DR Recordings - for the Reader or Editor to "fix"?

[carteki]

I've been doing a bit of editing of DR's recently (both plays and DR's) as well as PL'ing. Generally when I point out during the PL process that there is stuff wrong other than missing lines (incorrect and variable volume, long gaps in speeches, very audible breaths) the BC comes forward and says "don't worry - that'll be fixed in editing".
Then the end of the project comes and the BC requests for editing help ... So I land up fixing these issues - and not just for 1 person ... most of them (based on experience).

This would be unacceptable when submitting a section in the non-dramatic works section. Why is it acceptable in a DR? I have edited non-DR sections where the reader has gone away and it is recoverable, but it seems that there is a general attitude (there are exceptions) of "the editor will sort everything out" in a DR making much more editing work than is necessary.

Any comments, feedback, thoughts?

[Availle]

When I read for a dramatic work, I never noiseclean and I never do anything about volume. This is best left up to the editor since my recording has to match all the others of that scene, one way or the other.

Everything else I produce so that it can be just dropped into the right spot. The silences inside my own parts are part of my interpretation of the character I'm reading, and I need to be happy with it. I would not like anybody else to touch them. As an editor I would not touch audible breaths either (unless the reader sounds as if he has just won a marathon). This is also part of the appeal of a DR, and if this character is a bit short of breath, well, so be it.

I think the interpretation of the character should be left to the reader; the editor is just there to cut things up and put it together. As I said, volume and noisecleaning should be done by the editor (and has always been in the old dramas I have been part of). Everything else should be the work of the reader.

[Carolin]

When I read for a dramatic work, I never noiseclean and I never do anything about volume. This is best left up to the editor since my recording has to match all the others of that scene, one way or the other.

i havent done a ton of editing work so im not an expert, but i found that if i had to noiseclean i had to noiseclean individual parts, since you cant get a noise profile that matches the whole edited section anyway. so i actually found it very inconvenient when readers hadnt noisecleaned their part (properly). same with volume. if everyone just uploads ~89 db files, the editor can paste them together much more easily.

[carteki]

Thanks Ava.

Noise cleaning -> Not one I find missing often. Unless there is a specific background noise requirement isn't it easier to match the pieces if they're all at a minimum? I know that the cleaner will work better on a single piece that has 1 type of background noise, rather than trying to reduce multiple.

Breaths - I have left in breaths when editing when the character calls for it. In one piece the narrator had LOTS of audible breathing which I had to edit out. I'd probably do the same if it was a young lady or someone we don't expect to hear breathing from.

Pauses / Speed - Something that I've found is that people can interpret pieces differently and you land up with 2 characters talking to each other at 2 completely different rates. Easily happens when you're recording in isolation (literally and figuratively!) and don't have the cues from the other person. The movie Being Julia has a great example of how the same text can be interpreted in different ways.

Volume - I get your point. Most editors I've discussed this with run the leveller and adjust the volume BEFORE starting to put the different tracks together. This has the unfortunate side effect of reducing the dramatic quality of the readers interpretation of the character.

[annise]

Well I've only edited one Dramatic thing - and I was the BC as well (and probably MC) so I'm an outsider in this . But as an outsider I have seen too many Dr's and plays that when the reached editing stage some of the readers were no longer around and there were things that needed fixing which could not be, so I'm in favour of as much fixing as possible to be done as soon as possible after submission

Anne

[TriciaG]

I've edited several DWs.

Regarding volume: what the Checker shows and what the ear hears can be very different. I noise clean and ReplayGain/adjust parts before editing together a scene, but I still often adjust volume as I go along.

I see these two steps (noise removal and volume) as easy to do, so I wouldn't bother the reader about them if they didn't have them to my liking upon submission. If they don't noise clean, they're also not OVER cleaning (although I think that's harder to do nowadays). I do think the reader should fix really loud background noise - they'll probably have to re-record - but otherwise I'm OK with no noise removal when I edit.

[AdeledePignerolles]

I personally prefer fixing things myself. I had one person that tried to make the volume the right level and ended up making a lot of clipping (messed up the sound). I also had someone who overnoisecleaned. So as long as the files aren't missing any lines and there' aren't any overbearing background noises like people talking that I can't take out I like to do everything myself.
So, when I edit, I open up the text on one side of the screen and all the other files around it. Then I hand adjust the volume so that it sounds the same. I've found that the wave height can't be used to measure, because some people sound really quiet and have the same size peaks as someone who sounds really loud. I also hand adjust the noise removal. I also generally leave bangs for me to take out as long as they aren't long or very frequent.

[pdyer]

For general readings, not necessarily DR recordings -- perhaps this should be a new topic -- what do you expect or want from the reader in addition to corrections, noise reduction, and cleaning up clicks and pops, plus meeting the levels using the guidance from Checker or Replay gain? I typically apply a standard EQ (standard for my microphone and voice) and mild, perhaps 2:1, compression. Does the compression help or hinder or does it matter?

[MaryAnnSpiegel]

My opinion: We are a group of volunteers. We are not producing material for commercial ventures. Each reader should be satisfied with how they sound in their own file. If you want to buy better equipment or do more processing to your file, have at it, but don't think that is required by anyone other than yourself.

As for editing dramatic works, Auphonics is a good tool for post editing processing, to level out volumes and even out the sound a bit. That may save you some steps on the front end. But if it sounds like everyone was recording their part on different equipment in a different place, well, they were. As long as all the lines are there, I expect it to sound a bit choppy as a listener because I know how they are put together.

And lastly, I do re-read the FAQs every now and then to make sure I'm not straying from the vision behind LV. They are a good reminder that whatever I expect of myself, my expectations of others should focus on them having an enjoyable experience as a reader in a non-commercial all volunteer environment. (These were originally written in 2005):

Do you mean *anyone* can record for LibriVox?
Yes! All you need is a way to record into mp3 (most PCs and Macs can do that with a microphone and some free software, we assume if you are using GNU/Linux that you can figure it out -- but just ask if you need help!)

Don't you have any standards?
It depends what you mean by standards. Our feeling is this: in order for LibriVox to be successful we must welcome anyone who wishes to honour a work of literature by lending their voice to it. Some readers are better than others, and the quality of reading will change from book to book and sometimes from chapter to chapter. But we will not judge your reading, though we may give you some advice if you ask for it. This is not Hollywood, and LibriVox has nothing to do with commercial media's values, production or otherwise. However: we think almost all of our readings are excellent, and we DO try to catch technical problems (like repeated text etc.) with our Listeners Wanted/prooflistening stage. Maybe you'd like to help?

You mean your readings might have mistakes in them?
Project Gutenberg has a 99% accuracy rate target for its texts; we aim to equal or better that. For a 20 minute audio recording, 99% is equivalent to 12 seconds of error. Count to 12 and see how long that is. So, we think by a pretty objective measure, even with a few mistakes, we're doing pretty well.

MaryAnn

[Carolin]

i think what we are learning from this thread is that the dpl bc and editor have to talk about this before a dramatic reading or play is launched to avoid trouble later on

[carteki]

i think what we are learning from this thread is that the dpl bc and editor have to talk about this before a dramatic reading or play is launched to avoid trouble later on

Agree with you, but I also think that the editors take on too much "I can fix this in editing". Readers generally take pride in their works and want them to be the best that they can be (which I find is the case with readers fixing my optional "perfection comments" and not just the PL comments, although there are exceptions who refuse to even accept PL comments). I much prefer editing a piece with well read, easily defined lines and consistent volume levels rather than having to modify the volume of each phrase as I go along. If I have to modify the volume constantly sometimes I resort to Levelling which can affect the quality of the sound or at times or I just "Normalise" regardless of whether the person is shouting or whispering - all of which impact the quality of the output, but per LV's policy is still acceptable.

Anon.

[TriciaG]

In my editing experience, I rarely had readers' volumes fluctuate much from line to line. Either that or I am less picky than you. As an editor, I'll compress and amplify; perhaps that evens things out enough that I don't notice.

For the record, I'd run away screaming as a reader if I had someone tell me to amplify this line by 2, de-amplify this one by 2, ad infinitum.

[AdeledePignerolles]

For the record, I'd run away screaming as a reader if I had someone tell me to amplify this line by 2, de-amplify this one by 2, ad infinitum.

Me too--but that's just what I do (and enjoy doing) when I edit. I think it depends on the PL and editor. Perhaps if the PL was to ask the editor (if they're not the same person) what level of fixes they'd like so the editor doesn't have to find people for fixes later?

[Elizabby]

Slightly different POV here - as a BC who does NOT edit, I give the readers enough feedback and ask for corrections so that the editor should simply have to assemble the voices. So I give feedback to readers and ask them to correct the pauses in long speeches, or take out repeats etc - just as I would for a "normal" chapter reading. (I don't require complete silence between speeches though, if there are extraneous sounds between speeches those will be cut out anyway.)

I do it this way because if I am not going to edit the project myself, I want to make it as easy for the editor as possible! Where the BC and the editor are the same person, there is significantly more latitude - although I personally would still leave as much of the fixing as possible to the reader. The main problems being that if something actually can't be fixed the reader may not still be around, and also that it takes enough time to edit the voices together without going through and effectively re-PLing the whole thing again as well.

One thing I have noticed sometimes happens if the editor is also the BC, and that is that changes which we planned can be forgotten. When I've been doing the final PL I have occasionally had the editor say "Oh yes, I had meant to fix that in editing" which I think is a problem that arises because of the length of time it takes from recording and original PL to the final edited version.

So yes, since I always know from the start that I will not personally be fixing anything during editing I try to make sure it gets done up front by the reader so the editor has a "clean" recording to work from. However, my experience has also been that some editors like to "fine tune" more than others. So when I launch a project, I've often worked with the editor before or we agree on what is acceptable. If in doubt, I get the reader to fix it, unless the editor (being the person who will actually do the editing) says that it is OK - which is usually just overall volume of the file.

[Rapunzelina]

For general readings, not necessarily DR recordings -- perhaps this should be a new topic -- what do you expect or want from the reader in addition to corrections, noise reduction, and cleaning up clicks and pops, plus meeting the levels using the guidance from Checker or Replay gain? I typically apply a standard EQ (standard for my microphone and voice) and mild, perhaps 2:1, compression. Does the compression help or hinder or does it matter?

I use compression when I want to amplify and some waveform spikes don't let me.
I think that the main (if not the only) requirement from the readers is to stick to the Public Domain text; correct tech specs that allow cataloguing and hosting at archive.org, and volume within the suggested range for consistency among librivox recordings.
The rest, I would say, is a matter of taste and choice of the reader, as long as it doesn't interfere or distract from the comprehensibility of the reading. Even background noise can be subjective, depending on hearing sensitivity or listening equipment.