INTERACT FORUM

Please login or register.

Login with username, password and session length
Advanced search  
Pages: 1 2 3 [4] 5   Go Down

Author Topic: NEW: Improved audio analysis and volume leveling (R128)  (Read 173538 times)

MarkCoutinho

  • Regular Member
  • Citizen of the Universe
  • *****
  • Posts: 676
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #150 on: October 28, 2013, 11:06:45 am »

One more guy who does it just like Jack, BigCat etc. I always use MP3Gain first and then MC. Would be great if I could throw MP3Gain away and use just MC for this purpose. Remember, like Jack said: Will MC19 be able to modify the files so they can play normalized outside of MC?
I asked this question a couple of months ago and for the answer I was referred to another topic. However, I'm sorry to say: I still can't figure out what the answer to my question is. Must be because English is not my native language - sorry for that.

So let me rephrase my question, it's actually a simple one:
If I use MC's new analyze-tool for all my mp3's, will these files have the same volume outside of MC (for instance in my car)? If yes, that would be great - then I can throw good old MP3Gain away.
Logged
Mark Coutinho
Dutch Top 40 collector of lyrics, sleeves and bios

Vocalpoint

  • Citizen of the Universe
  • *****
  • Posts: 2007
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #151 on: October 28, 2013, 11:35:30 am »

If I use MC's new analyze-tool for all my mp3's, will these files have the same volume outside of MC (for instance in my car)? If yes, that would be great - then I can throw good old MP3Gain away.

No. Since MC does not alter the physical makeup of the MP3 file.

Whether using the older ReplayGain with v18 or prior - OR the new system within v19 - the end result is the same - MC writes it's volume/level data to tags within the files themselves. That tag data can then be picked up by apps (or devices) that can handle ReplayGain or similar - but if you car has no ability to interpret these tags - then the file will playback as originally rendered.

Hope that helps.

VP
Logged

MarkCoutinho

  • Regular Member
  • Citizen of the Universe
  • *****
  • Posts: 676
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #152 on: October 29, 2013, 04:11:41 pm »

No. Since MC does not alter the physical makeup of the MP3 file.

Whether using the older ReplayGain with v18 or prior - OR the new system within v19 - the end result is the same - MC writes it's volume/level data to tags within the files themselves. That tag data can then be picked up by apps (or devices) that can handle ReplayGain or similar - but if you car has no ability to interpret these tags - then the file will playback as originally rendered.

Hope that helps.

VP
Thanks VP. I guess I am stuck to MP3Gain, because in my car I play the mp3's from USB-stick and the player itself is not that fancy. Bummer..
Logged
Mark Coutinho
Dutch Top 40 collector of lyrics, sleeves and bios

mwillems

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 5233
  • "Linux Merit Badge" Recipient
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #153 on: October 29, 2013, 05:59:49 pm »

Thanks VP. I guess I am stuck to MP3Gain, because in my car I play the mp3's from USB-stick and the player itself is not that fancy. Bummer..

When writing your files to a handheld device (MP3 player or USB stick), you can tell JRiver to "apply DSP" to the file (including volume leveling).  So if your goal is to modify the files that you write to a portable device to all be roughly the same volume, you can do that with JRiver (I just did it myself over the weekend for a long car trip). It's just not part of the normal file analysis process (to avoid altering the files).
Logged

Vocalpoint

  • Citizen of the Universe
  • *****
  • Posts: 2007
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #154 on: October 29, 2013, 09:36:11 pm »

When writing your files to a handheld device (MP3 player or USB stick), you can tell JRiver to "apply DSP" to the file (including volume leveling).  So if your goal is to modify the files that you write to a portable device to all be roughly the same volume, you can do that with JRiver (I just did it myself over the weekend for a long car trip). It's just not part of the normal file analysis process (to avoid altering the files).

That's right. Completely forgot about that!

VP
Logged

MarkCoutinho

  • Regular Member
  • Citizen of the Universe
  • *****
  • Posts: 676
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #155 on: October 30, 2013, 04:09:56 am »

When writing your files to a handheld device (MP3 player or USB stick), you can tell JRiver to "apply DSP" to the file (including volume leveling).  So if your goal is to modify the files that you write to a portable device to all be roughly the same volume, you can do that with JRiver (I just did it myself over the weekend for a long car trip). It's just not part of the normal file analysis process (to avoid altering the files).
Now that sounds exactly what I want! However, where is this 'appy DSP' option situated? I always just use the 'Rename, move & copy files'-option to copy files to my USB-stick. But there is no DSP-thing there.
Logged
Mark Coutinho
Dutch Top 40 collector of lyrics, sleeves and bios

Vocalpoint

  • Citizen of the Universe
  • *****
  • Posts: 2007
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #156 on: October 30, 2013, 07:19:33 am »

Now that sounds exactly what I want! However, where is this 'appy DSP' option situated? I always just use the 'Rename, move & copy files'-option to copy files to my USB-stick. But there is no DSP-thing there.

I am now confused as well. I always thought you could only use the "Apply DSP" when burning a CD or using the Convert Format process...

I do not understand how one can "write DSP" to a file simply by moving it to a portable device. I load my iPod all the time but have never seen an "apply DSP" option.

Maybe mwillems can enlighten us?

VP
Logged

mwillems

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 5233
  • "Linux Merit Badge" Recipient
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #157 on: October 30, 2013, 07:40:09 am »

I am now confused as well. I always thought you could only use the "Apply DSP" when burning a CD or using the Convert Format process...

I do not understand how one can "write DSP" to a file simply by moving it to a portable device. I load my iPod all the time but have never seen an "apply DSP" option.

Maybe mwillems can enlighten us?

VP

The trick is to tell JRiver that the usb stick is a handheld device.   I'm not near JRiver right now so I can't describe the exact menu sequence, but in the options for configuring handheld devices there's an option to specify a file path as a handheld device.  The option probably exists because some MP3 players aren't detected as MP3 players and just show up as a drive (e.g. "N:\").  So you have to go into that option list and specify the drive letter for your USB stick as a handheld.  At that point JRiver will treat it as an MP3 player that can be synced.  You can then set up which files you want to sync through a number of different methods (drag and drop, send to, creating a playlist, etc.).  Once you've lined up all the files you want to transfer, you need to go into "sync options" or "sync details" before you press the sync button. In that options menu there is an "apply DSP" checkbox that will apply DSP (including volume leveling) to the files during the file transfer to the device.

It may sound complicated, but if you've used JRiver with a handheld device before it should be pretty easy.  I originally started using it before "apply DSP" was a thing to manually back up files to an external harddrive.  By configuring the harddrive as a handheld, I can take advantage of JRiver's "only sync files that have been changed" function to reduce the amount of file transfer that needs to happen for the manual backup. 
Logged

faster

  • Junior Woodchuck
  • **
  • Posts: 54
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #158 on: November 01, 2013, 08:39:22 am »

Hi Matt,

I am using current version of Media Center 19.0.66.

I think volume leveling is not working as expected. Now I have the album "Muddy Waters" - "Folk Singers" in two Versions. First one is an EAC CD Rip in 16Bit/44Khz FLAC. The other, newer one is 24Bit/192Khz AIFF from HD Tracks. Both albums analyzed with JRiver. Putting the same single Track from each album to a Playlist named "TestGain"). Adaptive Volume (Peak Level Normalize) and Volume Leveling is enabled.

- On playback the track form CD sounds much louder then track from HD Tracks. Why?

- And why is volume leveling in Audio Path the same for both tracks?


Tags for Track from CD:
Free Lossless Audio Codec (FLAC)
  44,1 kHz Sample Rate
  16 Bits Per Sample
  2 Channels

  Min Block Size = 4096
  Max Block Size = 4096
  Min Frame Size = 725
  Max Frame Size = 11669
  Padding Block = 7820 bytes

# of Pictures = 1
  Type 3 (21529 bytes)

SeekTable Block (576 bytes):
  # of points = 32

Vorbis Comment Block (621 bytes):
  Vendor String = S
  TOTALDISCS = 1
  TOTALTRACKS = 14
  TOOL NAME = Media Center
  TOOL VERSION = 19.0.66
  REPLAYGAIN_TRACK_GAIN = +0.52 dB
  BPM = 62
  VOLUME LEVEL (R128) = -4,480410099029541
  PEAK LEVEL (R128) = -0,1 dBTP; -0,8 Left; -0,1 Right
  REPLAYGAIN_TRACK_PEAK = 0.977000
  TRACKNUMBER = 3
  DATE = 1964
  GENRE = Blues
  DYNAMIC RANGE (R128) = 18,7092399597167969
  ALBUM = Folk Singer
  TITLE = My Captain
  RATING = 4
  PEAK LEVEL (SAMPLE) = -0,2 dB; -0,8 Left; -0,2 Right
  DYNAMIC RANGE (DR) = 14
  VOLUME LEVEL (REPLAYGAIN) = 0,5195900201797485
  DISCNUMBER = 1
  COMMENT = EAC
  ARTIST = Muddy Waters

Audiopath on playback:



Tags for Track from HDTracks:
Uncompressed Audio File (aif)
192,0 kHz, 24 bit, 2 ch

ID3v2.3 Tag: (524288 bytes)
  TIT2 (Name): My Captain
  TPE1 (Artist): Muddy Waters
  TPE2 (Album Artist): Muddy Waters
  TALB (Album): Folk Singer
  TRCK (Track #): 3
  TYER (Year): 1964
  TCON (Genre): Blues
  TCOM (Composer): Willie Dixon
  TBPM (BPM): 63
  APIC (Picture) (Cover): <too large to display>
  TXXX (Album Artist): Muddy Waters
  TXXX (Dynamic Range (DR)): 17
  TXXX (Dynamic Range (R128)): 19,3680591583251953
  TXXX (Peak Level (R128)): -3,4 dBTP; -4,0 Left; -3,4 Right
  TXXX (Peak Level (Sample)): -3,4 dB; -4,0 Left; -3,4 Right
  TXXX (replaygain_track_g..): +7.61 dB
  TXXX (replaygain_track_p..): 0.676000
  TXXX (Tool Name): Media Center
  TXXX (Tool Version): 19.0.66
  TXXX (Volume Level (R128)): 2,6068499088287354
  TXXX (Volume Level (Repl..): 7,6068501472473145

Audiopath on playback:




should i send you the two tracks for testing to logs at jriver dot com?



Thanks Erwin
Logged

Vocalpoint

  • Citizen of the Universe
  • *****
  • Posts: 2007
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #159 on: November 01, 2013, 08:57:39 am »

Adaptive Volume (Peak Level Normalize) and Volume Leveling is enabled.

I will let Matt tackle this one - but why do you have Peak Level Normalize on? That is pushing the level up as high as it can go without clipping (I believe)...

"If you use Volume Leveling AND Peak Level Normalization, you will get the loudest playback of the current playlist that maintains equal volume between tracks and prevents all clipping.  The volume between tracks in the playlist will be the same, but the volume between different playlists could be different (since each playlist will have a different peak level normalization value)."

I only have Volume Levelling engaged in my current testing of v19 and that Muddy Waters records sounds fantastic.

VP

Logged

faster

  • Junior Woodchuck
  • **
  • Posts: 54
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #160 on: November 01, 2013, 09:26:00 am »

but why do you have Peak Level Normalize on? That is pushing the level up as high as it can go without clipping (I believe)...

Wy not? The reason is what you have said: 2That is pushing the level up as high as it can go without clipping"

There should be no SQ problem with this option, right?

Now I tested playback without Peak Level Normalize. Same problem: CD Track sounds much louder, and Audiopath shows same value (- 4,5 dB) for volume leveling. shouldend it be different for playing the different files with the same loudness?
Logged

Vocalpoint

  • Citizen of the Universe
  • *****
  • Posts: 2007
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #161 on: November 01, 2013, 10:11:30 am »

Why not? The reason is what you have said: That is pushing the level up as high as it can go without clipping" There should be no SQ problem with this option, right?

Pushing the level up as high as it can go without clipping -  is probably not what Muddy (or any other artist ever intended). SQ quality will be effected since playback can and will be unnatural with some material.

Adaptive Volume was really designed for playlists where you could have tracks that have wide volume swings. It then glues together a playlist volume wise BUT at the expense of cranking ALL tracks to the highest peak value (max). Kinda like listening to a poorly mastered 2013 pop album on your iPod where the volume is goosed to the max and every track is at the same relentless volume. Of course Adaptive Volume has it's place if you were in a car, a party or other noisy environment where you having everything "leveled" to the max would be beneficial. Read here for more..

http://yabb.jriver.com/interact/index.php?topic=82023.0

While everyone has their own tastes - for me personally - I would never "normalize" anything. Just let the natural dynamics of the source material do the talking - and the new Volume Leveling and R128 enhancements do an excellent job of that in v19 on their own without resorting to Peak Level Normalize..

Cheers,

VP
Logged

Matt

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 42323
  • Shoes gone again!
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #162 on: November 01, 2013, 10:19:20 am »

Pushing the level up as high as it can go without clipping -  is probably not what Muddy (or any other artist ever intended). SQ quality will be effected since playback can and will be unnatural with some material.

I think you might be thinking of dynamic range compression.  That does not apply here.

Faster's screenshots say "Peak level normalize (fixed)".  This means it's no different than turning the volume up (or down) at the beginning of the song.  I think Muddy would be happy to let you listen at any volume :)
Logged
Matt Ashland, JRiver Media Center

Matt

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 42323
  • Shoes gone again!
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #163 on: November 01, 2013, 10:22:21 am »

should i send you the two tracks for testing to logs at jriver dot com?

Yes, please.

I can't explain why the second track (aif) is showing -4.5dB for volume leveling.
Logged
Matt Ashland, JRiver Media Center

Vocalpoint

  • Citizen of the Universe
  • *****
  • Posts: 2007
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #164 on: November 01, 2013, 10:29:02 am »

I think you might be thinking of dynamic range compression.  That does not apply here.

I guess I am misunderstanding how Adaptive Volume works then. Will study some more.

Cheers!

VP
Logged

mojave

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 3732
  • Requires "iTunes or better" so I installed JRiver
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #165 on: November 01, 2013, 10:30:12 am »

Since both songs have the same album tag there won't be an adjustment since "intentional volume differences between tracks on the same album are preserved." It doesn't matter that both songs are from different labels.

Regarding Peak Level Normalize - it adds back in the lost volume due to volume leveling. It rarely increases it even further. For example, if an album is played that has -4.5 dB in the Volume Level (R128) tag, the volume will be reduced by 4.5 dB when just that album is added to the playlist and Volume Leveling is on. If you also turn on Peak Level Normalize, then 4.5 dB will be added back in for a net change of 0 dB. Since the album is the only one in the playlist, its volume obviously does not need to be reduced for Volume Leveling. Basically Adaptive Volume (Peak Level Normalize) intelligently looks at the items in the playlist and uses the least amount of volume reduction necessary to still achieve Volume Leveling.

Logged

InflatableMouse

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 3978
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #166 on: November 01, 2013, 10:48:07 am »

Since both songs have the same album tag there won't be an adjustment since "intentional volume differences between tracks on the same album are preserved." It doesn't matter that both songs are from different labels.

This.

Many of us have different versions of the same album (Dark Side of the Moon, I think I have like 12 versions) and volume leveling doesn't consider [fields] other than album, artist.

@Faster: As a test, just name one of the albums "Folk Singer (CD)" and test again.

PS. Try and find the the MFSL UDCD-593 version ;).
Logged

Matt

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 42323
  • Shoes gone again!
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #167 on: November 01, 2013, 10:51:39 am »

Thanks guys for pointing out that the two files were on the same album.  That had flown right over my head, and I'm sure that's the explanation.

If they're different albums, I'd recommend giving them different names.  We might also be able to leverage the album analyzer which considers file path, but this opens up a new set of problems with users that don't reliably use album folders.
Logged
Matt Ashland, JRiver Media Center

mojave

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 3732
  • Requires "iTunes or better" so I installed JRiver
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #168 on: November 01, 2013, 11:02:14 am »

I use the Publisher tag to differentiate albums, but I don't know if that would be standard with other users.

I really can't imagine that anyone listens to the same song from different versions of the same album in one playlist.  ? I think this was just for testing.
Logged

Matt

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 42323
  • Shoes gone again!
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #169 on: November 01, 2013, 11:17:16 am »

I guess I am misunderstanding how Adaptive Volume works then. Will study some more.

Hopefully this will help:
http://yabb.jriver.com/interact/index.php?topic=82023.0

The peak level normalize mode of Adaptive Volume does no dynamic range compression.  For audio, it's mostly adding back the volume you'd lose from Volume Leveling.

The other two modes of Adaptive Volume do dynamic range compression, but the user interface and link above try to explain why you might want that in some cases.
Logged
Matt Ashland, JRiver Media Center

Vocalpoint

  • Citizen of the Universe
  • *****
  • Posts: 2007
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #170 on: November 01, 2013, 11:34:50 am »

The peak level normalize mode of Adaptive Volume does no dynamic range compression.  For audio, it's mostly adding back the volume you'd lose from Volume Leveling.

Understood. I wasn't implying "compression" but it came out that way wit the iPod comparison. I understand "normalization" in the sense of determining a peak value and then cranking everything else up to that level.

But to me that's never a good thing (to me personally) - unless of course one has a special need for it - like a specific playlist, late night viewing etc. Even with this option available - I still use just Volume Levelling alone for all my playlists and they all sound great - especially with the R128 additions in v19.

VP
Logged

InflatableMouse

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 3978
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #171 on: November 01, 2013, 11:44:30 am »

I use the Publisher tag to differentiate albums, but I don't know if that would be standard with other users.

I really can't imagine that anyone listens to the same song from different versions of the same album in one playlist.  ? I think this was just for testing.

Yeh, that's true. I use several custom fields, for instance Source (CD, HDTracks, SACD) and Album Version (MFSL, 1991 Remaster, or Cat Number).

It would be convenient if I didn't have to edit the album tag, which is also a give-away as to which version is playing when I'm switching back and forth. But as you say, its often only for testing so not that important.
Logged

faster

  • Junior Woodchuck
  • **
  • Posts: 54
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #172 on: November 01, 2013, 12:08:47 pm »

Since both songs have the same album tag there won't be an adjustment since "intentional volume differences between tracks on the same album are preserved." It doesn't matter that both songs are from different labels.

That's the case!
After i have renamed albumtag in one of the files everything works as expected!
Same loudness and adjust volume by -4,5 for CD Version and + 2,4 dB for the HD Tracks Version in audiopath.

Thanks!! Great and clever Community here, an of course a very great product!

Now i better understand how volume leveling and adaptive volume works!
And this (my) is a great example especially with different Masters (compressed vs better dynamic range recordings) from the same album.
Track "My Captain" on the CD Version is 14 (which is already well) vs HD Track is 17 (witch is exellent)!

btw. highest DR on this musthave album is track "Long Distance Call" with DR18 on the HD Tracks version (CD Version DR14).

Erwin
Logged

mojave

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 3732
  • Requires "iTunes or better" so I installed JRiver
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #173 on: November 01, 2013, 12:52:18 pm »

I have the HD Tracks Elton John - Madman Across the Water. The Madman Across the Water track is DR8.

The MCA Records version I have is DR10.

Earlier this week I purchased the DJM Records 825 487-2 made in West Germany that is supposed to have it at DR13. I haven't received the disc yet.

It will be interesting to compare.
Logged

InflatableMouse

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 3978
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #174 on: November 01, 2013, 03:29:48 pm »

Track "My Captain" on the CD Version is 14 (which is already well) vs HD Track is 17 (witch is exellent)!

btw. highest DR on this musthave album is track "Long Distance Call" with DR18 on the HD Tracks version (CD Version DR14).

Erwin

Glad to hear you figured it out.

Those DR numbers are the same as the MFSL version I mentioned. If the HD Tracks is a true high res version (ie, not upsampled) I might have to get it too.

I have the HD Tracks Elton John - Madman Across the Water. The Madman Across the Water track is DR8.

The MCA Records version I have is DR10.

Earlier this week I purchased the DJM Records 825 487-2 made in West Germany that is supposed to have it at DR13. I haven't received the disc yet.

It will be interesting to compare.

One of my favority EJ albums.

That number is better than the MFSL version I have. Not that means its better but its enough reason to look into it. I may have to get that too :D.

Logged

Vocalpoint

  • Citizen of the Universe
  • *****
  • Posts: 2007
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #175 on: November 01, 2013, 04:10:22 pm »

I have the HD Tracks Elton John - Madman Across the Water. The Madman Across the Water track is DR8. The MCA Records version I have is DR10. Earlier this week I purchased the DJM Records 825 487-2 made in West Germany that is supposed to have it at DR13. I haven't received the disc yet. It will be interesting to compare.

I just analyzed my MFSL Madman (1989) CD...and I get DR 13 in 19.0.67

Anyhoo - that WG DJM is awesome. All of the mid 80's DJM pressings are a wonderful thing.

VP
Logged

mrpro

  • Recent member
  • *
  • Posts: 19
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #176 on: November 04, 2013, 10:50:15 am »

I think I would like to use this feature. (Improved audio analysis and volume leveling (R128))
However, I have two concerns:
1. That R128 volume leveling NOT alter dynamic range or alter sound quality in any other way, and
2. Do the calculated levels alter the actual file in any way, other than to create a tag value?

It sounds like the R128 method (volume leveling), without Peak Normalize is what I want to or do I have it backwards?
Sorry if this has already been answered, but the posts here seem somewhat contradictory.
Logged

Matt

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 42323
  • Shoes gone again!
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #177 on: November 04, 2013, 10:52:16 am »

1. That R128 volume leveling NOT alter dynamic range or alter sound quality in any other way, and
2. Do the calculated levels alter the actual file in any way, other than to create a tag value?

Neither Volume leveling or Peak Level Normalize alter dynamic range.

No changes are made to the audio data in your files.
Logged
Matt Ashland, JRiver Media Center

Vocalpoint

  • Citizen of the Universe
  • *****
  • Posts: 2007
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #178 on: November 04, 2013, 11:30:02 am »

It sounds like the R128 method (volume leveling), without Peak Normalize is what I want to or do I have it backwards?Sorry if this has already been answered, but the posts here seem somewhat contradictory.

This is exactly what you want.

And as Matt mentioned - "No audio is harmed during the making of this R128 volume analysis" :)

VP
Logged

Matt

  • Administrator
  • Citizen of the Universe
  • *****
  • Posts: 42323
  • Shoes gone again!
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #179 on: November 08, 2013, 08:19:41 am »

The Dynamic Range (R128) is the R128-derived Loudness Range (LRA).

Semantics discussion split here.
Logged
Matt Ashland, JRiver Media Center

bobkatz

  • World Citizen
  • ***
  • Posts: 213
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #180 on: November 08, 2013, 02:15:40 pm »

I have the HD Tracks Elton John - Madman Across the Water. The Madman Across the Water track is DR8.

The MCA Records version I have is DR10.

Earlier this week I purchased the DJM Records 825 487-2 made in West Germany that is supposed to have it at DR13. I haven't received the disc yet.

It will be interesting to compare.

Thanks, Matt, for incorporating ITU standard loudness normalization in JRiver. Another obstacle to ending the loudness race has fallen!

Let's stop using these non-standard terms that mean nothing. The so-called terms "DR" which have been coined by a non-standards organization mean nothing. They use it to describe the level, not the dynamic range. The proper internationally-standardized loudness measurement is called "LUFS", for loudness units full scale. As Matt described earlier in the thread, the proper term for "dynamic range" is NOT DR, but rather LRA, for loudness range. This is also a standardized measurement by the ITU.

I do not believe that LRA is a good measure of sound quality, however. There will be something better down the pike, but until then, don't try to rely on LRA measurements as a judgment of whether you should buy a recording or not. For example, a good Steely Dan might have a rather small LRA, but be a very open-sounding, uncompressed recording with clean transients. But Tool's Aenima is purposely highly compressed---it has a very poor transient response and the sound is squashed to death. But loudness range?  it has plenty of it. How did they get that?  By manipulating the gain in mastering AFTER the compressor was applied. So I can make a recording which has a ton of measurable range but still sounds squashed and compressed. That's why LRA is a poor measurement of sound quality.

Stick with LUFS and that's about as close as you can get to an idea of whether a pop recording is overcompressed. But still far from perfect. No time to get into that today!
Logged

6233638

  • Regular Member
  • Citizen of the Universe
  • *****
  • Posts: 5353
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #181 on: November 08, 2013, 03:11:07 pm »

Thanks, Matt, for incorporating ITU standard loudness normalization in JRiver. Another obstacle to ending the loudness race has fallen!

Let's stop using these non-standard terms that mean nothing. The so-called terms "DR" which have been coined by a non-standards organization mean nothing. They use it to describe the level, not the dynamic range. The proper internationally-standardized loudness measurement is called "LUFS", for loudness units full scale. As Matt described earlier in the thread, the proper term for "dynamic range" is NOT DR, but rather LRA, for loudness range. This is also a standardized measurement by the ITU.
I made a big push for this on the beta testing forum. Matt felt that it would be easier for people to understand "Dynamic Range (R128)" than "Loudness Range (LRA)" or something similar to it.

We had the same argument about using LU rather than LUFS. I was also of the opinion that Media Center should match the ITU as much as possible.

I do not believe that LRA is a good measure of sound quality, however. There will be something better down the pike, but until then, don't try to rely on LRA measurements as a judgment of whether you should buy a recording or not. For example, a good Steely Dan might have a rather small LRA, but be a very open-sounding, uncompressed recording with clean transients. But Tool's Aenima is purposely highly compressed---it has a very poor transient response and the sound is squashed to death. But loudness range?  it has plenty of it. How did they get that?  By manipulating the gain in mastering AFTER the compressor was applied. So I can make a recording which has a ton of measurable range but still sounds squashed and compressed. That's why LRA is a poor measurement of sound quality.
I agree - the crest factor is generally a much better indicator of track quality than the loudness range - that's why we have both. Dynamic Range (DR) is a measure of crest factor, and the results should match the TT-DR meter.
Logged

mojave

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 3732
  • Requires "iTunes or better" so I installed JRiver
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #182 on: November 08, 2013, 03:21:59 pm »

Let's stop using these non-standard terms that mean nothing. The so-called terms "DR" which have been coined by a non-standards organization mean nothing. They use it to describe the level, not the dynamic range.
From Sound On Sound's Sept. 2011 Article on Dynamic Range:  "For instance, one criterion commonly used to describe the dynamic behaviour of a piece of recorded music is the 'crest' factor. Put simply, the crest factor is the difference between the RMS level and the peak level over the course of the song. Intuitively, it measures the amplitude of the emerging 'peaks' in the audio stream. It's considered a good marker of the amount of dynamic compression that was applied to the music: more compression generally means a lower crest factor. Some professionals consider good handling of the crest factor as the cornerstone of successful mastering. Also, still generally speaking, the lower the crest factor, the louder the music."

JRiver's Dynamic Range (DR) is a measure of the crest factor and, in my library, seems to represent compression or lack of compression quite well.

Here are the specifications for Dynamic Range (DR) implemented by JRiver. 

Quote
That's why LRA is a poor measurement of sound quality.
I haven't read any posts here of anyone that thinks LRA is a measure of sound quality.
Logged

Vocalpoint

  • Citizen of the Universe
  • *****
  • Posts: 2007
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #183 on: November 08, 2013, 03:37:39 pm »

JRiver's Dynamic Range (DR) is a measure of the crest factor and, in my library, seems to represent compression or lack of compression quite well.

Despite the plentiful amount of measurement and analysis columns available in v19 - the TT-DR meter column (DR) is the only number I care about when considering a purchase or when considering inclusion into our library.

If that specific column value is painfully high (7,6,5,4) then regardless of what kind of sound quality the source has - it's totally irrelevant to me. Any source with a DR higher than maybe 9 (using TT-DR) gives me a headache after 2 tracks so it's immediately dismissed.

These DR numbers driven by TT have never failed me yet. I have yet to encountered ANY source file with a deafening DR (8,7,6,5) that either sounds good or is worth putting in the library and I have never heard a source file with a nice DR (14,13,12,11) that sounds bad or causes me to stop playback and reach for an Advil.

The R128 stuff is a technical dream to have in this class of program - but unless you know what all those columns mean in a professional sense - I think the TT-DR numbers are a nice easy way for Average Joe to instantly know if he's in the right ballpark and the source is not crushed to death.

VP

Logged

bobkatz

  • World Citizen
  • ***
  • Posts: 213
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #184 on: November 08, 2013, 03:38:20 pm »

Mojave. Using "DR" for crest factor is a stupid, misleading, misuse of the English language. For over 100 years, the term "dynamic range" has referred to the distance between loudest and softest passages, NOT the distance between average level (or loudness) and peak level. I thoroughly detest this reinvention and reuse of one term to substitute for another. I thoroughly endorse the use of a new term such as (invented by Thomas Lund) "PLR" for the quantity you wish to describe. PLR standing for "Peak to Loudness Ratio". But please cease and desist using the term "DR" to represent something which it is not. Thank you.

By the way, I'm glad that no one except some of those in the European Radio production community has mistakenly latched on LRA as a measure of sound quality. As you say, (but using a misleading term), PLR is a much better measure of sound quality. Let's call a spade a spade and tell the people at Sound On Sound and a few other places that they are opening a can of worms by trying to reapply "dynamic range" to another quantity.

BK

From Sound On Sound's Sept. 2011 Article on Dynamic Range:  "For instance, one criterion commonly used to describe the dynamic behaviour of a piece of recorded music is the 'crest' factor. Put simply, the crest factor is the difference between the RMS level and the peak level over the course of the song. Intuitively, it measures the amplitude of the emerging 'peaks' in the audio stream. It's considered a good marker of the amount of dynamic compression that was applied to the music: more compression generally means a lower crest factor. Some professionals consider good handling of the crest factor as the cornerstone of successful mastering. Also, still generally speaking, the lower the crest factor, the louder the music."

JRiver's Dynamic Range (DR) is a measure of the crest factor and, in my library, seems to represent compression or lack of compression quite well.

Here are the specifications for Dynamic Range (DR) implemented by JRiver. 
I haven't read any posts here of anyone that thinks LRA is a measure of sound quality.
Logged

bobkatz

  • World Citizen
  • ***
  • Posts: 213
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #185 on: November 08, 2013, 03:39:32 pm »

Tell the people at TT to rename their misleading and incorrect terminology. Peak to loudness ratio is NOT dynamic range. And the TT meter is NOT a dynamic range meter.

BK


Despite the plentiful amount of measurement and analysis columns available in v19 - the TT-DR meter column (DR) is the only number I care about when considering a purchase or when considering inclusion into our library.

If that specific column value is painfully high (7,6,5,4) then regardless of what kind of sound quality the source has - it's totally irrelevant to me. Any source with a DR higher than maybe 9 (using TT-DR) gives me a headache after 2 tracks so it's immediately dismissed.

These DR numbers driven by TT have never failed me yet. I have yet to encountered ANY source file with a deafening DR (8,7,6,5) that either sounds good or is worth putting in the library and I have never heard a source file with a nice DR (14,13,12,11) that sounds bad or causes me to stop playback and reach for an Advil.

The R128 stuff is a technical dream to have in this class of program - but unless you know what all those columns mean in a professional sense - I think the TT-DR numbers are a nice easy way for Average Joe to instantly know if he's in the right ballpark and the source is not crushed to death.

VP


Logged

bobkatz

  • World Citizen
  • ***
  • Posts: 213
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #186 on: November 08, 2013, 03:50:46 pm »

My recommendation is that Matt truly stick to the ITU specs. ITU is an international standards organization. "DR" is not an internationally-recognized term. I recommend the term "PLR" or "Crest Factor" to apply to the distance between the average loudness and the highest measured peak of the material. It's not in the ITU book, but at least it won't confuse readers and users as DR clearly does.

The so-called "DR" meter is NOT a measure of dynamic range.  Not in the least.

BK
Logged

Vocalpoint

  • Citizen of the Universe
  • *****
  • Posts: 2007
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #187 on: November 08, 2013, 03:53:33 pm »

Tell the people at TT to rename their misleading and incorrect terminology. Peak to loudness ratio is NOT dynamic range. And the TT meter is NOT a dynamic range meter.

Bob,

100% agree with you.

However - since you are the "man" in some circles regarding this wide topic (I still read your "Mastering Audio" book with purpose :) - it would probably have more weight if you told the TT people to correct it. :)

Just sayin...

VP
Logged

6233638

  • Regular Member
  • Citizen of the Universe
  • *****
  • Posts: 5353
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #188 on: November 08, 2013, 04:03:22 pm »

My recommendation is that Matt truly stick to the ITU specs. ITU is an international standards organization. "DR" is not an internationally-recognized term. I recommend the term "PLR" or "Crest Factor" to apply to the distance between the average loudness and the highest measured peak of the material. It's not in the ITU book, but at least it won't confuse readers and users as DR clearly does.

The so-called "DR" meter is NOT a measure of dynamic range.  Not in the least.
I'd definitely like to see this change; I was never very keen on the "Dynamic Range (DR)" naming.
Is there a standard for this Peak to Loudness Ratio? I'm wondering if the Dynamic Range (DR) field should be renamed, or if it would be a new addition.
Logged

bobkatz

  • World Citizen
  • ***
  • Posts: 213
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #189 on: November 09, 2013, 10:52:57 am »

I'd definitely like to see this change; I was never very keen on the "Dynamic Range (DR)" naming.
Is there a standard for this Peak to Loudness Ratio? I'm wondering if the Dynamic Range (DR) field should be renamed, or if it would be a new addition.

There is no official standard yet for the "PLR". But it is becoming a de facto standard being pushed by Thomas Lund of TC Electronic. It is the simple difference between the ITU highest true peak and the average loudness so it is not a big stretch to call it by a name. I intend to publicize it and recommend that "PLR" be used as the designation in the next edition of my book, "Mastering Audio." Maybe it will catch on.
Logged

bobkatz

  • World Citizen
  • ***
  • Posts: 213
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #190 on: November 09, 2013, 10:54:07 am »

Bob,

100% agree with you.

However - since you are the "man" in some circles regarding this wide topic (I still read your "Mastering Audio" book with purpose :) - it would probably have more weight if you told the TT people to correct it. :)

Just sayin...

VP

I haven't gotten any budge from the TT people. Someone's being very stubborn  :-(.
Logged

Vocalpoint

  • Citizen of the Universe
  • *****
  • Posts: 2007
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #191 on: November 09, 2013, 11:18:10 am »

I haven't gotten any budge from the TT people. Someone's being very stubborn  :-(.

Probably hard for them to admit they are wrong.

On another level tho - whilst I totally agree with your technical angle to this - the current "DR" measurement (via TT meter) - right or wrong - does give the layman out there an easy way to quickly determine whether a specific recording has had the crap kicked out of it via dynamic range compression.

Seeing a nice simple number like DR 12 - is much easier to process in one's brain than "6.4 LU" or many of the other values being displayed in MC v19. Looking at the Audio Analysis dialog in v19 - I sometimes feel like I am in an engineering or statistics class :)

When I visit my audio forums and want to gather some VERY quick "intel" on a specific recording prior to purchase - the DR analysis (previously via Foobar or the TT standalone meter and now - thankfully via MC) is really the only weapon out there to give us any indication on how "good" or "bad" the actual "listenability" is on any CD release. If someone has purchased a new CD and has ran it thru the DR meter - and is kind enought to post the results - that's perfect by me.

While I am all about technical "standards"- I do not think that moving to "ITU" and "PLR" and so on - will actually mean a whole lot to those simply wanting to find out if a CD is crushed to death. I agree the naming (or what the TT should be actually measuring) could/should be changed up to be more technically correct...but the pure simplicity of a single number popping out per track is what sells it for me.

At the end of the day - I really don't think a lot of folks actually care what the TT meter measuring - they just want to take comfort in the fact that if they see a 5 on the DR analysis for a specific CD - there's a very high probability that it's a sonic disaster and to stay away from it.

VP

Logged

bobkatz

  • World Citizen
  • ***
  • Posts: 213
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #192 on: November 09, 2013, 11:28:04 am »

Vocalpoint......  Yeah, use the TT meter, it gives us something useful. But please tell everyone that you recommend the meter to that it's not measuring "dynamic range", it's measuring a sort of crest factor. And the two terms are very different and should not be confused.

Logged

Vocalpoint

  • Citizen of the Universe
  • *****
  • Posts: 2007
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #193 on: November 09, 2013, 11:39:32 am »

Yeah, use the TT meter, it gives us something useful. But please tell everyone that you recommend the meter to that it's not measuring "dynamic range", it's measuring a sort of crest factor. And the two terms are very different and should not be confused.

Unfortunately - most folks that I usually yak with regarding the current "DR" meter wouldn't know crest factor from Crest toothpaste :). Most of these folks are simply trying to get thru life trying to find decent music to listen to and not get burned by the current "remaster" myths out there. This little doodad may not be technical correct - but it's easy. That's all these folks really want.

And if no one has mentioned it yet - very nice to see you on this board BTW.

Cheers!

VP
Logged

InflatableMouse

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 3978
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #194 on: November 09, 2013, 12:41:28 pm »

I'm sorry I got to ask ... what is dynamic range if its not what Dynamic Range (DR) or Dynamic Range (R128) are measuring? I'm confused ...
Logged

6233638

  • Regular Member
  • Citizen of the Universe
  • *****
  • Posts: 5353
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #195 on: November 09, 2013, 12:56:08 pm »

I'm sorry I got to ask ... what is dynamic range if its not what Dynamic Range (DR) or Dynamic Range (R128) are measuring? I'm confused ...
Dynamic range is the difference between the loudest and quietest part of the track.
 
Dynamic Range R128 is a measure of dynamic range, but a statistically weighted one to provide more useful results, that the EBU call "Loudness Range". (LRA)
 
Dynamic Range (DR) is a measure of Crest Factor, which is not related to Dynamic Range at all - it compares the peak loudness to the average loudness. It's a good measure of how dynamic a track sounds though. (as opposed to sounding "flat" and "compressed")
Logged

InflatableMouse

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 3978
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #196 on: November 09, 2013, 03:33:49 pm »

I think I get what the 2 measurements are but aparently some people seem strongly opposed to calling DR "dynamic range".

It's probably all lost on me anyways, I'm the kind that can't tell cresh factor from toothpaste  ;D, but seriously, isn't it just a matter of interpretation?
Logged

bobkatz

  • World Citizen
  • ***
  • Posts: 213
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #197 on: November 09, 2013, 03:57:25 pm »

I think I get what the 2 measurements are but aparently some people seem strongly opposed to calling DR "dynamic range".

It's probably all lost on me anyways, I'm the kind that can't tell cresh factor from toothpaste  ;D, but seriously, isn't it just a matter of interpretation?

As long as you know what a term means to you, then you can call it "cupcakes" if you like. But for clear communication amongst professionals and serious lay people, it pays for a high quality application like JRiver to use correct, unambiguous terminology.

My two cents,


Bob
Logged

6233638

  • Regular Member
  • Citizen of the Universe
  • *****
  • Posts: 5353
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #198 on: November 09, 2013, 04:37:18 pm »

As long as you know what a term means to you, then you can call it "cupcakes" if you like. But for clear communication amongst professionals and serious lay people, it pays for a high quality application like JRiver to use correct, unambiguous terminology.
Thanks, your new rating system is really helping me understand my library.
 
Unfortunately it just tends to highlight the sorry state of music releases these days. :(
Logged

mwillems

  • MC Beta Team
  • Citizen of the Universe
  • *****
  • Posts: 5233
  • "Linux Merit Badge" Recipient
Re: NEW: Improved audio analysis and volume leveling (R128)
« Reply #199 on: November 09, 2013, 10:57:29 pm »

Thanks, your new rating system is really helping me understand my library.
 
Unfortunately it just tends to highlight the sorry state of music releases these days. :(

I appreciate the underlying dilemma about terminology, but that screencap is the funniest thing I've seen all week.  Not sure why, but it just completely creased me up  ;D
Logged
Pages: 1 2 3 [4] 5   Go Up