How long does it take you to produce a section?

[andyglover]

I'm curious how long other librivoxers take to do their recordings. For me from start to finish, a 10-30 minute long section will take around 4 hours of work. I'm a bit obsessive with accuracy and editing - I do at least a couple rounds of listening and error-correcting so when it gets to the DPL it might be an error free, easy "PL OK" for them. I don't say this to brag, I'm only curious, is anyone else as obsessive as I am about having a word-perfect crisp section before posting?

(This ruminant message brought to you after an evening of exhaustive editing)

[moniaqua]

I can not tell how long exactly I need, also it depends. But I am very sure that editing takes a couple of times the time I need for recording. Except if I record a role in a play, then lucky me doesn't need to edit that much because there is another poor guy doing this for me . Of course, I still need to get my errors out there.

[mightyfelix]

Nope, I'm not willing to spend that long. If I'm spending more than an hour on a 30-minute recording, something is wrong. (It used to take me longer, but that was before I discovered Audacity's punch and roll feature. It cuts editing time down to practically zero.)

I've come to accept that not every recording is going to be PL ok the first time, and that's ok. Proof listeners have an important job to do, and who am I to do it for them? The way I see it, they'll be listening to the whole section anyway, so I'm not saving them that much time by PLing my own sections before submitting. If they need to pause the playback for a few moments to make a note, it might take them 30 seconds or so to do it. If they give me a note to correct, it might take me a few minutes to correct it. That's a net gain, compared with the extra twenty minutes or whatever that I'd be taking to PL my own section. I'd rather spend that time recording the next one.

Also, in my experience as DPL, it can be somewhat lackluster never to be able to give any notes. Don't get me wrong, I can still enjoy listening to the book, but then my role just becomes one of listener, rather than my really being involved in the process of creating a good audio book. Kind of like my dental hygienist said at my last cleaning, "Go ahead and eat breakfast before you come in. Don't even brush your teeth. Gives me something to do!"

I'm not trying to change your mind necessarily, just hoping to explain my view on the subject.

[moniaqua]

just hoping to explain my view on the subject.

An interesting view indeed.

That's a net gain, compared with the extra twenty minutes or whatever that I'd be taking to PL my own section. I'd rather spend that time recording the next one.

Do I have to understand this like "I don't edit at all, I just throw it in and PL may get the errors."?

The way I see it, they'll be listening to the whole section anyway, so I'm not saving them that much time by PLing my own sections before submitting. If they need to pause the playback for a few moments to make a note, it might take them 30 seconds or so to do it.

Of course it saves time, for them and for me, when I edit my sections and this always includes some kind of "PLing". For the PLer, stopping, making the notes, writing the notes in the thread, going back to the file, find the spot(s), listen again, all takes time and I think more of it than just listening and saying "Wonderful, PL OK". For me, going back to the section, getting the sense again, find the spot(s), punching in, find maybe that the setup was different last time, re-upload etc. etc., all takes time and I thing more of it than editing properly in the first run. Plus, it is really hard to get the same sound again. So, for me doing all to - hopefully - get an error-free file on the first run has highest priority.

[Availle]

That's a net gain, compared with the extra twenty minutes or whatever that I'd be taking to PL my own section. I'd rather spend that time recording the next one.

Do I have to understand this like "I don't edit at all, I just throw it in and PL may get the errors."?

No. Punch-and-roll means you edit during the recording phase. Basically, whenever you notice having make a mistake, you stop the recording, rewind to before the mistake (the beginning of the sentence for example) and start reading and recording from there again.

Obviously cuts way down on editing during post-processing, but will increase recording time accordingly.

[tovarisch]

To answer the original question, my time has been at least 6-fold that of the finished section (when I cared to estimate).

I don't use Audacity, I use Reaper (except for small poems). I do repeat portions in which a mistake has been made (if I notice), after a click, rarely stopping, but sometimes pausing the recording (when a truck rumbles by our house, for instance). Clicks do help in editing, but not much, I have not developed the quick technique like some pros have, I suppose.

I still listen through the entire track, sometimes at higher playback speed (like x1.3 or so), removing repeats, re-recording what does not sound OK to me, adjusting pauses between speeches in conversations. My goal is word-perfect reading. It does not save the PLer much time, though. They still need to listen to the entire section and make their judgement, and if I let something get through, they will need to spot-PL (which does not really take as much time as PLing the whole section again, of course).

Which approach to take is up to you, and you only - it's your time, you decide how to spend it. Justifications can be given in all approaches, and the importance or precedence of every one of them is subjective.

[TriciaG]

I record (with all the repeats, flubs, sneezes, and coughs included), then go back and listen to the whole thing while I edit. It does take a long time - not 4 hours for a 30-minute section, though. If I had to estimate, I'd say, for a 30-min final file:
Original recording: 35-40 minutes
Editing: maybe an hour or a little more (about double the recording time)
Post-processing: 5 minutes (noise removal, exporting to MP3)

I don't listen to the whole thing after having edited, but I do jump back a few seconds on every edit to make sure it sounds smooth and the pace is okay.

I do like to listen to my whole recording while editing, for the pacing and such. One time I tried punch-and-roll and disliked it. And sometimes when I listen to the recording during editing, I'll understand a sentence or phrase that I didn't while recording! And then I realize I emphasized the wrong word or phrased it wrong, so I re-record that sentence or phrase to make it better. I suppose I might catch more of these if I did one last listen before uploading, but there is that effort-time-quality balance to maintain.

I will say that I don't submit a section that I don't think is ready to go straight to the catalog. I brush my teeth before going to the dentist. Any PL notes I get are edits I missed, which does happen occasionally, but I intend to submit a "perfect" recording the first time.

[philchenevert]

I do it the way Tricia does; just record with many repeats, coughs and flubs. Editing takes about twice as long as the finished track. E.G. a finished 30 minute track takes about 1.2 hours to edit.

[KevinS]

The question really is: How long does it take Kevin to edit out all his swear words?

[tovarisch]

The question really is: How long does it take Kevin to edit out all his swear words?

Time for editing out the swear words depends on how many swear words are introduced into the recording during re-recording of some passages... I wonder if there is a mechanism that listens (and tries to recognize spoken words) and then pauses the recording when it hears a swear word, after which it continues. Now, what if the text contains a swear word?

As I recall there used to be a system that, once trained to your voice/manner/accent, could match what you say to a set of known words/phrases to convert them into commands; I worked with one in the 1990s. Given that we often read from a text file, if that text is supplied to the voice recognition system, could it not catch all the mistakes made while narrating, thus helping editing and/or PLing?

[KevinS]

The question really is: How long does it take Kevin to edit out all his swear words?

Time for editing out the swear words depends on how many swear words are introduced into the recording during re-recording of some passages... I wonder if there is a mechanism that listens (and tries to recognize spoken words) and then pauses the recording when it hears a swear word, after which it continues. Now, what if the text contains a swear word?

As I recall there used to be a system that, once trained to your voice/manner/accent, could match what you say to a set of known words/phrases to convert them into commands; I worked with one in the 1990s. Given that we often read from a text file, if that text is supplied to the voice recognition system, could it not catch all the mistakes made while narrating, thus helping editing and/or PLing?

"Alexa, PL section 10, please."

[DACSoft]

I'm probably still a statistics outlier. This was a post from about 5 years ago, when I was first starting out.

... After recording my first solo, and two more in progress, I find that, on average (even for small colaborative pieces), it takes me about an hour per minute of finished recording to record, edit and polish (compression, noise removal, etc.) and upload. Partly because I find it difficult to read more than one or two paragraphs without making mistakes (I have to work on that), so I may record a paragraph twice (or more) and cut and paste the best versions. Also, I may be too much of a perfectionist (I am by nature) in cleaning up all the clicks and mouth noises. I tend to try for word perfect even though I set the PL level for standard (and I also PL all my recordings in addition to my DPL)....

What keeps me going is that I really enjoy recording. I expect (hope) that I'll get better with experience, but it is still an enjoyable hobby, and I feel I'm contributing to the greater good.

How has this changed in 5 years? Well now, on average, it only takes me about 35 minutes per minute of finished recording (I've cut my time in half, and editing is about 80% of the total time). All the rest is still the same, including the primary reason why I do this. I continue to PL all my recordings before submission, because I want to save PLs/DPLs time in their task. IMHO, they shouldn't have to spend needless effort finding obvious errors that I should have identified PLing my own work.

FWIW,
Don

[lurcherlover]

Like a lot of readers I also take about 40 minutes to record a 10 minute section, and about 40 minutes to edit it. In other words about 8 hours to record and edit 1 hour of recorded material - so a seven hour book with take about 56 hours plus the time to upload etc.

I think I may have started off reading quite well, back in 2016, but it has deteriorated in the 3 years I have been doing it, and I often wonder if I should continue reading, as my readings have become pretty poor, especially this year. So 2020 may be a make or break year where I either need to stop, or in the short term only record very short, short stories or a bit of poetry. The last book I recorded I finished in June, and it was just about OK. The present book that I have nearly finished recording has been a bit of a nightmare with lots of mistakes, mis-pronounciations, and quite poor reading, so I will be pleased when it's finished, and I haven't enjoyed doing it.

[mightyfelix]

Do I have to understand this like "I don't edit at all, I just throw it in and PL may get the errors."?

I may not have explained as well as I thought I did. No, it doesn't mean I don't do anything at all, it just means that I don't obsess over it. As I briefly mentioned, I now use punch and roll for most of my sections, which means that you edit as you record. So editing time is drastically reduced.

But even before I started using it, I would edit my file as usual, trying of course to remove all of the errors. But I never have (and probably never will) PLed my own recording after it is edited and before it is submitted. That's just doubling someone else's work, without much reducing their work. It seems to me a waste of time for two people to do the same job. I hope that makes my philosophy on this a bit more clear. I rarely do have PL notes, but when I do, I don't see this as a failure on my part, and it's generally a quick note and a quick fix. Basically, my philosophy boils down to "don't sweat the small stuff."

[LCaulkins]

I do my recording the way Tricia describes. I tried punch-and-roll, and it doesn't float my boat. It might if I gave it more practice, and I can certainly see how people like it. I just really like the freedom of freely recording. If I need to get up to refill my drink, do it, leaving the recording rolling. If a truck rolls by or the cat goes vocal-crazy behind me, stop, wait, start again at a good previous pause point, never pausing the recording. If I stumble on a word, realize the sentence has a different structure than I first anticipated, need to look up a foreign pronunciation - leave it rolling, and when ready start again back at a good pause point. For a 20-minute end file, it will take anywhere between 23 minutes and 40 minutes for me to get it recorded. All depends on my mistakes, changes of mind about inflection, and interruptions.

Editing - I go through from start to finish meticulously. Cut out the gaps or repeated portions. Snip out clicks, etc, between words, try to surgically remove bad clicks/plosives that are embedded in words (although most of those are let go as a leave-well-enough-alone situation), neg-amp overly loud breaths, etc. During this process, I'll backtrack the last little bit each time I've done something, to check the pacing and tone/loudness of splices. I try to get it all laid out nicely. Sometimes I'll do a punch-in recording if I have a bad mismatch of tone that can't be adjusted with pitch/tempo/amp type functions, or if I notice that I did a meaningful word swap. I work carefully this way all the way through. When I get to the end, I don't normally go back and play the entire track again. The method works great for me as long as I'm not distracted by someone, say my husband is in the kitchen talking to me, for example. Editing for a 20-minute end file will take anywhere from 40 minutes to 90 or so. I'm not a professional broadcaster; my recordings are never going to go as smoothly as theirs and will probably always need significant editing.

I'm only a couple months old on LV, so my process will likely evolve. I've been doing a lot of DPL and bit-work PL'ing, as well as recording. And from the PL side, I echo the sentiment that it can be boring when you proof file after file after file and there's nothing to note for folks. I'm not saying be lazy about your editing and upload half-baked recordings - just saying there's no need to go overboard with multiple replays in an attempt to always send in PLOK-first-time files. My philosophy is that I should be turning in my Reasonable-Best Effort, not killing myself pursuing Absolute Perfection.

That said, one of my recent files sure was jinxed or something, and with things that never should have been sent in to start with. It was a file from working while Scott was home, though, so the initial mess-up was clearly a distraction issue.