PLing and dynamics

Comments about LibriVox? Suggestions to improve things? News?
Post Reply
habasud
Posts: 283
Joined: June 22nd, 2020, 11:30 am
Location: S-H, Germany
Contact:

Post by habasud » September 14th, 2021, 3:11 am

I recently had some discussions about the dynamics of my readings. Maybe some of you would like to discuss this topic and share your opinions about that.

The recording I made started very dramatic -a ghost calls the souls of those who lost their lives at sea- and then turns down to a more calm scene in the second half -a wife giving toys to children.
My recording was technically o.k. and clean - no distortion, no pops, no loudness steps caused by editing, normalized with ReplayGain to the correct level. Nevertheless I got a PL-Note: I should reduce the level of the loud parts and should think about aligning the loudness of the beginning and the end of the recording.

My opinions on that are:

It is the readers part to interpret a text as he thinks, it should be. It is not the DPLs part to argue on that as long as the technical specs are o.k. On the contrary I would consider this as PLing by taste, which is not desirable - even when done with the best intentions.

Librivox's claim is "Acoustical liberation of books ..."
Since about the 80's of the last century the audio world saw a phenomenon known as "loudness war" (...you may search the web for that term). Productions were more an more compressed and normalized to the max in order to gain attention of the listeners. Think of music and voices on FM stations.
In order to end this people invented tools like ReplayGain (...others to mention: EBU R128, ITU BS1770). On one hand these algorithms align the loudness of different recordings. But the main benefit was: They are bringing back headroom to the audio world - that means: dynamics. Sound engineers and creative people could use this means to bring back life to their recordings.

Thinking of this I would strongly disagree when DPLs force the readers to avoid dynamics. If the loud/calm voice suits the story and the recording is o.k. in tech. terms, in my opinion there is nothing to complain about that. On the contrary: On the long run many good recordings may be spoiled forever when the readers are advised to "correct" their work.

I'm interested on what you think about that :-)

Regards
Claus
Life is What Happens To You While You're Busy Making Other Plans (John Lennon)

https://librivox.org/reader/14583

annise
LibriVox Admin Team
Posts: 34893
Joined: April 3rd, 2008, 3:55 am
Location: Melbourne,Australia

Post by annise » September 14th, 2021, 5:15 am

The LV answer to that is clear - if the text is correct and the volume is within the range you thank the PLer, but leave it the way it is if you are happy with it. No one forces you to change anything that conforms to the technical requirements.

Anne

TriciaG
LibriVox Admin Team
Posts: 51886
Joined: June 15th, 2008, 10:30 pm
Location: Toronto, ON (but Minnesotan to age 32)

Post by TriciaG » September 14th, 2021, 5:40 am

If the recording is within the volume range we need, then it's OK. However, another view:

Keep in mind that people may be listening to these on the commuter train, in a car, or in other places where it may be hard to hear the calmer parts if they're too quiet. And if they've got ear buds in, they may hurt their ears if the loud parts are really loud, taking them by surprise. :)

So yes, dynamics are a matter of interpretation and taste, but there is also the listener to think about. If the recording goes from 80 dB in the quiet parts to 95 dB in the louder parts, it's really hard for the listener to adjust.

david wales
Posts: 6227
Joined: May 2nd, 2011, 5:46 pm
Location: Los Angeles

Post by david wales » September 14th, 2021, 7:35 am

"... in other places where it may be hard to hear the calmer parts if they're too quiet."

And to supplement Tricia's comment: as a person with some hearing loss, I am constantly aware in my recording of trying to keep recording levels, well... level, so that the listener can adjust their hearing level to something that works for them.
Peace, David

DACSoft
Posts: 1524
Joined: August 17th, 2013, 8:51 am
Location: Connecticut, US

Post by DACSoft » September 14th, 2021, 12:40 pm

annise wrote:
September 14th, 2021, 5:15 am
The LV answer to that is clear - if the text is correct and the volume is within the range you thank the PLer, but leave it the way it is if you are happy with it. No one forces you to change anything that conforms to the technical requirements.

Anne
To expand on Anne's comment ...

Claus' discussion brought up an additional question for me regarding different volume levels in different sections of a recording and the technical requirements.

According to the LV Wiki (https://wiki.librivox.org/index.php?title=Questions_and_Answers - second question), it is the overall volume of the recording that needs to be in the technical specs range. So, for example, if half of one's recording is, let's says 99.0dB and the other half 79.0dB, but the overall recording registers within the range (let's say 89.0dB), it would pass technical specs (although half the recording was well above, and half well below the volume range, if each half were a separate recording).

Although, please let me know if I misunderstood this.

FWIW,
Don (DACSoft)
Bringing the Baseball Joe series to audio!

In Progress:
Baseball Joe in the Big League; Whispering Tongues; The Story My Doggie Told to Me
Next up:
Baseball Joe on the Giants; Pickett's Gap

TriciaG
LibriVox Admin Team
Posts: 51886
Joined: June 15th, 2008, 10:30 pm
Location: Toronto, ON (but Minnesotan to age 32)

Post by TriciaG » September 14th, 2021, 12:45 pm

All I'll say to that is, don't try it. :roll:

EDIT: Nah, I'll say a little more. The PL specs aren't supposed to be letter-of-the-law. Yes, that would technically pass, but would anyone WANT to put out a recording like that?

DACSoft
Posts: 1524
Joined: August 17th, 2013, 8:51 am
Location: Connecticut, US

Post by DACSoft » September 14th, 2021, 12:47 pm

I wouldn't want to try ... but making sure I understand the requirements as written. :D :D

TriciaG
LibriVox Admin Team
Posts: 51886
Joined: June 15th, 2008, 10:30 pm
Location: Toronto, ON (but Minnesotan to age 32)

Post by TriciaG » September 14th, 2021, 12:52 pm

PL checklist:
For what types of error should I be listening?
There are some standard things that need to be listened for in each section, with possible additions depending on the nature of the text. The first post of each proof-listening thread should include a list of the requirements; if it doesn't, please ask the BC or soloist. By far the most common level of PL required is 'standard PL':
-- Do the intro and disclaimer match the instructions in the first post of the project thread?
-- Are there any long silences or pauses that ought to be edited out? If so, note the time.
-- Are there any repeats, or stumbles that ought to be edited out? If so, note the words and the time.
-- Is there excessive background noise, a constant hiss or buzz that detracts from the reading?
-- Did you need to turn your volume up unusually high to listen to the recording? (Or did you find the recording too loud?)
-- Are the correct closing words used at the end of the recording, as per the first post of the project thread?
-- Are there about 5 seconds of silence at the end of the file?
So a PL would probably flag a file that was half/half 70 & 99 dB. The link you quoted is assuming the volume is fairly level throughout.

DACSoft
Posts: 1524
Joined: August 17th, 2013, 8:51 am
Location: Connecticut, US

Post by DACSoft » September 14th, 2021, 1:04 pm

TriciaG wrote:
September 14th, 2021, 12:52 pm
So a PL would probably flag a file that was half/half 70 & 99 dB. The link you quoted is assuming the volume is fairly level throughout.
Thank you. I didn't interpret any assumptions in the statement in the link, so your perspective was helpful. :thumbs:

habasud
Posts: 283
Joined: June 22nd, 2020, 11:30 am
Location: S-H, Germany
Contact:

Post by habasud » September 14th, 2021, 1:37 pm

TriciaG wrote:
September 14th, 2021, 12:52 pm
PL checklist:

So a PL would probably flag a file that was half/half 70 & 99 dB. The link you quoted is assuming the volume is fairly level throughout.
The extremes you describe are undoubtfuly technical errors. This is not what I was aiming at. You probably can imagine that a ghost forcing the dead to come out of the sea will do this quite different from a woman talking to little kids. Just as you never would critizize if a peace of music has a soft intro and ends in a crescendo you should allow that a dramatic text may have a dynamic developement.
It may be true, that most listeners of today use our recordings while driving their car or on a train. But those recordings may be listened to in decades. Who knows, what the demands on audio are then? How do cars sound in 20 jears. Are there any?
It's true that every listening environement has it's own demands regarding to dynamics. But its not only a matter of function. Its also a matter of expression.
Life is What Happens To You While You're Busy Making Other Plans (John Lennon)

https://librivox.org/reader/14583

annise
LibriVox Admin Team
Posts: 34893
Joined: April 3rd, 2008, 3:55 am
Location: Melbourne,Australia

Post by annise » September 14th, 2021, 4:01 pm

Just to expand on my answer - if I was listening for pleasure and the recording jumped about like that in volume I would pull out my earplugs and never listen to another recording you made :D . Good readers learn to make emotional changes without changing volume . Just as actors do. But it is a skill that takes practice.
However, the fact I would never listen doesn't mean it should be rejected if it fits the guide lines - the PLer has the right to comment, the reader the right to not follow the advice.

Anne

Availle
LibriVox Admin Team
Posts: 20388
Joined: August 1st, 2009, 11:30 pm
Contact:

Post by Availle » September 15th, 2021, 12:19 am

I think I have found the section in question, it's in one of the German projects.
I don't have time right now to check this myself, but according to the DPL there, there is a jump of 8 dB somewhere in the file.

Even though overall this seems to even out quite nicely and fit our technical requirements, if these were two successive files, I would definitely comment on so big a change.
And according to this Q/A on quora, a change of 6 dB is a doubling in volume.
That's a bit much for a simple "artistic expression". :D
I wouldn't even know how to achieve this while recording, without post processing, and that's taking it a bit far in my opinion.

Plus, the section is apparently in a group project. While I would definitely turn off a solo like this and avoid the reader in the future, it's a bit unfair to the other readers who put in a lot of work as well. Nothing annoys me more when I see comments about a group project where people can't stand the first reader and don't listen any further.

So, while I am all for "artistic expression" and "it matches our tech specs, so it's fine" and "we accept all reading styles", in this particular case I would definitely side with the DPL.
Cheers, Ava.
Resident witch of LibriVox, inspired by
Granny Weatherwax: "I ain't Nice."

--
AvailleAudio.com

habasud
Posts: 283
Joined: June 22nd, 2020, 11:30 am
Location: S-H, Germany
Contact:

Post by habasud » September 15th, 2021, 3:03 am

Availle wrote:
September 15th, 2021, 12:19 am
So, while I am all for "artistic expression" and "it matches our tech specs, so it's fine" and "we accept all reading styles", in this particular case I would definitely side with the DPL.
I tried to avoid discussing the work of a special DPL. But as you share that point of view in this special case, it's doing nor harm...

Probably you're right with your estimations. I also invested much time in order to create an "outstanding" endpoint of the book - and definitelly I myself liked the impact of that scene and found it suitable. It had "Wumms" one might say in german :D o.k. :shock:

My opinions on dynamics in general I stated before. I could elaborate on that - but, although it would make my intentions clearer, it might sound a bit like "Grandpa telling 'bout the war" :mrgreen:

Thx - Claus
Life is What Happens To You While You're Busy Making Other Plans (John Lennon)

https://librivox.org/reader/14583

moniaqua
Posts: 1519
Joined: April 11th, 2013, 4:48 am
Location: Somewhere in the south

Post by moniaqua » September 15th, 2021, 10:13 am

Availle wrote:
September 15th, 2021, 12:19 am
I wouldn't even know how to achieve this while recording,
Ähm, well, äh, ja, ähm - I didn't measure it but I think it might have happened to me more than once :oops: Usually I level it out; you still get the dramaturgy (it still sounds different - when singing you back up from the mic on loud passages, too, that's the same effect).

habasud
Posts: 283
Joined: June 22nd, 2020, 11:30 am
Location: S-H, Germany
Contact:

Post by habasud » September 15th, 2021, 10:24 am

moniaqua wrote:
September 15th, 2021, 10:13 am
Availle wrote:
September 15th, 2021, 12:19 am
I wouldn't even know how to achieve this while recording,
Ähm, well, äh, ja, ähm - I didn't measure it but I think it might have happened to me more than once :oops: Usually I level it out; you still get the dramaturgy (it still sounds different - when singing you back up from the mic on loud passages, too, that's the same effect).
I'm not quite shure if I understand Ava right. What I did, is:

I recorded in 24bit with plenty of headroom on a Tascam DR-40. In Audacity I did make a measurement with the ReplayGain plugin. Then I limited that famous ghost about one dB more with the HardLimiter plugin so that the recording is not clipped when normalizing.

Claus
Life is What Happens To You While You're Busy Making Other Plans (John Lennon)

https://librivox.org/reader/14583

Post Reply