, etc. As for Amadeus Pro, my best results have been to normalize using RMS power at peak level -8 dB. However, there is a new standard called EBU-R128 that's being developed for broadcast audio - a lot of it in reaction to the so-called "loudness war" between TV programs and commercials. The standard seems to have been pretty intelligently arrived at, looking carefully at how loudness is actually perceived by the human ear. I certainly don't understand it all, but apparently there is serious consideration to using it at the final stages for digital audio authoring for DVDs, etc. Perhaps this might end up being a "normalization" option in future versions of Amadeus Pro? The best discussion I found was here http://forum.doom9.org/showthread.php?t=171472.

Ray Foret Jr

Equalize Volume Across Many Tracks

Post by Ray Foret Jr »

Well, for what it's worth, my understanding is that the acceptible level for mastering NLS talking book audio files is no higher then -16DBSPL using RMS, I think. Not sure about that one whether it's at peak levels or RMS. For me, I like to get things to just the point where the peak level is not louder than Apple Voice Over at it's highest volume. That's kind of a crazy way to do it, but, that's sort of the way I proceed for now. Your suggestion of -8 using RMS seems quite sensible to me.
Sent from my Mac, The Only computer with full accessibility for the blind built-in
Sincerely, The Constantly Barefooted Ray,Still a very happy Comcast XFinity Voice Guidance, Mac, Verizon Wireless iPhone6+ and Apple TV user!




On Feb 27, 2016, at 4:28 PM, JonLawry <forum2mail@hairersoft.com (forum2mail@hairersoft.com)> wrote:
In case anyone's interested I did a little further research on this. Apparently I'm not the only one wanting to arrive at some sort of "average" standard for loudness levels. The problem is that what's "average" can depend on different listeners, at different ages, different styles of music, etc. As for Amadeus Pro, my best results have been to normalize using RMS power at peak level -8 dB. However, there is a new standard called EBU-R128 that's being developed for broadcast audio - a lot of it in reaction to the so-called "loudness war" between TV programs and commercials. The standard seems to have been pretty intelligently arrived at, looking carefully at how loudness is actually perceived by the human ear. I certainly don't understand it all, but apparently there is serious consideration to using it at the final stages for digital audio authoring for DVDs, etc. Perhaps this might end up being a "normalization" option in future versions of Amadeus Pro? The best discussion I fou nd was here http://forum.doom9.org/showthread.php?t=171472.

User avatar
Jim Edgar
Posts: 118
Joined: Mon Mar 01, 2010 7:20 am
Contact:

Equalize Volume Across Many Tracks

Post by Jim Edgar »

ACX specifies recording/mastered levels to fall between -23 and -18 RMS Average.
Jim Edgar

jimedgarvoices@gmail.com (jimedgarvoices@gmail.com) - @jimedgarvoices
http://www.jimedgarvoices.com  -  http://about.me/jimedgar




On Sat, Feb 27, 2016 at 2:31 PM, Ray Foret Jr <rforet7706@comcast.net (rforet7706@comcast.net)> wrote:
Well, for what it's worth, my understanding is that the acceptible level for mastering NLS talking book audio files is no higher then -16DBSPL using RMS, I think.  Not sure about that one whether it's at peak levels or RMS.  For me, I like to get things to just the point where the peak level is not louder than Apple Voice Over at it's highest volume.  That's kind of a crazy way to do it, but, that's sort of the way I proceed for now.  Your suggestion of -8 using RMS seems quite sensible to me.


Sent from my Mac, The Only computer with full accessibility for the blind built-in


Sincerely, The Constantly Barefooted Ray,

Still a very happy Comcast XFinity Voice Guidance, Mac, Verizon Wireless iPhone6+ and Apple TV user!





On Feb 27, 2016, at 4:28 PM, JonLawry <forum2mail@hairersoft.com (forum2mail@hairersoft.com)> wrote:

In case anyone's interested I did a little further research on this. Apparently I'm not the only one wanting to arrive at some sort of "average" standard for loudness levels. The problem is that what's "average" can depend on different listeners, at different ages, different styles of music, etc.  As for Amadeus Pro, my best results have been to normalize using RMS power at peak level -8 dB. However, there is a new standard called EBU-R128 that's being developed for broadcast audio - a lot of it in reaction to the so-called "loudness war" between TV programs and commercials.  The standard seems to have been pretty intelligently arrived at, looking carefully at how loudness is actually perceived by the human ear. I certainly don't understand it all, but apparently there is serious consideration to using it at the final stages for digital audio authoring for DVDs, etc. Perhaps this might end up being a "normalization" option in future versions of Amadeus Pro?  The best discussion I fou
nd was here http://forum.doom9.org/showthread.php?t=171472.







Garth Humphreys

Equalize Volume Across Many Tracks and a Feature request.

Post by Garth Humphreys »

Just wanted to add my voice to this feature request, that being the addition of a tool / normalisation option that allowed for normalising to an LUFS value. If I understand it correctly, it is the RMS value for the whole of the audio, however any audio that is below a certain threshold is ignored in the calculation so that portions of silence don’t effect the result. In addition, there is some EQ curve applied to the audio prior to measuring to make it more relevant to the perceived loudness. Probably should have done some more research prior to writing but I think that’s the basic principle. Seems to make a lot of sense, and I’d love to see it as an option in Amadeus Pro.

Regards,
On 28 Feb 2016, at 8:28 AM, JonLawry <forum2mail@hairersoft.com> wrote:

In case anyone's interested I did a little further research on this. Apparently I'm not the only one wanting to arrive at some sort of "average" standard for loudness levels. The problem is that what's "average" can depend on different listeners, at different ages, different styles of music, etc. As for Amadeus Pro, my best results have been to normalize using RMS power at peak level -8 dB. However, there is a new standard called EBU-R128 that's being developed for broadcast audio - a lot of it in reaction to the so-called "loudness war" between TV programs and commercials. The standard seems to have been pretty intelligently arrived at, looking carefully at how loudness is actually perceived by the human ear. I certainly don't understand it all, but apparently there is serious consideration to using it at the final stages for digital audio authoring for DVDs, etc. Perhaps this might end up being a "normalization" option in future versions of Amadeus Pro? The best discussion I fou
nd was here http://forum.doom9.org/showthread.php?t=171472.



JonLawry
Posts: 7
Joined: Tue Dec 29, 2015 3:39 pm

Post by JonLawry »

I think the issue, and I'm certainly no expert on anything here, has to do with sound purity. In several of the articles I read, the suggestion was to do the EQ, processing, etc. at the "front end" of the recording. This makes sense as it is part of the art of recording. However, when consumers are getting compilation CDs (as I have in my Billie Holiday project) with wide-ranging levels, it's a bit disconcerting. I guess the value of the EBU-R128 standard is that it has an imperceptible effect on the "front end" aspects of the recording and brings it to a consistent level. I'm not so sure it could do this in theory because technically it's an EQ-related thing. I'm having good results with RMS -8, so I guess that will be my standard, but I'd sure like to hear how the EBU-R128 sounds in comparison.

Tere Codina

Can not open Amadeus

Post by Tere Codina »

Hi Martin and all,

I had not worked on Amadeus for a while and now I can not open the program.

I have El Capitan. As far as I remember, after purchasing Amadeus on the Mac
appstore I downloaded the version of the web.

What can I do?

Thank you very much.

User avatar
Martin Hairer
Site Admin
Posts: 1975
Joined: Wed Nov 08, 2006 11:49 am
Contact:

Can not open Amadeus

Post by Martin Hairer »

Please make sure you are using the latest version (2.2.2). OS X 10.9.5 broke the licensing function in versions 2.1.7 and earlier, this is why they would revert to their unlicensed state. Regards,

Martin

--
HairerSoft
http://www.hairersoft.com/

JEGG
Posts: 122
Joined: Sat Feb 24, 2007 6:33 pm

Post by JEGG »

Keep in mind that these figures are for replay. The reason you adjust your submitted material to those standards (below) is so it won't be modified in unintentional or unacceptable ways on replay.

There are various LKFS standards in use.

Broadcast is generally from -22-24 LKFS with true peaks of -1 or -2 dB.

Streaming is different with several standards.

Generally, iTunes and YouTube is about -16LKFS.

Spotify has hugely reduced dynamic range, perhaps -6 or -9 LKFS, memory is not serving me well!

An AES paper recommends -16-19 LKFS for streaming. This assures a bit more uniform listening experience for those on portable devices and in cars, and the wider range of -19 is a bit more useful for classical repertoire.

The idea re: YouTube and iTunes, is that you should upload at these specifications in to ensure that there won't be modifications on replay. And a live classical recording should be at the broadcast standard, or it may be modified automagically or just rejected

If you're providing material on a website, it's helpful to stick with those specifications to ensure you material is perceived in the same way as material fromiTunes or YouTube. It's the same as before, no one wants to be "softer" (or louder) than iTunes or YouTube.

Spoken word recordings, consider: For mixed (as in VO withing a music program, the VO should be lower that the LKFS. For VO only material, the level should be at the nominal LKFS level. See "Short Term LKFS."

To comply with these standards, in Amadeus you can use various plugins, such as those from Nugen. Or other products, such as Rx Advanced, incorporate this function.

The Sound Check function in iTunes is the playback function that ensures a wide variety of material will be perceived at approximately the same level.

Post Reply