How do you PL?

Post your questions & get help from friendly LibriVoxers
annise
LibriVox Admin Team
Posts: 38654
Joined: April 3rd, 2008, 3:55 am
Location: Melbourne,Australia

Post by annise »

Hokuspokus wrote:Hm, I begin to doubt if I do prooflistening careful enough.

I download the file and put it into a special folder where all relevant details are shown and check the tags.
Then I listen to it with winamp player. It shows bit rate, sample rate and mono/stereo on one glance.
I don't "check" the volume but I use a standard volume setting. When I need to change this setting much I know that the volume is to low/loud.

I only read the text along when it is asked or when I proof in a language other than German.

I never change anything in the recordings I proof unless I'm asked for. Sometimes I offer advice or help in a PM.
I personally would not be very happy if someone would change my recordings without asking me first.

And I must confess that I love to do ironing while prooflistening longer parts. I turn up the volume much more than I normally would and wind back when ever I here something strange. When it is still strange after listening again, I check the text. Some texts are strange.
Oh, and I note down the mistakes on paper.
If I change "German" to English this is exactly what I do - I never change a file unless particularly asked - well I have in an antique project when the reader had not been heard of since . My feeling was that the main purpose is so the file is listenable to without going "huh" so that is what I do - and I use paper too

Anne :)
Illiterati
Posts: 454
Joined: June 19th, 2007, 2:34 pm
Location: The Icy North

Post by Illiterati »

This is a great learning process here... :D

I'm especially grateful of the thoughts on editing files, which I - being the ignorant technocrat I am - hadn't really thought fully through. I now see the error of some of my ways.
lomond
Posts: 3127
Joined: July 7th, 2008, 6:22 pm
Location: Portland, Oregon, area

Post by lomond »

PaulW wrote:There are 453 projects that are in the database in an active stage (open, subscribed, proofing, validation) at this time. If we have 20 active MCs (I don't think that many are currently active), that works out to 22.65 porjects per MC ... which is a lot of projects, especially for new MCs (I don't even want that many!) [/shameless pitch]
Greetings -

Okay, Paul. I am a sucker for shameless pitches. What are the steps in becoming an MC?

Lomond
PaulW
Posts: 2272
Joined: February 24th, 2007, 5:00 am
Location: Austin, Texas

Post by PaulW »

A lot of experience (including being a BC), being helpful (all over LV), and being chosen by the current admin team. We've just added 4 new admins (russiandoll, RuthieG, smijen, Steampunk), but they're still trying to assimilate all of the new information, and are feeling their way around the MCing end of things. When I started MC'ing (just a few months ago, actually) I only took on a couple of projects at a time, and stuck to solos at first. It can seem a bit intimidating to a new MC...I felt that way a little at first.

My shameless pitch was not for new MCs, but for more detailed PL'ing. Some think that you can just listen to the recording for stumbles and repeats, but a conscientious PLer will also check the sample rate, the MP3 bit rate, stereo vs mono, is there enough silence at the end, make sure the intro, disclaimer and outro match what's requested in the first post, check the ID3 tags, and really take a good listen (and look) at the volume and check for DC offset using a program that displays the waveform of the file. That kind of in-depth, detailed check of the file saves much effort on the part of the MC at validation time. And it can help the reader correct technical problems with their recording setup, too, making them a more productive reader.

Maybe we need to copy into the templates the list of what to check for when PLing the work for each level of proofing, instead of just referring back to the thread about PL levels. I've seen many works that just state "Standard PL" without specifying exactly what should be listened for.

Got to go, I can continue this later if needed.
Paul
[b]DPL: [url=http://librivox.org/forum/viewtopic.php?t=12852]Brigands of the Moon[/url]; [url=http://librivox.org/forum/viewtopic.php?t=13151]Brain Twister[/url][/b]
TriciaG
LibriVox Admin Team
Posts: 60753
Joined: June 15th, 2008, 10:30 pm
Location: Toronto, ON (but Minnesotan to age 32)

Post by TriciaG »

I saw there was a wiki page designated for PL advice/tips, but no info is in it. :?

I'm a newbie PL'er (and reader, if truth be known). Generally I open the file in Real Player, then have the original text open and a Notebook page open. I have the Real Player as small as I can and still see the time, and arrange my windows so I don't have too much overlap.

I do check the MP3 bit rate, stereo vs mono, and ID tags (these I do in RealPlayer - do a Control-I for clip information). The sample rate is harder - it seems like the only way I can check that is by importing the file into Audacity and looking there.

I do check the intro and outtro and the opening/closing silences, too. Question on that: how close to 5/10 seconds does it need to be?

If I'm PL'ing more than one section in a work, I do listen to hear if the volumes are similar. Besides downloading and listening to all the sections, though, I'm not sure how one would go about seeing how uniform the volume is over the entire span of the work. If there were some standard or guideline we could go by on this (maybe going on the wiki page) about how to check the volume in a more objective manner, it'd be helpful.

Paul - what is DC offset?
School fiction: David Blaize
America Exploration: The First Four Voyages of Amerigo Vespucci
Serial novel: The Wandering Jew
Medieval England meets Civil War Americans: Centuries Apart
Leni
LibriVox Admin Team
Posts: 16269
Joined: July 27th, 2008, 9:10 pm
Location: Lexington, KY, USA

Post by Leni »

PaulW wrote:My shameless pitch was not for new MCs, but for more detailed PL'ing. Some think that you can just listen to the recording for stumbles and repeats, but a conscientious PLer will also check the sample rate, the MP3 bit rate, stereo vs mono, is there enough silence at the end, make sure the intro, disclaimer and outro match what's requested in the first post, check the ID3 tags, and really take a good listen (and look) at the volume and check for DC offset using a program that displays the waveform of the file. That kind of in-depth, detailed check of the file saves much effort on the part of the MC at validation time. And it can help the reader correct technical problems with their recording setup, too, making them a more productive reader.

Maybe we need to copy into the templates the list of what to check for when PLing the work for each level of proofing, instead of just referring back to the thread about PL levels. I've seen many works that just state "Standard PL" without specifying exactly what should be listened for.
Actually, I think the template in the "Listeners and Editors Wanted FAQ" needs to be updated then. Because it does NOT specify a lot of things that, apparently, the PL'er should do, from what I gather by reading this thread. Here is what it says:
Standard Proof-listening
-- 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 stumbles or repeats 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 5 seconds of silence (10 if the recording is more than 30 minutes long) at the end of the file?
Some of the problems pointed out by MCs and other people in this thread will not be noted by the most helpful PL'er who sticks to these items - and the wiki doesn't add much to that either. The volume, for example - they just ask the PL'er to say if they had to turn the volume up or down, which is pretty simple and straightforward, but doesn't mention anything about checking waveforms, spikes, etc. Also, that is VERY vague - and in another thread (I think it was the Glorious Qu'ran one) I have read about now many dB it should be, minimum. That's a very useful information, I wish I had it before, though. Another example: nowhere, in the FAQ or in the wiki, they say I should check bit sample, or even if it's mono or stereo! So, I'm sorry, but I think it's unfair with the PL'ers, who will follow exactly what's asked for Standard PL thinking they are doing a great job, to later find out the poor MC had to download and fix all the files by himself, and make it look like the PL'er didn't do a good job. That seems to me a huge waste of time and energy, both the MC's and the PLer's, simply because no one asked the PLer to check some of the things that are being mentioned here.

Don't get me wrong, I think the PLer can and should check those. It's not very complicated, and takes one more load off the MC's shoulder, which is obviously what we all want. I don't want any MCs feeling overloaded, mainly because I think that if LV becomes a chore, it isn't fun anymore, and no volunteer work should be anything less than fun and rewarding. But, personally, only recently I found out one project I PL'ed following the PL requirements strictly had some of its files recorded in stereo. I felt like an idiot, because, since I had downloaded all the files, I could have pointed it out and even edited some myself, if needed. But I didn't because no one told me to, and, as all newbies, I strived to do exactly what I was told to, and now I feel my work was poorly done, and the MC will have more unnecessary work.

So, all these tips here are nice, but I think we should avoid letting them be forgotten in the middle of all other threads that will be written soon. I suggest these things be added to the FAQ (and possibly the wiki), where they can be accessed directly by all new PLers.
Leni
=================
PaulW
Posts: 2272
Joined: February 24th, 2007, 5:00 am
Location: Austin, Texas

Post by PaulW »

Sample rate can be found in several places. If you are running Windows XP (NOT Vista), you can set the folder's view options to "Detailed", then select View->Choose Details. On the menu that pops up, you can select Title, Artist, Album Title, Bit Rate and Audio Sample Rate (not size). XP will only display to the nearest full kHz, ie it will show 44, meaning 44.1 kHz. Same thing as 44100 hz, which is what Audacity shows. Vista, unfortunately, took out the Audio Sample Rate detail option. I'm pretty sure there's a way to get the information in Mac OSX, but not having used a Mac since OS 9, I don't know how to get it.

You can also import the file into iTunes, and the summary tab under Get Info will show the technicals as well.

DC offset is where the center of the recording is offset from the center, or zero, line. It happens most frequently in recordings done with an analog microphone connected to the sound card of the system. If offset is there, it causes an audible click at the beginning and end of the recording, and every time there is an edit, where the audio instantly goes from the offset to the zero and back to the offset. You can visually see DC offset in the Audacity waveform...the blue waveform line is either higher or lower than the black zero line. It's most visible when there's silence, becuase you have a black line either below or above the waveform line.

DC offset can be easily in Audacity with the Normalize effect. Just select the entire recording, choose Effect->Normalize, check the "Remove DC offset" checkbox, UNcheck the "Normalize to" checkbox, and click OK. Then re-export the MP3 file when it's done. Fast, simple, and I recommend that it be done by every reader as the first step after saving the recording and exporting a clean WAV or AIFF file before any editing is done.
Paul
[b]DPL: [url=http://librivox.org/forum/viewtopic.php?t=12852]Brigands of the Moon[/url]; [url=http://librivox.org/forum/viewtopic.php?t=13151]Brain Twister[/url][/b]
annise
LibriVox Admin Team
Posts: 38654
Joined: April 3rd, 2008, 3:55 am
Location: Melbourne,Australia

Post by annise »

TriciaG wrote: Paul - what is DC offset?
Hi Tricia
I agree with what you have said - and I also have found that quite often if I say - "the recording was soft" the BC/MC leaps in and says "that doesn't matter we will fix it at validation " so now I often don't say

Anne
chocoholic
LibriVox Admin Team
Posts: 14190
Joined: January 16th, 2007, 9:23 am
Contact:

Post by chocoholic »

Stereo is actually OK. It just takes up more space than mono in an Audacity file.

This is how I PL about 2/3 of the time:

Download and quickly check bit/sample rates
Transfer file straight to mp3 player
Listen while I go about my business (sometimes in the car)
Jot down times of errors (sometimes at traffic lights)
Post in the thread

The other 1/3 of the time I PL in front of the computer reading along with the text.

That's it. I keep a few completed files on my computer so I have something to compare volume to, in case I think a file might have been a bit soft.

I don't worry much about ID tags much anymore since we now have the power to transfer them directly to the files from the Magic Window. I like having readers put them in just in case somebody reads the wrong chapter by accident, but it's not such a big deal if they aren't uniform.

I usually don't worry about spikes and waveforms and offset either. But if a file is soft and I have to amplify it, THEN I have to pay attention to those things because they mess with the amplification.

I do always pay attention to intros and endings. As for how close something needs to be to the 5 or 10 sec silence -- actually I don't know because I'm not sure how much gets cut off by the players that do that. I try to stay pretty close to 5 or 10 sec since I'm not sure.
Laurie Anne
annise
LibriVox Admin Team
Posts: 38654
Joined: April 3rd, 2008, 3:55 am
Location: Melbourne,Australia

Post by annise »

With the end silence , at what point would you ask someone to fix and reload ?

Say for the 10sec time at 9 sec/11sec or when ? I had always assumed that the 5 and 10 sec was a sort of guide line and say if the file was 29 mins long and 10 secs was used , considering that if the file had been a minute longer the 10 secs would be correct to comment would be nit-picking . But have I been wrong?

Anne
Starlite
Posts: 16548
Joined: April 30th, 2006, 2:17 pm
Location: Thunder Bay Ontario, Canada

Post by Starlite »

annise wrote:With the end silence , at what point would you ask someone to fix and reload ?

Say for the 10sec time at 9 sec/11sec or when ? I had always assumed that the 5 and 10 sec was a sort of guide line and say if the file was 29 mins long and 10 secs was used , considering that if the file had been a minute longer the 10 secs would be correct to comment would be nit-picking . But have I been wrong?

Anne
I think what you are doing is ok. I am not hard and fast on silence. 4.5 seconds, I won't make the reader fix, likewise 11 seconds. When it gets annoying is when there is 11 seconds at the end of one file and 5 seconds silence at the beginning of the next. That equates to 16 seconds silence. :evil: I'm just too darned impatient to handle that!

We only need a half second silence at the beginning of a file! We add the ending silence because some players cut off the end. One of our diligent volunteers actually studied it and found the longer the file, the more it was chopped off.

Esther :)
"Reasonable people adapt themselves to the world. Unreasonable
people attempt to adapt the world to themselves. All progress,
therefore, depends on unreasonable people." George Bernard Shaw
Illiterati
Posts: 454
Joined: June 19th, 2007, 2:34 pm
Location: The Icy North

Post by Illiterati »

[ snipped ]
Last edited by Illiterati on October 18th, 2008, 8:46 pm, edited 2 times in total.
PaulW
Posts: 2272
Joined: February 24th, 2007, 5:00 am
Location: Austin, Texas

Post by PaulW »

Starlite wrote:
annise wrote:With the end silence , at what point would you ask someone to fix and reload ?

Say for the 10sec time at 9 sec/11sec or when ? I had always assumed that the 5 and 10 sec was a sort of guide line and say if the file was 29 mins long and 10 secs was used , considering that if the file had been a minute longer the 10 secs would be correct to comment would be nit-picking . But have I been wrong?

Anne
I think what you are doing is ok. I am not hard and fast on silence. 4.5 seconds, I won't make the reader fix, likewise 11 seconds. When it gets annoying is when there is 11 seconds at the end of one file and 5 seconds silence at the beginning of the next. That equates to 16 seconds silence. :evil: I'm just too darned impatient to handle that!

We only need a half second silence at the beginning of a file! We add the ending silence because some players cut off the end. One of our diligent volunteers actually studied it and found the longer the file, the more it was chopped off.

Esther :)
Ditto. I'm flexible about the ending silence as well, if it's close to the needed amount. I will mention it if it's a 30+ minute file and there's only 3-4 seconds at the end. or if there's 15+ seconds.

As far as the stereo v. mono issue, I'm one who prefers stereo, but I personally have proofed a section (don't ask me which one, I don't remember) where the left and right channels were significantly different in volume. In fact the left channel was only about 25% of the level of the right channel, both in listening, and visually in the Audacity waveform. That definitely needs mentioning, and if you listen using your speaker system, you might not catch that kind of error. Stereo imbalance and background noise can sometimes be hard to pick up through speakers, that's why we ask for proofers to use headphones or earbuds.

Volume levels: I just recently found out that the mp3gain program used by the validator did not actually modify the sound levels, and had also been taking it for granted that the validator actually fixed the levels. I can't speak for any of the other MCs, but I don't think many of us knew that, we just assumed that when we used the "make all tracks the same", it really modified things.I, for one, would appreciate the PL letting me know about volume problems on any project that I am MC or BC for .

Technicals (bit rate, sample rate, offset, etc): No, these aren't in the list of things for standard level proofing, but the finished product will be much better if they are caught early on, while the reader still has his or her original uncompressed recording and can re-export the mp3. If you take an mp3 that was recorded at 22.05 kHz, and exported at 64 kbps, then import it, fix the problems and re-export it as a 128 kbps mp3, it's not going to sound any better than the original file. Whereas if the reader re-samples the original and re-exports as 128 kbps, it may sound considerably better. It may be time for us to update the list of things to check for when proofing.
Paul
[b]DPL: [url=http://librivox.org/forum/viewtopic.php?t=12852]Brigands of the Moon[/url]; [url=http://librivox.org/forum/viewtopic.php?t=13151]Brain Twister[/url][/b]
catchpenny
Posts: 1020
Joined: October 28th, 2007, 10:20 pm
Location: The here and now

Post by catchpenny »

Personally, I'd like the option to have the Pl'er note mispronounced words. I said "The king's ad-dress" the other day instead of "The king's a-ddress." (I wonder what his zip code is?) And once I said Ill-a-noise instead of Ill-a-noy. Those I caught listening to my own recording, but there is one time I said a-size instead of a-seize (or was it the other way around?) and looked it up in the next recording. I also say yo-grit instead of yo-ghurt, but that is my family dialect.
Anyone can read accurately. [i]I[/i] read with great expression.
Sibella
Posts: 4858
Joined: April 28th, 2007, 7:16 pm
Location: Kingdom of Meridies
Contact:

Post by Sibella »

I agree with Catchpenny. Sometimes a "mispronunciation" is a regionalism, and sometimes it's just wrong. For example, there doesn't seem to be any way to mispronounce QUAY, but there are a number of others I've heard mispronounced...
[size=100][b]It is the mark of an educated mind to be able to entertain a thought without accepting it. -Aristotle

[/b][/size]
Post Reply