Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[InflatableMouse]

Funny.

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

Of course, and I appreciate that. I didn't mean to imply the opposite. My previous post didn't quite come out the way I intended. I was tired, I rushed it and wanted to go to bed. And the toothpaste joke was just a silly joke, nothing more. I'm actually trying to understand this.

I've read this before and I just read it again with the wiki page on crest factor. I'm honestly trying to understand . I went back to the RR DR site and found a link to this which was also a site referenced above.

From that EBU – TECH 3342 pdf I found:

Loudness Range should not be confused with other measures of dynamic range or crest factor, etc.

So there's that. Crest factor is not dynamic range. I get that. Then what is dynamic range (which is actually what I meant to ask all along)?

I found this:

There remains the question of whether one should use such a term as 'dynamic range' at all: there is no official definition for it, and it may be confused with the dynamic range of a recording medium, which is basically the difference between the highest and lowest level it can handle. During the course of this article, therefore, I won't talk about 'dynamic range' in relation to a piece of music. Instead, I will be using 'RMS variability', or more generally 'dynamic variability'. The term 'dynamic range' will be reserved for the measure of signal-to-noise ratio of a recording medium. I will use the term 'loudness range' in strict reference to the EBU 3342 document, and the term 'loudness variability' in other cases involving loudness instead of RMS.

Is that an accurate description of dynamic range then?

[InflatableMouse]

I seem to have killed this thread somehow ... 

[chrisjj]

for clear communication amongst professionals and serious lay people, it pays for a high quality application like JRiver to use correct, unambiguous terminology.

Hear, hear.

[chrisjj]

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

Some mistake? That's a manual for "TT DYNAMIC RANGE METER VST Plugin", not a specification.

[Vocalpoint]

Back to audio analysis now

I want to get some feedback (specifically from Matt) on a oddity I am noticing. This past weekend - I complete our whole home move to v19 and everything went great. Of course - it took may hours to analyze the entire library using the new v19 toolkit - but I have noticed an issue - that may be related to a thread I brought up several weeks back where MC was choking on MP3 conversions with 4 files at a time.

Since I had 60000 tracks to analyze over this weekend - I spent a bit of time watch MC do it's thing and noticed a peculiar "lag" in the first thread of a 4 thread analysis cycle. While the first track (Thread) seems to start with the others - it then starts to slowly grind along 5%, 9%, 11% sometimes stopping and waiting for awhile etc etc while the other 3 active rip right through from 0% to 100%....then when the next group of four are targeted - another thread seems to always lagging behind the other three.

Wondering if anyone else has noticed this and if there was an speed issue or not. I realize that all track analysis is unique but I  am wondering why that first thread is so slow in completing it's pass while the other three threads seem to fly.

This behavior is consistent on my workstation and can be better seen when doing 30 or 40 at a time...one thread will always be noticeably lagging way behind the other three.

Appreciate any comments from the field...

VP

[6233638]

Since I had 60000 tracks to analyze over this weekend - I spent a bit of time watch MC do it's thing and noticed a peculiar "lag" in the first thread of a 4 thread analysis cycle. While the first track (Thread) seems to start with the others - it then starts to slowly grind along 5%, 9%, 11% sometimes stopping and waiting for awhile etc etc while the other 3 active rip right through from 0% to 100%....then when the next group of four are targeted - another thread seems to always lagging behind the other three.

This happens all the time - I suspect it's when MC is limited by disk/network speeds.
I have often requested that MC tries to assign a drive to each CPU core rather than working down the list of files sequentially to try and move the limiting factor back to your CPU rather than disk/network access.
I'd also like to see MC be "smarter" about how many files it's analyzing when limited by disk I/O.

[Vocalpoint]

This happens all the time - I suspect it's when MC is limited by disk/network speeds.

But I still see this behavior on local files...certainly not limited by disk I/O...that first thread always stands around like it has no idea what it should be doing Over the network - yes - could be anything - but that said - the other three threads - network or local - still burn through with lighting speed...

Something is going on with that first thread...maybe it's normal or by design?

VP

[Hendrik]

I've not seen this, i typically use 3 threads and they all run just fine at about the same speed. More threads caused disc IO issues when analyzing video files, so i settled on 3 (it slowed down all analysis tasks, not just one, though)

[Vocalpoint]

I've not seen this, i typically use 3 threads and they all run just fine at about the same speed. More threads caused disc IO issues when analyzing video files, so i settled on 3 (it slowed down all analysis tasks, not just one, though)

I will crank it back to 3 today and see what happens.

VP

[jamesecox50]

Funny.

Of course, and I appreciate that. I didn't mean to imply the opposite. My previous post didn't quite come out the way I intended. I was tired, I rushed it and wanted to go to bed. And the toothpaste joke was just a silly joke, nothing more. I'm actually trying to understand this.

I've read this before and I just read it again with the wiki page on crest factor. I'm honestly trying to understand . I went back to the RR DR site and found a link to this which was also a site referenced above.

From that EBU – TECH 3342 pdf I found:

So there's that. Crest factor is not dynamic range. I get that. Then what is dynamic range (which is actually what I meant to ask all along)?

I found this:

Is that an accurate description of dynamic range then?

There may be no official standard but we know good DR when we hear it! The Loudness war movement is real and correlates well with my taste, R128 seems worthless. Thanks for incorporating DR, I was just starting to think the upgrade was farce, you squeaked her out though.

[6233638]

There may be no official standard but we know good DR when we hear it! The Loudness war movement is real and correlates well with my taste, R128 seems worthless. Thanks for incorporating DR, I was just starting to think the upgrade was farce, you squeaked her out though.

I'm not sure I understand why you think R128 Leveling seems worthless - it directly counteracts the one of the main problems caused by the loudness war. (tracks being played too loud)

[vladacasa]

Hi, having in my playlist a mix of red-book, 24/88, 24/96, 24/192 and DSD (128 bit stream) files, I've noticed that in the new 87 version, the DSD files have a lower volume than before. And that for DSD the volume leveling is not active. I keep my rule to introduce no changes in the signal path, but I was hit by a large swing on a red-book file after several DSD ones.
I'm missing something?
Regards,
Vlad

[Hendrik]

If you use DSD Bitstreaming, Volume Leveling would be disabled.

If you have volume leveling active for PCM tracks, and disabled for DSD tracks, you might experience changes in the level.

[tcman41]

I want to copy 30 tracks from my collection to a temp directory and then audio analyse and then volume level all the tracks to each other.

I am playing the tracks via winamp with a broadcast module thru a third party server site to a chat forum that i DJ at (if you will). Therefore I need MC 19 to physically change the volume level data inside each of the 30 tracks, once again so that all tracks are volume leveled to each other when they are played via winamp.

Can MC 19 do what I am requesting?, any help would be appreciated. ( I am using MC 18)

Terry 

[GreggP]

I have a question about using the data from audio analysis to create playlists.

I have created several "mellow" smartlists based on the following parameters from the old audio analysis:

BPM <= 75 and Intensity <=2

Now that 'Intensity' is gone, all of these smartlists are empty. How do I use the new tools to create similar "mellow" playlists?

I understand that the dynamic range measures the difference between the quietest and loudest levels in a track and that if the dynamic range is small, the track is probably more intense. Tracks that have some quiet passages and some loud passages will have a large dynamic range. This doesn't really help me find "mellow" music. For example, a track of acoustic guitar played quietly will probably have a low dynamic range if there aren't any silent parts.

I think the volume or peak levels would also relate to the mellowness of a song. The volume or peak level of a mellow song should not be as large as the volume or peak level of an intense song. However, the values shown in MC19 don't look like they are absolute values. They look like corrective values. So I am not sure how I could use them in this case.

Can someone help me come up with a formula to create a "mellow" playlist?

[tcman41]

ok, first time using replay gain and auto leveling.

I took a new country album and an old country album with vastly different sound volume levels. I audio analzed the two albums and then during playback used volume leveling with automatic function. The two albums now play at the same level, great it works.

From what I understand the albums now will play at the same volume in another player like winamp, I played them in winamp and sure enough the volume levels are the same, how is this done, winamp doesnt have a volume leveling checkbox to check.

thanks
Terry

[6233638]

From what I understand the albums now will play at the same volume in another player like winamp, I played them in winamp and sure enough the volume levels are the same, how is this done, winamp doesnt have a volume leveling checkbox to check.

If the player has support for ReplayGain, it will read the tags Media Center has written to the file and use them. (though this "ReplayGain" data is actually derived from the R128 analysis)

[tcman41]

If the player has support for ReplayGain, it will read the tags Media Center has written to the file and use them. (though this "ReplayGain" data is actually derived from the R128 analysis)

If I audio analyze a bunch of mp3's and then format convert them to flac, is the replay gain / volume leveling info still retained?

[Trumpetguy]

If I audio analyze a bunch of mp3's and then format convert them to flac, is the replay gain / volume leveling info still retained?

Out of curiosity, why would you do that?

[mykillk]

I'm starting to feel like I need a PhD in audio production to understand this 

[lendall]

Since DR is coming next build (see above), you might wait to analyze.

I'm lost.  Did this ever happen?

[MrC]

19.0.26 (8/13/2013)

3. NEW: Added DR dynamic range analysis to the audio analyzer.

See also:

   http://wiki.jriver.com/index.php/File_Properties_%28tags%29#Dynamic_Range_R128
   http://yabb.jriver.com/interact/index.php?topic=82025.0
   http://yabb.jriver.com/interact/index.php?topic=81980.0

[theoctavist]

will analyzing audio for 8000 tracks (all flac with cover art 16 bit or higher) take an extremely long time? because every time i open jrmc it is still on track one.


 

ideas?

[MrC]

My analysis over under 20k tracks took took most of the night; I don't recall how long exactly, but it was done by morning.  The time will also depend on your machine's speed.

The analysis should not keep restarting with track 1 each time.  The calculated values are stored in the library.  Are your files imported into the main library, or is this some DLAN library from a DLNA server?

[theoctavist]

My analysis over under 20k tracks took took most of the night; I don't recall how long exactly, but it was done by morning.  The time will also depend on your machine's speed.

The analysis should keep restarting with track 1 each time.  The calculated values are stored in the library.  Are your files imported into the main library, or is this some DLAN library from a DLNA server?

so after it analyzes them once it has the values it needs, yes? im trying to figure out why, when I open JRiver every single time, it seemingly starts with the same "cache" of files(same count, etc) and takes forever and a month of sundays to do them...... (ie seemingly several days) would a log help, i mean could yall tell me if everything is hunky dory

[MrC]

I corrected my typo above.  Keyword NOT was missing.

Are you able to tag files and have the values remain?

Please describe how you import files, any import settings you've configured, and where the files are located.

[theoctavist]

yes, I can tag files, they are fine. 


I get cds so frequently(or vinyl rips) that I am constantly using the "run auto import" feature(so the new stuff will show up.


the bulk of the audio /video files IS on an external 2tb disc. program files are on the C drive.

two folders on the external drive are in the library, and two folders in the system(the system pictures folder, and also  a temporary location I name "limbo" where I keep downloaded stuff from HD tracks that ive bought(if I dont like it it doesnt go to the main library on the external drive)

[jkrzok]

When this feature first rolled out I created a smartlist that showed files that were as yet not analyzed. The file count most recently was negligible and limited to files recently imported. With the release of .116 my count is now up to some 46,000 files (of approx. 377,000) needing analysis. The files that need reanalysis are without exception audio files. Any ideas?


Smartlist includes Volume Level (R128) is empty and Dynamic Range (R128) is empty.

[Crom]

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.

Yes please. It would be great to be able to configure what is defined as an album. I've found that if I give the album a different name it doesn't get cover art correctly and if I give it the correct name then MC has problems with differentiating during audio analysis.

[theoctavist]

i still have no idea how to ID which files in my library are giving jriver audio analysis problems.

if there is a way to filter them, or some app for that purpose, that would be fantastic. would prefer to do it with jriver tho

[6233638]

i still have no idea how to ID which files in my library are giving jriver audio analysis problems.

if there is a way to filter them, or some app for that purpose, that would be fantastic. would prefer to do it with jriver tho

If you manually run analysis, it will either error out, or stop progressing on bad files.

[retro]

I'm trying this feature for the first time, and so far I have analyzed about 20000 of my ca 90000 files.
I have added the relevant columns like "volume level (replay gain)" etc. to see whats happening during the process. Its very interesting.

But, so far it has only analyzed my normal flac's and ape's with new values, none of my 24/48, 24/96, 24/192 or my sacd iso's have been analyzed or show any values in the new columns..?!? Why is that..?!?

Am I missing something..?!?


Edit.
A few of my 24/48 and 24/96 does get analyzed. But the majority wont...?!?

Edit no 2  .
Manual analyze works! That makes me happy..!! I see now that analyzing sacd iso's takes a long time, at least on my Core2 server. So maybe this issue was a non-issue. Sorry..

[Rafal Lukawiecki]

This expression will let you see how much headroom your tracks require. (it could probably be cleaned up - but it works)

Code: [Select]

Delimit(if(isempty([Peak Level (R128)]),,formatnumber(math(removecharacters(left([Peak Level (R128)],5),/ /+,0)+RemoveCharacters([Volume Level (R128)],/ LU,0)+1),1)),/ dB,)

Sorry for resuscitating an older thread, but I have a question regarding this expression and Volume Levelling in general. I have analysed audio in my library, and if I understand this expression it adds Peak Level R128 to Volume Level R128 plus 1. I have a track on Grofe Grand Canyon Suite album, track "Crickets and Distant Thunder". It shows:

Volume Level R128: 12.9 LU
Dynamic Range R128: 26 LU
Peak Level R128: -5.1 dBTP, -5.1 Left, -14.6 Right
Dynamic Range: 12
Headroom (from above expression): 8.8 dB

If I understand it correctly, this means it should clip heavily if it is volume levelled at the analysed setting of 12.9. My question is why Mac MC 20.0.40 (oops, should I start a thread in the new forum, Mac or Win?) calculate this value at a level that would lead to such high amount of clipping? Also, why is the Dynamic Range number consistently different from the Dynamic Range R128? 

[mwillems]

Sorry for resuscitating an older thread, but I have a question regarding this expression and Volume Levelling in general. I have analysed audio in my library, and if I understand this expression it adds Peak Level R128 to Volume Level R128 plus 1. I have a track on Grofe Grand Canyon Suite album, track "Crickets and Distant Thunder". It shows:

Volume Level R128: 12.9 LU
Dynamic Range R128: 26 LU
Peak Level R128: -5.1 dBTP, -5.1 Left, -14.6 Right
Dynamic Range: 12
Headroom (from above expression): 8.8 dB

If I understand it correctly, this means it should clip heavily if it is volume levelled at the analysed setting of 12.9. My question is why Mac MC 20.0.40 (oops, should I start a thread in the new forum, Mac or Win?) calculate this value at a level that would lead to such high amount of clipping? Also, why is the Dynamic Range number consistently different from the Dynamic Range R128?  

You have it backwards; the 12.9 number is the number of dB by which the track is attenuated for volume leveling, not boosted.  Volume leveling only works by attenuation, it never rarely adds boost [EDITED: because I was incorrect).

The two dynamic ranges are different because they measure very different things. I explained the difference between the two in detail here: http://yabb.jriver.com/interact/index.php?topic=84489.msg577389#msg577389

[Rafal Lukawiecki]

You have it backwards; the 12.9 number is the number of dB by which the track is attenuated for volume levelling, not boosted.  Volume levelling only works by attenuation, it never adds boost.

If I understand you correctly, this might then be a bug, as in this case the Volume Level R128 is, actually, a positive figure of 12.9, not a negative number, like for majority of my tracks. I understood that negative meant attenuation but positive values meant volume boosting. I have 332 tracks for which Volume Level is a positive number. Why would this happen?

[mwillems]

If I understand you correctly, this might then be a bug, as in this case the Volume Level R128 is, actually, a positive figure of 12.9, not a negative number, like for majority of my tracks. I understood that negative meant attenuation but positive values meant volume boosting. I have 332 tracks for which Volume Level is a positive number. Why would this happen?

I've never seen a track that required a positive adjustment of 12 dB, so I misunderstood what you posted; my apologies.  

Volume levelling is trying to achieve a specific level (83 dB at -20dBFS) and some tracks will have positive values because the tracks in question are very, very quiet tracks. In order to make them "the same volume" as other tracks they would need to have volume added to them.  I misspoke before when I said volume leveling never boosts; in practice volume leveling very rarely adds boost because the vast majority of tracks have a peak level at or near 0dBFS, so no volume can be safely added.  Your track is unusual in that it has a very low peak level as well, so about 4 dB could be safely added, but that's it.  

So the track would need 12 dB of boost to sound "the same" volume as other tracks, but because it peaks at -5.1dB, only 4 dB can be safely added, and that's the most volume leveling would ever add.  Sorry for the miscommunication/information!

[Rafal Lukawiecki]

Your track is unusual in that it has a very low peak level as well, so about 4 dB could be safely added, but that's it.  

So the track would need 12 dB of boost to sound "the same" volume as other tracks, but because it peaks at -5.1dB, only 4 dB can be safely added, and that's the most volume leveling would ever add.  Sorry for the miscommunication/information!

Many thanks for your explanations, they are very helpful. Let me check if I still understand it the way you have explained, though: in this track, even though the Volume Level R128 has been calculated as +12.9 LU, but because the Peak Level R128 is -5.1 dBTP, Volume Levelling would not add the 12.9 boost, but only the smaller amount, that is of 5.1 dB, in order to prevent clipping. Is that correct or have I still misunderstood? Many thanks for your patience...

[mwillems]

Many thanks for your explanations, they are very helpful. Let me check if I still understand it the way you have explained, though: in this track, even though the Volume Level R128 has been calculated as +12.9 LU, but because the Peak Level R128 is -5.1 dBTP, Volume Levelling would not add the 12.9 boost, but only the smaller amount, that is of 5.1 dB, in order to prevent clipping. Is that correct or have I still misunderstood? Many thanks for your patience...

You're very close, and you have the principle correct, but with one tweak: volume leveling typically leaves a 1dB "margin of error" to make absolutely sure no clipping occurs.  So the most I would expect volume leveling would add is 5.1 - 1 = 4.1dB

[Bigguy49]

While aware of its impact audibly, definitely a newbie to the world of measured DR! 

Recently found out about the <dr.loudness-war.info/> database and, after googling a bit on using Audacity to generate DR values, learned that MC will generate them.  Very nice, BTW!  Not sure how comparable the results might be but I am sure they are close enough for my purposes since I am not doing any automatic Volume Leveling.  Information in the Wiki on Analyze Audio was a bit lean.  It would be helpful for us that are less technically savvy to have a (simple?) explanation of what the various results are and/or mean to the average listener.  Skimming this thread made my head hurt!  :-)

I see where I can select track(s) of an album, etc., and then activate the Analyze Audio option.  Is there some way to select ALL tracks in my library and generate the data "enmasse"?  IF so and I chose to analyze the full database, then I guess I would need to keep track of what was added subsequently and batch analyze?!

Can the Analyze Audio be used on PCM formats other than 16/44?   How about DSD?

Thanks in advance.

[ferday]

While aware of its impact audibly, definitely a newbie to the world of measured DR! 


I see where I can select track(s) of an album, etc., and then activate the Analyze Audio option.  Is there some way to select ALL tracks in my library and generate the data "enmasse"?  IF so and I chose to analyze the full database, then I guess I would need to keep track of what was added subsequently and batch analyze?!

Can the Analyze Audio be used on PCM formats other than 16/44?   How about DSD?

Thanks in advance.

go to your audio view, select all the tracks (shift-click), and analyze the audio.  if you have a lot of tracks it will take some time!

in the analyze audio menu, there is a checkbox "skip analyzed files".  check this, and never think about which files have been done or not, as MC will skip previously analyzed files

you can analyze any PCM bit/sample rate, and DSD .dff files.  you can also analyze the audio for video files using the same tool

[Bigguy49]

go to your audio view, select all the tracks (shift-click), and analyze the audio.  if you have a lot of tracks it will take some time!

in the analyze audio menu, there is a checkbox "skip analyzed files".  check this, and never think about which files have been done or not, as MC will skip previously analyzed files

you can analyze any PCM bit/sample rate, and DSD .dff files.  you can also analyze the audio for video files using the same tool

Thank you, ferday.

From info you provided, figured out that I had to open up the AUDIO tree to get to the folder where the source files are stored, double click on folder to show all files, do CTRL A to select all files, and then ANALYZE AUDIO.  The "skip analyzed files" was already checked (default?)
Does it matter whether you select "1 file at a time", "2 files at time", etc.?

Tried googling on the ANALYZE AUDIO data column headings but not finding a relatively understandable (to me) description of what they mean and their implication on listening.  Again, it would be great if the Wiki were expanded to include this information.

I am using MC at probably close to its simplest implementation but continue to be impressed with its capabilities! 

[Bigguy49]

Just found the TAGS section of the Wiki which has a glossary of all terms including those in the column headers of the ANALYZE AUDIO data screen.

Given site thoroughness with most things, figured that there WAS a description of terms somewhere but did not associate the ANALYZE AUDIO with the TAGS function!

Thanks.

[zoom+slomo]

MC19 makes several important improvements to audio analysis and volume leveling:

• Adoption of the R128 industry standard to analyze the loudness and dynamic range of content
  
• Ability to analyze audio for video files, including surround sound
• Smarter Volume Leveling that automatically respects intentional between track levels when playing from an album
• Volume Leveling works together with Adaptive Volume's peak level normalization
  
• Peak level is reported in decibels, measured as an R128 compliant True Peak, and reported per channel

This is a big topic, so I'll try to revisit it and post more details about the changes and the motivation at a later date.

Regarding this not uncommon problem https://benchmarkmedia.com/blogs/application_notes/intersample-overs-in-cd-recordings?_pos=1&_sid=0eeb1f150&_ss=r   

https://gearspace.com/board/mastering-forum/1401406-intersample-clipping-audible.html

I had recently asked about it here. https://yabb.jriver.com/interact/index.php?topic=135725.0 , But it's still not clear to me, which if any player improvements allow correction of this specific problem. If yes, how can JRiver be set up to reduce or eliminate this kind of distortion?