Making your reading sound Great

Post your questions & get help from friendly LibriVoxers
lurcherlover
Posts: 1206
Joined: November 10th, 2016, 3:54 am
Location: LONDON UK

Post by lurcherlover »

bubbaspeare wrote:
I did my Librivox test, got distracted, and still haven't done the short poem Librivox Admin recommend doing. Would like to give Reaper a try for that :D

Cheers

Bubba
A short poem would be a good test for using Reaper from the free trial download. You could record it several times and edit between different versions. If you want advice on shortcuts, setting up a customised version, or anything else, as well as the forum at http://forum.cockos.com/index.php I can give you tips as well.

You may need to import the edited file into Audacity (which takes just seconds) to then export it again as an MP3. This is very quick and easy.

If you download the Checker it will take your final file and check it to pass it or fail it on certain criteria. Once passed Librivox will accept the file (subject to no reading or editing errors) and you are home and dry.

Interesting reading of Shakespeare's Sonnet 130 - surprise after the cockney accent in the intro. I liked the slow pace too, that's hard to do!

Peter
bubbaspeare
Posts: 30
Joined: December 17th, 2016, 10:21 am

Post by bubbaspeare »

annise wrote:As far as LibriVox is comcerned , the big advantage in using Audacity is that it is free and most people on the forum can help you with any problems you have. If you aren't new to recording and use another program - go ahead with it. But if you are new , use what's on offer , find out how you record and then look for something that suits you .

Anne
Oops! did it again. Sent reply as a pm instead of right here. Apologies. Anyway, I said:

Cheers Annise, you are all so helpful and friendly.

A big thanks

Bubba
bubbaspeare
Posts: 30
Joined: December 17th, 2016, 10:21 am

Post by bubbaspeare »

lurcherlover wrote:
bubbaspeare wrote:
I did my Librivox test, got distracted, and still haven't done the short poem Librivox Admin recommend doing. Would like to give Reaper a try for that :D

Cheers

Bubba
A short poem would be a good test for using Reaper from the free trial download. You could record it several times and edit between different versions. If you want advice on shortcuts, setting up a customised version, or anything else, as well as the forum at http://forum.cockos.com/index.php I can give you tips as well.

You may need to import the edited file into Audacity (which takes just seconds) to then export it again as an MP3. This is very quick and easy.

If you download the Checker it will take your final file and check it to pass it or fail it on certain criteria. Once passed Librivox will accept the file (subject to no reading or editing errors) and you are home and dry.

Interesting reading of Shakespeare's Sonnet 130 - surprise after the cockney accent in the intro. I liked the slow pace too, that's hard to do!

Peter
Hi Peter, I've enjoyed reading your posts on Reaper and ribbon microphones. You certainly know your stuff. Anyway, as it's now up and running, and as there is a very active Reaper community... I think I'll go with it. I'm not techy like you, but I'm sure I'll get there in the end. The main issue I find with performance, is not merely 'doing it', but doing it efficiently. In this case, I've become aware that for every minute I put into producing audio, some clever clogs out there (you :wink: ) is getting the same result with only 15 seconds of effort.

I can see the simplicity of having everyone on the same DAW (I know Audacity ain't really that). Especially when thing go wrong, audio disappears from sight, and we start crying like babies.

Off-Topic: I'm staring at a pristine 1960's era Uher Report 4000 -L reel-to-reel tape recorder (portable and fits into a small daysack). Belonged to my mother in her youth. Takes 13cm reels of tape. It's the strangest thing; listening to your mother reading sonnets/plays as a young woman.

Sonnet 130... was trying to steer clear of # 18 :)

Thanks for the offer. Yep, I might have to call upon your expertise one day

Regards

Bubba
RecordingPerson
Posts: 229
Joined: May 18th, 2017, 5:53 am

Post by RecordingPerson »

Just as a quick one, as someone who is used to dBs, I do find the '89dB' very confusing as it doesn't really mean anything.

dBFS = deciBels Full Scale as mentioned by lurcherlover are IMHO a better way of describing volume levels, and are widely used. We could use peak levels, RMS, or other measures.

I'm new here, so don't want to do the arrogant b'tard thing. So, I deleted most of a post that I had written.
lurcherlover
Posts: 1206
Joined: November 10th, 2016, 3:54 am
Location: LONDON UK

Post by lurcherlover »

RecordingPerson wrote:Just as a quick one, as someone who is used to dBs, I do find the '89dB' very confusing as it doesn't really mean anything.

dBFS = deciBels Full Scale as mentioned by lurcherlover are IMHO a better way of describing volume levels, and are widely used. We could use peak levels, RMS, or other measures.

I'm new here, so don't want to do the arrogant b'tard thing. So, I deleted most of a post that I had written.
It can be confusing if you are used to the methods normally used in the recording/engineering world.

I think the 89dB reference is really a peak reading of -11dB (although I could be wrong). Librivox allows a maximum peak reading of about 92dB which i would say is approx. -8 dB peak reading. In other words, have a at least -8dB headroom and keep away from zero DBFS by 8 to 11 dB. You can go as low as -14dB but the PL may then ask for 3 or 4 dB to be added. The minimum high level of recordings is so that people listening on trains,busses, and any noisy environment can hear reasonably well.

The CHECKER program which can be downloaded will give you readings on MP3 files that will cover all the requirements including levels, clipping, file format etc - and can be quite useful.

I keep my recordings at the top level between -8dB and -11 dB so that I avoid having to boost the gain in Audacity when I convert the files from Reaper, where I edit.

Hope this helps.

Peter
RecordingPerson
Posts: 229
Joined: May 18th, 2017, 5:53 am

Post by RecordingPerson »

I am used to professional audio measurements. dBFS peak, dBFS RMS, crest factor, etc. I've only been using Audacity for the recordings I've done for here, but am used to using ProTools and Logic Pro X.

I've got the checker program, but haven't yet reverse engineered how it calculates its dB measure. Is source code available? At present I do a bit of trial and error to get into the correct range.
lurcherlover
Posts: 1206
Joined: November 10th, 2016, 3:54 am
Location: LONDON UK

Post by lurcherlover »

RecordingPerson wrote:I am used to professional audio measurements. dBFS peak, dBFS RMS, crest factor, etc. I've only been using Audacity for the recordings I've done for here, but am used to using ProTools and Logic Pro X.

I've got the checker program, but haven't yet reverse engineered how it calculates its dB measure. Is source code available? At present I do a bit of trial and error to get into the correct range.
I've no idea about the source code, but I also wonder how it works. I would think it just measures the peaks and if they fall in the range -8DBFS to -14DBFS and do not clip then it passes the file, everything else also being OK. I imagine the source code will be fairly simple and quite brief.

Reaper is a free download and it can be evaluated of several weeks and even months in its full state, and then for small business use it is relatively cheap at around $50

Peter
GrayHouse
Posts: 638
Joined: October 6th, 2012, 3:27 pm

Post by GrayHouse »

The loudness measure used by Librivox is the ReplayGain standard. It's used by lots of audio programs and mp3 players. It wasn't invented by Librivox. It's a fairly sophisticated measure of perceived loudness. The full specification is here.

The volume-checking part of the Librivox Checker program is an implementation of ReplayGain - more details and full source code can be obtained here.
-Ian
lurcherlover
Posts: 1206
Joined: November 10th, 2016, 3:54 am
Location: LONDON UK

Post by lurcherlover »

Many thanks. Very interesting.
RecordingPerson
Posts: 229
Joined: May 18th, 2017, 5:53 am

Post by RecordingPerson »

Yes, that makes sense now.
Algy Pug
Posts: 6962
Joined: December 26th, 2009, 10:07 pm
Location: Perth, Western Australia

Post by Algy Pug »

I don't use Checker or Replay Gain. I just run a couple of rounds of compression and then normalize to a peak level of -3.0 db (commercial audiobook standard). Keep it simple.
Algy Pug

My Librivox page



_________________________
tovarisch
Posts: 2936
Joined: February 24th, 2013, 7:14 am
Location: New Hampshire, USA

Post by tovarisch »

Algy Pug wrote:I don't use Checker or Replay Gain. I just run a couple of rounds of compression and then normalize to a peak level of -3.0 db (commercial audiobook standard). Keep it simple.
Simplicity is in the eye of the beholder. I don't use either compressor or normalizer, but instead ask ReplayGain to tell me how much to amplify, then do amplify by that amount, and noise-clean. Occasionally my peaks cross the -3.0 level, I am sure. But then I do not read commercial audiobooks :wink:
tovarisch
  • reality prompts me to scale down my reading, sorry to say
    to PLers: do correct my pronunciation please
Algy Pug
Posts: 6962
Joined: December 26th, 2009, 10:07 pm
Location: Perth, Western Australia

Post by Algy Pug »

As a rule, I don't noise clean - unless I am editing other people's recordings.
Algy Pug

My Librivox page



_________________________
TriciaG
LibriVox Admin Team
Posts: 60512
Joined: June 15th, 2008, 10:30 pm
Location: Toronto, ON (but Minnesotan to age 32)

Post by TriciaG »

It just goes to show that no one system or set of effects will work for everyone. :)
Serial novel: The Wandering Jew
Medieval England meets Civil War Americans: Centuries Apart
Humor: My Lady Nicotine
lurcherlover
Posts: 1206
Joined: November 10th, 2016, 3:54 am
Location: LONDON UK

Post by lurcherlover »

I do hit peak values of -3dB occasionally - my recording machine often tells me I'm in the red during playback (but not so much when recording) - but in Reaper the values are not peaking quite so high. The most important thing is to listen for any distortion rather than watch meters which are only a rough guide. If you are close to zero DBFS in Reaper one can reduce the render by a dB or so on the master slider. This ensures that nothing goes over, and yet you have a pretty full on recording. I do not use compressors as they alter the sound in varying degrees, depending on the compressor. I just try to keep the peaks and troughs not too far apart by avoiding extremes of low voice and extremes of loud voice. It's not easy though. There is little need for noise cleaning as it is a full blooded recording with a good S to N ratio, and the environment is quiet at 6.00am or after midnight.
Post Reply