What's Your Time Budget for Recording?

Comments about LibriVox? Suggestions to improve things? News?
Kaffen
Posts: 2530
Joined: February 7th, 2006, 3:35 pm
Location: Simpsonville, SC
Contact:

Post by Kaffen » June 26th, 2008, 1:50 pm

As an engineer, I just have to submit repeated actions to scrutiny to determine their efficiency! And I have discovered that it takes me 3.6 minutes to produce 1 finished minute of reading. :roll:

My time budget works out this way:

1.1x Recording, including repeats to cover mistakes
1.5 x Editing
1.0 x Noise reduction, equalizing, exporting as MP3, ID3 tagging, uploading
3.6 x Total

This total obviously excludes reading & posting in the fora, noodling around Gutenberg, and other fan-boy stuff like listening to other contributors!

I've always had a preference for 20-minute chapters; they're a tidy piece of work that're done in about an hour. My last three books have leaned towards 40-50 minute chapters; now, that's starting to feel like real work!

I often batch chapters together, but I find that Audacity sometimes locks up if I'm trying to do operations on more than one file at a time, so I can't double-up much. I used to use Roxio's Sound Editor, which had no problem with that, but it didn't noise-clean, and when I got a more recent version that did, it had other faults that threw me to Audacity.

I wish we could batch-transmit files to the server. When I used to FTP to Anita's server, I could just line 'em up and let go! The current arrangement works very well, but requires tending if you have more than one file to send.

So, what's it take for you to make a minute of PD history??! :)
- Mark

"In narrating everything is simple, but it's the simple things that are difficult." (Apologies to von Clausewitz!)
Mark's Librivoxings

Jc
Posts: 3541
Joined: May 22nd, 2007, 10:25 pm
Location: Montreal, Qc, Canada

Post by Jc » June 26th, 2008, 3:03 pm

I'm pretty sure there was a thread on this, but I can't seem to find it ...

It takes me about 1.5 to 2 times the time of the final recording.
Put yourself in the Readers' Accents Table. See this post.
(Busy real life & traveling, sorry if not here often.)

Cloud Mountain
Posts: 4064
Joined: June 30th, 2006, 8:42 pm
Location: Jersey Shore, N.
Contact:

Post by Cloud Mountain » June 26th, 2008, 3:20 pm

:hmm:

A completed 40 minute chapter takes you 2 hours and 24 minutes to get to the uploader?
A completed 40 minute chapter takes you 40 minutes for Noise reduction, equalizing, exporting as MP3, ID3 tagging, uploading?

Wow, that's amazing!


I haven't kept a regular log, but a guesstimation, based on many hundreds of hours of experience, tells me that my per minute time budget for a completed 40 minutes is something like:

3.000x for Recording, including repeats to cover (imagined) mistakes or reinterpretations of the text,
3.000x or supplying alternate pronunciations for questionable words (usually foreign)
2.500x for Editing
0.200x for re-recording missing parts, mispronunciations and re-editing
0.125x for Noise reduction, equalizing, exporting as MP3, ID3 tagging, uploading
5.825x TOTAL (about 3 3/4 to 4 hours for a 40 minute chapter.)

Have not taken into account (did you?) pre-reading (rehearsing and becoming familiar with the text) and research for understanding, particularly pronunciations of archaic and foreign words/phrases. Except for long works, I usually cold read. Either way, things equal out.

(Now, my comments on photographic manipulation and time values there is going to take a LOT more thought and time than this response.)


Hi Jc! I think this is a much different approach than I've seen elsewhere, as it takes into account ALL time spent creating a particular file. I usually see discussions on just the WPM reading speed. But if it IS the same conversation, please let me/us know. I'd be interested to see just how dyslexic I is. Am guessing it'd take an engineer or an English teacher to get this nitpickity. :lol:

.
[url=http://librivox.org/newcatalog/people_public.php?peopleid=254]Alan's LV catalog[/url]

CalmDragon
Posts: 289
Joined: January 20th, 2007, 5:30 pm
Location: LA.CA.USA.NA.Earth.Sol
Contact:

Post by CalmDragon » June 26th, 2008, 4:00 pm

I'm new at this, but I've noticed this.
2x recording time
4x editing time
so a half hour recording of three to four thousand words will take me an hour to record and two hours to edit. Recording time is lengthened waiting for the cooling fan of my laptop to stop spinning. I ignore the fan during editing.

Starlite
Posts: 16606
Joined: April 30th, 2006, 2:17 pm
Location: Thunder Bay Ontario, Canada

Post by Starlite » June 26th, 2008, 4:19 pm

1.25 recording
1.5 editing (I proof the text at the same time as editing)
0.25 amplifying
0.25 tagging
No noise reduction needed.

Total: 3.25x :oops:

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

RuthieG
Posts: 22022
Joined: April 17th, 2008, 8:41 am
Location: Kent, England
Contact:

Post by RuthieG » June 26th, 2008, 5:12 pm

Jc wrote:I'm pretty sure there was a thread on this, but I can't seem to find it ...
Here it is:
http://librivox.org/forum/viewtopic.php?t=14113

I dread to think how long I take. The more recording I do, the longer I am taking to come up with the finished article. I am getting less and less easy to please with quality of sound and nuances of interpretation. :roll:

RuthieG
My LV catalogue page | RuthieG's CataBlog of recordings | Tweet: @RuthGolding

Jc
Posts: 3541
Joined: May 22nd, 2007, 10:25 pm
Location: Montreal, Qc, Canada

Post by Jc » June 26th, 2008, 5:27 pm

RuthieG wrote:
Jc wrote:I'm pretty sure there was a thread on this, but I can't seem to find it ...
Here it is:
http://librivox.org/forum/viewtopic.php?t=14113
That's exactly the thread I was looking for! Thanks for finding it! (I swear I had gone over help forum and I hadn't seen the thread. I must be going blind :S )
Last edited by Jc on June 26th, 2008, 5:27 pm, edited 1 time in total.
Put yourself in the Readers' Accents Table. See this post.
(Busy real life & traveling, sorry if not here often.)

Kaffen
Posts: 2530
Joined: February 7th, 2006, 3:35 pm
Location: Simpsonville, SC
Contact:

Post by Kaffen » June 26th, 2008, 7:27 pm

Cloud Mountain wrote: Have not taken into account (did you?) pre-reading (rehearsing and becoming familiar with the text) and research for understanding, particularly pronunciations of archaic and foreign words/phrases. Except for long works, I usually cold read. Either way, things equal out.

.
I'm filling in the gaps in my literary education with this gig. Only 3 of my solos have been books I've read before. The only one that required looking ahead was "Hans Brinker", where I had to get Anna to record Dutch words for me to learn to say. So most of my reads are cold, and if I sense I've missed the proper emphasis, I re-read the passage immediately. Reading poetry is different; I always rehearse poetry before recording.

My editing covers the whole file - I don't spike the recording to signal errors. My goal is to bore Anna by giving her nothing to catch! :wink:
- Mark

"In narrating everything is simple, but it's the simple things that are difficult." (Apologies to von Clausewitz!)
Mark's Librivoxings

Kaffen
Posts: 2530
Joined: February 7th, 2006, 3:35 pm
Location: Simpsonville, SC
Contact:

Post by Kaffen » June 26th, 2008, 7:29 pm

RuthieG wrote:
Jc wrote:I'm pretty sure there was a thread on this, but I can't seem to find it ...
Here it is:
http://librivox.org/forum/viewtopic.php?t=14113
Well, I missed that one too. :oops:
Still, considering how many replies have already been appended to this thread, it doesn't look like the topic was exhausted!
- Mark

"In narrating everything is simple, but it's the simple things that are difficult." (Apologies to von Clausewitz!)
Mark's Librivoxings

Cloud Mountain
Posts: 4064
Joined: June 30th, 2006, 8:42 pm
Location: Jersey Shore, N.
Contact:

Post by Cloud Mountain » June 26th, 2008, 8:22 pm

Jc wrote:
RuthieG wrote:
Jc wrote:I'm pretty sure there was a thread on this, but I can't seem to find it ...
Here it is:
http://librivox.org/forum/viewtopic.php?t=14113
That's exactly the thread I was looking for! Thanks for finding it! (I swear I had gone over help forum and I hadn't seen the thread. I must be going blind :S )
Ah, I hadn't see this. Thank you Jc and Ruthie!
[url=http://librivox.org/newcatalog/people_public.php?peopleid=254]Alan's LV catalog[/url]

Cloud Mountain
Posts: 4064
Joined: June 30th, 2006, 8:42 pm
Location: Jersey Shore, N.
Contact:

Post by Cloud Mountain » June 26th, 2008, 8:34 pm

Kaffen wrote: I'm filling in the gaps in my literary education with this gig. Only 3 of my solos have been books I've read before. The only one that required looking ahead was "Hans Brinker", where I had to get Anna to record Dutch words for me to learn to say. So most of my reads are cold, and if I sense I've missed the proper emphasis, I re-read the passage immediately. Reading poetry is different; I always rehearse poetry before recording.

My editing covers the whole file - I don't spike the recording to signal errors. My goal is to bore Anna by giving her nothing to catch! :wink:
Thanks. This is intriguing. Interesting how you do it.

I always read poems cold. They lose their life for me if I'm not discovering them. Which is why they take so long to record. I have to be certain what I'm reading and I almost always read them as prose. Which is to say, I don't want to be captured by a poems form —particularly very formal poetry. I'd rather meaning be the obvious element, not the form. I try to treat a poem like I treat dialog in a story: someone is speaking to me and I simply have to to have a conversation with the poet. Nothing fancy. Poets like Pope are the most difficult because everything is rhyming couplets. Argh! But I like that challenge sometimes too, particularly if I happen to agree with what the poet's saying, or it is particularly meaningful to me.

As for spiking, I never understood how to do it —or saw how it was a service to me. Doesn't make sense —I'm going to heart it all anyway. And with my error ratio, a 40 minute recording will have hundreds of spikes in it. Above all, I envy people who can read long passages errorless. Others are blessed with mild or greater forms of dyslexia. I think my greatest gift, or skill, is never being discouraged by continual misreadings. The only thing that discourages me is all of the sounds a city can make when you're in the reading zone.

Interestingly, be it poem or prose, I can read a phrase or line or sentence 3-4 different ways, trying to understand it and then convey it's meaning. Often I surprise myself when I finally, sometimes after 5-6-7 attempts to get the correct phrasing, believe I have it just right, but on editing the file, hear it all freshly and discover that #1 or #2 were actually closer to the mark than something I, upon reading, said to myself, with self congratulation: "Ah, Now I've got it! Finally! No need to read that one again."

Thanks to everyone contributing here. I'm really getting some wonderful insights. Don't think I'd get bored is every LV'er responded to the head post. Fascinating.
[url=http://librivox.org/newcatalog/people_public.php?peopleid=254]Alan's LV catalog[/url]

Steampunk
Posts: 2466
Joined: January 23rd, 2008, 1:41 pm
Location: Exile

Post by Steampunk » June 26th, 2008, 9:05 pm

Kaffen wrote: My editing covers the whole file - I don't spike the recording to signal errors. My goal is to bore Anna by giving her nothing to catch! :wink:
This has been a point of opinion in the editing process I've always found interesting. Several how-to's I've read, both at Librivox and elsewhere, recommend "clucking," using a clicker, thumping the mic, or otherwise "spiking" to indicate where you've made an error to mark it for later editing.

To my mind, this method could, at best, be a back-up reminder during the editing process. To rely on "editing at the spike" as the only or main method implies that no--or at least less--attention will be give to the rest of the file.

The potential effectiveness of this method, of course, is predicated on realizing that you've made an error at the time you've made it. My personal experience is that many (maybe half, give or take) of the errors I make are made without my noticing them while recording. These tend to be smallish mistakes like mispronunciations, transposition of words, and other similar syntax errors. In addition to spiking not being very helpful to me, I find that the interruption of flow caused by 1) noticing the error, and 2) remembering to spike it causes me to make more errors!

Again, I'm not saying that spiking is the wrong way to edit, just that it's wrong for me. Whatever works for others, works...and more power to them!

Oh, and just to keep on the subject here... I haven't crunched any numbers, but a 3:1 ratio of recording time + editing to actuial audio seems about right. Unless my dogs (Mastiffs!) are snoring, of course, then it can be much higher - and, if you count the occasional cussing jag, much much higher. :D


Jim


EDIT: Does pounding on the desk afther the umpteenth re-read count as "spiking"? If so...

RuthieG
Posts: 22022
Joined: April 17th, 2008, 8:41 am
Location: Kent, England
Contact:

Post by RuthieG » June 27th, 2008, 12:27 am

The only thing that discourages me is all of the sounds a city can make when you're in the reading zone.

Oh, yes, I know that feeling! I never realised how many vehicles passed my house until I started LV'ing. I am now "tuned in" and generally manage to pause in my speech until it has gone - and recognise the ugly thick line on the wav if I have missed one. :cry:

And I always edit the whole file (especially to remove nasty 'mouth' sounds and breaths, and to adjust the length of gaps between phrases), proof-listen against the text and then listen to the whole thing to make sure it reads smoothly. Nothing would make me more cross than to listen at some point in the future and realise that it could have been so much better.

RuthieG
My LV catalogue page | RuthieG's CataBlog of recordings | Tweet: @RuthGolding

Hokuspokus
LibriVox Admin Team
Posts: 8012
Joined: October 24th, 2007, 12:17 pm
Location: Germany
Contact:

Post by Hokuspokus » June 27th, 2008, 3:47 am

Cloud Mountain wrote: Thanks to everyone contributing here. I'm really getting some wonderful insights. Don't think I'd get bored is every LV'er responded to the head post. Fascinating.
Dito :D

Well, editing isn't the fun part, it always takes to long. But I never looked on the watch. You really made me curious about that so looked on the watch this morning.

:shock: It takes twice the time of the recording!

For one minute I need:
0,75 reading silent ( I don't do this all the time, but regret it later, when I leave it out)
1,25 recording
2,00 editing
0,50 (at least) exporting, importing to another progi, noise cleaning, polishing the sound and exporting again
0,25 for tags and uploading
4,75 in total

I too listen to the whole recording while editing. I can't get used to mark the mistakes. I produce only more mistakes when I try.
And I produce horrible clicks and smacks (must be horrible to listen to my unedited voice :wink: ) and leave really long silents between the sentences.

I thought about cutting out all my mistakes and things and past them in another track. But I was not jet masochistic enough to do so. :lol:

LeonMire
Posts: 2403
Joined: February 21st, 2007, 11:23 am
Location: Effingham, IL, USA

Post by LeonMire » June 28th, 2008, 5:16 am

The amount of time I take usually depends on how important the text is to me. If it's something I thought would just be kind of fun, all I do is skim the text for any difficult words, look 'em up, and then cold read. If it's something I care a lot about, I'll do all the things everyone else mentioned. I'll read the text several times, proof-listen against the text, listen to the whole file for errors and the best reading, and so on. My 9-minute reading of "The Raven" took longer than anything else I've done, because I care about the poem. Otherwise, it's about an hour per 20 minutes like everyone else.

Here's something I'd really like the developers of Audacity to create: Edit-as-you-go. If you make a mistake, you could rewind a few seconds, to the last natural pause, preview it, and start over, with the new take automatically overriding the old one. That would entirely cut out the Editing stage of producing an LV recording, which is like 2/3 of the time people spend.

I would even pay money for something like that, provided it also had good noisecleaning. But it'd be nice if it was free so that everyone could use it.

Oh, and by the way, I've increasingly found that warming up my voice helps eliminate a lot of the mistakes I make. I sometimes go through the exercises on this page to loosen up my lips and jaws. Or sometimes I'll just read a couple of paragraphs really slowly, articulating everything carefully. If I don't do this, I make tons of mistakes.
I remember how, in college, I got that part-time job as a circus clown, and how the children would laugh and laugh at me. I vowed, then and there, that I would get revenge.
-[url=http://www.deepthoughtsbyjackhandey.com/][u]Jack Handey[/u][/url]

Post Reply