Making your reading sound Great

[ArleneJoyce]

I was PL-ing one of JeremiahSutherland's recordings today, and really appreciated the quality of the recording. His voice sounded so warm and natural, with an even volume, and no mouth noise -- just totally enjoyable. I asked him about his mic setup, as I'm looking for a better one, and he PM'd me all this really great advice. So, with his permission, I share it with you:

I’m not sure that results are determined by just the microphone set up. So here is everything I can think of that might have an effect on the end product.

The Equipment:

I use a Bluebird microphone by Blue (http://www.bluemic.com/bluebird/). I bought it used as they are a bit pricey when new. The mike is pretty directional so there seems to be less chance of reflection from other surfaces.

For interfacing from mike to computer, I use a Saffire interface (http://us.focusrite.com/usb-audio-interfaces/saffire-6-usb): also bought second hand. The Saffire is very solidly built and seems to be good value for the money.

I live in a very quiet building, distant from traffic noises. I typically wait until about 8 PM before recording anything. In theory, this ensures that people are mostly settled in and not likely to make much noise.

I turn off everything that might cause background noise (fridge, dishwasher, etc).

The mike is on a stand and I position it 6-8 inches from my mouth, slightly off-centre so that I’m not speaking directly into it. There are a couple of metal screens attached to the mike to prevent pops and such. I’m not convinced they do much good, but they came with the mike.


Reading:

I spent 3 years volunteering for an organization that recorded news items for people with eyesight deficiencies. I’ve had a lot of experience, so I’m fairly comfortable when reading. One thing I learned was not to speak too fast.

I’ve also studied classical voice for the last 10 years or so. Aside from being able to hit the correct notes, teachers put a lot of emphasis on clear enunciation of the lyrics as an aid to timbre. It also helps listeners understand the lyrics.

Teachers also focus on being relaxed. If you’re stressed or in a hurry, your speech will suffer.

You mentioned “mouth noise”. I try to cut down on this by keeping my mouth and tongue fairly relaxed. I also drink water when I’m reading. Singers always avoid food/drink that tends to cause phlegm in the throat or mouth. Chocolate and milk are two things to avoid.

I think if you are stressed or speaking quickly, there is a tendency for the vocal cords and jaw and lip muscles to tighten up.

Don’t be afraid to pause the recording and rest for a bit.

I also keep the recording volume low so that I have to amplify the results so they conform to Librivox’s standards. This might contribute to having less excess noise showing up on the result.

Editing:

You mentioned that you wanted to be able to reduce the amount of time it takes to edit recordings: I take deep breaths between phrases and I’m a noisy breather. So there is a lot of breath noise between phrases.

When I’m listening to podcasts, I find the readers’ breathing to be quite distracting. Hence, when I proof listen my recordings, I reduce the time between phrases and cut out or silence a lot of the breath sounds.

I’m not certain if this approach shortens the editing process at all.

I also use the “notch filter” to cut out the 60 cycle interference caused by electricity feeds. This probably chops out some other sounds too.


Finally:

This is probably too much information. I’m not sure that any single thing is a recipe for good recording, but other people might find one or two of the items above to be useful for their work.

This is a useful conversation to have and it would be useful to hear from the those who are also professional voice over artists and readers about best practices.

[ganym3d3r]

I'm new and wondering how to read certain parts of text. For instance, when reading "XVII Meditation" do I say X-V-I-I-Meditation or 17th Meditation or Meditation 17?

How do I handle abbreviations? When I come across credits or sources are cited, how do I read that - at the beginning, middle, or end of the actual text?

In Project Gutenberg, there is extensive credit and agreement text both before and after the actual material, do I need to read all of it?

For example, I might like to read " THE BUDDHIST CATECHISM BY HENRY S. OLCOTT"
https://www.gutenberg.org/files/30216/30216-h/30216-h.htm
But I have no idea how to read all the annotated bits. There will numbers in the text which point to notes at the end further explaining the text. Do I read the numbers and the notation before I go on, even though the notes are at the end of the text? I'm confused.

I did skim a lot of sections but didn't see these questions addressed anywhere. If I'm in the wrong section to ask this, please forgive my mistake and point me in the right direction, since I'm new to all this. Thank you!

[gypsygirl]

I'm new and wondering how to read certain parts of text. For instance, when reading "XVII Meditation" do I say X-V-I-I-Meditation or 17th Meditation or Meditation 17?

How do I handle abbreviations? When I come across credits or sources are cited, how do I read that - at the beginning, middle, or end of the actual text?

In Project Gutenberg, there is extensive credit and agreement text both before and after the actual material, do I need to read all of it?

For example, I might like to read " THE BUDDHIST CATECHISM BY HENRY S. OLCOTT"
https://www.gutenberg.org/files/30216/30216-h/30216-h.htm
But I have no idea how to read all the annotated bits. There will numbers in the text which point to notes at the end further explaining the text. Do I read the numbers and the notation before I go on, even though the notes are at the end of the text? I'm confused.

I did skim a lot of sections but didn't see these questions addressed anywhere. If I'm in the wrong section to ask this, please forgive my mistake and point me in the right direction, since I'm new to all this. Thank you!

A lot of it depends on whether you are recording a solo or are contributing to a collaborative project. In a collaborative project the BC (book coordinator) will lay out guidelines for things like footnotes and abbreviations. If you have any questions, you refer them to the BC.

If you are recording a solo, then you are the BC, so you make those decisions yourself, and if you have any questions, you refer them to your MC (meta-coordinator). For something like the question about how to read the roman numeral, you have the choice of what you think sounds best to you (as a personal opinion, I'd probably go with one of your two latter options, "17th Meditation" or "Meditation 17")

In general, the rule is that we record the whole of the text, as written. Prefaces, introductions, etc, if written by someone other than the main author, can be said not to be part of the main text. In that case, it's up to you whether you decide to include those or not. In the text you referenced, several of those prefaces were added to later editions of the text. So even though they were written by the author, I'd say they fall in the category of "not part of the main text" and would be optional.

Footnotes are usually also optional and up to the BC's discretion. If the footnote is clarifying a point, or otherwise adding useful information, I would include it (usually something like "I'm reading the text. Footnote. This is the footnote. End footnote. I'm continuing the text"). If the footnote is more bibliographic in nature, then it's not usually necessary to include it.

We don't read Project Gutenberg's notes.

[RuthieG]

For instance, when reading "XVII Meditation" do I say X-V-I-I-Meditation or 17th Meditation or Meditation 17?

If you're reading it as part of a group project, then your Book Co-ordinator (BC) will mention the preferred way in the top post. If you're reading it as a solo, then it is your choice how to read it, though 17th Meditation or Meditation 17 certainly seems preferable to me .

How do I handle abbreviations? When I come across credits or sources are cited, how do I read that - at the beginning, middle, or end of the actual text?

Use your own common sense, really. Again, if it is part of a group project, your BC may have specified whether or not to read footnotes. Generally I read them in the place where they interfere least with the flow of the text. In the Buddhism example, I think I would read footnotes 1 and 2 after the answer to each question. It is often not necessary to read the footnotes, if they add nothing to the text except a reference. Something like "see p. 84" isn't much use in an audiobook .

You do not ever read the Project Gutenberg license material.

Ruth

Gypsygirl types faster.

[tovarisch]

You're asking the right questions, Cathy. Yet, there may not necessarily be a single "correct" answer to each of those. Readers do differ in their manners of handling some of the quirks of the texts.

I don't have statistics, nor can I claim encyclopedic knowledge of how things are handled here. I can only tell you about my own experiences.

Assuming 'XVII Meditation' is a chapter title... Chapter titles that have numbers in them I read "<number> . <title> .", so the text in your instance I'd read "Seventeen. Meditation." If it isn't, then please provide the context.

Credits cited -- I'd probably say or "cited from" and then the title, and usually at the end of the [sub-]clause, IOW, as soon as reasonably possible, without allowing the listener to lose the thread of the sentence. And I don't think the numbers matter when you name the source of citation.

Abbreviations - it depends. If they are obvious ones, I'd read the entire words. For example, you wouldn't read "etc." as "ee-tea-sea", would you? Or 'i.e.' as "Eye-ee"? The 'i.e.' ought to be read as "that is", I think. Same with other similar abbreviations. Of course, the name of the author should probably be read as it is written, "hen-ry Ess Ol-cott", even if you know what the 'S.' stands for.

I probably didn't answer all your questions. Perhaps you could split them into single ones, and have a separate thread of discussion for each. That way it's much easier to have a discussion...

[ganym3d3r]

Thank you, those replies did help me a lot!

[carteki]

The World of Sound

Nice listening too and talks about the things we should be taking into account, like a room where the sound of a clap can be heard for 3-4s is good for music, but not for public speaking.

[AOrlans]

and in the immediately preceding passage from Leaves of Grass we find:

My voice goes after what my eyes cannot reach,
with the twirl of my tongue I encompass worlds and volumes of worlds
Speech is the twin of my vision . . . . it is unequal to measure itself.

[lurcherlover]

OK, here is my take on dB levels that Librivox requires.

I know I should shut up, and keep my big mouth shut, but I do question why levels which are only 1 or 2 dB lower than zero dbfs are required? I know we are in the territory of 128kb.mp3 which is less than hi-fi standards (although acceptable for spoken voice reproduction) - but even here it is definitely best to keep levels down to -6dB - well this is the professional standard. (In fact, generally, the accepted level may often be -12dB below zero).

But OK - I know some people get a bit anxious when we talk about professional audio standards - but we should consider that in this digital age where audio can be (and should be ) of pretty high quality.

There are big discussions in the audio world about "Loudness Wars." This is where the pop world competes to get the absolute loudest album CD, with the sound permanently on zero dB - and where there is no dynamic range at all.

So, maybe we should re-think these standards.

[philchenevert]

OK, here is my take on dB levels that Librivox requires.

I know I should shut up, and keep my big mouth shut, but I do question why levels which are only 1 or 2 dB lower than zero dbfs are required? I know we are in the territory of 128kb.mp3 which is less than hi-fi standards (although acceptable for spoken voice reproduction) - but even here it is definitely best to keep levels down to -6dB - well this is the professional standard. (In fact, generally, the accepted level may often be -12dB below zero).

But OK - I know some people get a bit anxious when we talk about professional audio standards - but we should consider that in this digital age where audio can be (and should be ) of pretty high quality.

There are big discussions in the audio world about "Loudness Wars." This is where the pop world competes to get the absolute loudest album CD, with the sound permanently on zero dB - and where there is no dynamic range at all.

So, maybe we should re-think these standards.

Hi Lurcherlover (why do I picture love struck zombies shambling towards each other? anyway, an interesting forum name). I am not a professional which is why I am not certain what you mean by levels 1 or 2 dB lower than zero as the LV standard. We always talk about 89dB as our 'goal', with some variation allowed. Is 89 dB close to 0? or is it another standard that I'm not familiar with. In any case, it is always good to talk about improving the quality of our recordings so don't hesitate. Let me know what I'm missing here, OK?

[lurcherlover]

Hi Lurcherlover (why do I picture love struck zombies shambling towards each other? anyway, an interesting forum name). I am not a professional which is why I am not certain what you mean by levels 1 or 2 dB lower than zero as the LV standard. We always talk about 89dB as our 'goal', with some variation allowed. Is 89 dB close to 0? or is it another standard that I'm not familiar with. In any case, it is always good to talk about improving the quality of our recordings so don't hesitate. Let me know what I'm missing here, OK?

Lurcher is a cross breed of dog i.e. canine.

89 dB is rather meaningless and in professional recording circles they talk of zero (0 dB) as being the point where distortion begins if this threshold is crossed. (Into the red zone). Digital distortion is very ugly and instantaneous. That is why -6dB is fine, but +6db is horribly distorted. Normally it is recommended that you should record at a maximum of -12 dB - but sometimes peaks may rise to -6dB or even -3dB.

Because 24 bit digital recording has such a great signal to noise ration (i.e. the dynamic range is around 140dB - which means that the noise floor is so far away that you can record with no loss of signal to noise at even -30dB) it means you can leave a lot of "headroom" so if an unexpected very very loud sound occurs you will still not be in the red, or beyond zero dB. Having -30dB as a maximum peak, you would be talking of about 70dB and not 89 dB which is about 19dB lower.

With 128 kb mp3 files I would not suggest that you should record at this lower level, but maybe at -12 dB which means 82 dB in Librivox language.

The use of compressors has been mentioned in another post. What this does is to reduce the high peaks and increase the quieter (lower) peaks in the waveform. So it compresses the dynamic range. Radio broadcasts use compression so that for car radios the quietest parts are still loud enough to hear, and the loudest sounds do not blast you out of the car. But you lose the dynamic range and in classical music this is a big loss for those at home listening to music on a hi fi system.

To give you an example, CD's have a dynamic range of 90dB (not often achieved) and old vinyl records used to have 70db dynamic range - but never really achieved - as it was usually about 60 dB. Old shellac 78's were about 40 dB dynamic range.

So what I am saying is that in Librivox language it may be better to aim for 82 dB rather than 89 dB (plus or minus 1 or 2 dB).

Hope this helps.

[philchenevert]

Wow. yes, that does indeed help. I can't claim to undeerstand all of that but your point does seem valid. With that said, and with your knowledge, how would you translate that into a reading or scale that we all could use to evaluate 'loudness'? Almost all of our recorders have little technical audio training and use what is to hand, namely, free software like Audacity and our Checker program. Most admins too I might add including me.

Our scale of does not seem to match up at all with what you are saying is the professional scale of 0 dB being the loudest possible witho9ut distortion if I am understanding it right. How in the world did we end up with what looks like a simple linear scale where 89dB is louder than 86dB or 80 dB, and quieter than 95dB? Is it just because it is something we can understand? This seems directly opposite to the professional scale you mention.

Anyway, what technique do you use to give the dB level of a recording?

[tovarisch]

I wonder what prompted this discussion of loudness. What problem are we trying to solve here?

[lurcherlover]

Wow. yes, that does indeed help. I can't claim to undeerstand all of that but your point does seem valid. With that said, and with your knowledge, how would you translate that into a reading or scale that we all could use to evaluate 'loudness'? Almost all of our recorders have little technical audio training and use what is to hand, namely, free software like Audacity and our Checker program. Most admins too I might add including me.

Our scale of does not seem to match up at all with what you are saying is the professional scale of 0 dB being the loudest possible witho9ut distortion if I am understanding it right. How in the world did we end up with what looks like a simple linear scale where 89dB is louder than 86dB or 80 dB, and quieter than 95dB? Is it just because it is something we can understand? This seems directly opposite to the professional scale you mention.

Anyway, what technique do you use to give the dB level of a recording?

Using the peak meters. If you look at Audacity it has two meters along the top. (Not sure how accurate these are). They start at -54dB far left and the far right finishes at 0 dB. (One is for recording and the other for playback). Most professional meters can start at -64dB and will certainly go to about +12 dB or +20dB. When they get near 0 dB they change colour to yellow and once over 0 dB they are red for danger - i.e. distortion.

Of course the best instrument is our ears, so we listen for signs of distortion. But if we record too low, where the top of the green signal is only just reaching say -40 dB, then we have a quiet recording which will require the listener to wind up the gain control and which in turn brings up the noise floor to a more disturbing level.

The only difference with the way dB is used here is that instead of saying 89 dB is the level to aim for, it really is the minus sign which is missing. In other words -54 dB is very quiet, whereas -6dB is very loud (48 dB more sound). Zero dB is the loudest sound before distortion starts. Anything reading plus is in the distortion area. (i.e +6 dB)

If I was recording using 24 bit wave files I would not let the highest peak go over -10 dB and if it did I would reduce the recording level to be more on the safe side. Often I would record at a lower level than this, by about another 4 dB making the peaks hit a maximum of -14 dB (which is a normal broadcasting standard here in the UK). (It might be slightly different in the US broadcasting world, but still close to the UK standard).

There is a very good book in the US by an author called Bob Katz called "Mastering Audio" - but it is quite expensive. It is also rather technical. I'm sure there are other books and even articles on the Internet which explain more simply and certainly better than me.

By the way, the decibel scale is not linear. This also makes life in the audio world a bit more difficult. To quote "The decibel (dB) is a logarithmic unit used to express the ratio of two values of a physical quantity. One of these values is often a standard reference value, in which case the decibel is used to express the level [a] of the other value relative to this reference." (Wiki)

Don't worry, it gives me a headache as well!

[philchenevert]

I wonder what prompted this discussion of loudness. What problem are we trying to solve here?

Lurcherlover just brought this up

I know I should shut up, and keep my big mouth shut, but I do question why levels which are only 1 or 2 dB lower than zero dbfs are required? I know we are in the territory of 128kb.mp3 which is less than hi-fi standards (although acceptable for spoken voice reproduction) - but even here it is definitely best to keep levels down to -6dB - well this is the professional standard. (In fact, generally, the accepted level may often be -12dB below zero).

Since It is confusing to me I'm just trying to understand what he is talking about. I don't think he is trying to solve a problem, more questioning why we use what we use. Could be wrong here.