Freesound Forums

Hi,

I'm looking for voice samples where numbers (at least from 1 to are spoken. It would be great if the following considerations were met:

- it is for a click track, so the numbers should be spoken out quite clearly, precise and neutral (a rhythmic "feel" would be important)
- male / female speaker
- several languages - english, german, french, italian, ...

I've looked already here, but couldn't find any pack with something. Does anyone knows/wants to do something?

Greets,

Joao

that was ( at least from 1 to 8 ).

http://www.naturalvoices.att.com/demos/

Sound like a great idea, it would be fun to do it in lots of different languages too. I'm pretty busy at the moment, but If i have the chance I will record een tot en met acht (1 thru in dutch.

I have just uploaded "Numbers_Male.flac" with the numbers 1 to 20 spoken in English.
Will try and get my wife's assistance later for a female version.

Thank you for your work, acclivity. But I would like to make some suggestions, if I may. I find your "interpretation" very good, neutral enough (which in this context is good). But there's an important point that could still be improved. The samples will help musicians to follow complex scores (changing tempos, irregular values, etc). It is therefore important, that they are quite precise, with a certain rhythmic edge. If you have the will to do another round, I guess that what could make it better is for example to speak the numbers a bit faster (but keep the linear expression you already achieved).
Think of it as if you would be on a workout session, and the trainer would count the numbers to you to give you the tempo - always accelerating a bit. When you think that the numbers are as fast as possible to be understood (or even a tad above it), that's the right point. A good way to train it would be also by snapping your fingers, or with a metronom. Try to get the exactitude of the metronom in a 3-silable word like eleven, for example (stressing the middle silable).
Thank you for your work,

Joao

Dear Dunks,

thanks for the link. I've tried already something similar (that one's better, though). The thing is that is not only about getting "a" number sample, but they're for a very specific goal (as I just tried to explain in my just sent post).

Thanks,

Joao

jmmmp
I guess that what could make it better is for example to speak the numbers a bit faster (but keep the linear expression you already achieved).

Joao

OK Joao, just to clarify, do you want each individual number spoken faster regardless of gap between, or is the whole thing to be faster? I had intentionally left a bit of a gap between each, so you could cut them up as you liked. Will you be chopping up the sample, or using the full sequence as recorded?

Mike

acclivity
do you want each individual number spoken faster regardless of gap between, or is the whole thing to be faster? I had intentionally left a bit of a gap between each, so you could cut them up as you liked. Will you be chopping up the sample, or using the full sequence as recorded?

the way I'm programming the patch, all samples will be separated and loaded into different tables, so that they can be called upon only when necessary. so, it really doesn't matter how big the gap is, because I'll but them all up (I can do that, don't waste your time with it). I am not making a soundfile with all numbers, they'll just be there to be read by the "score".

I meant "faster" for each sample (11 was an extreme case). But with "faster" I just mean more exact, or precise. I found your voice tone/expression really appropriate, that should remain so. As I described in the last mail, the musicians are going to hear these samples while they play their instrument. What's necessary is just what's enough for them to understand in which beat they are. Also important is that the samples should indicate the precise rhythmical structure, so that's why they should more 'precise' as 'expressive'.

Unfortunately I can't myself make an example now (with my built-in micro), I am travelling tomorrow and packing as I type. If you want to wait I can give you a "live" example asap (so that you don't waste your time).

Joao

This might not be at all useful for what you have in mind, but I've had this sitting on my Mac for a few weeks waiting for some free time to come here and upload it. I saw this thread after returning from a short vacation and decided to finally get it into the database:

http://freesound.iua.upf.edu/samplesViewSingle.php?id=21916

The original intent (as per the description) was for IVR phone systems and things like that.

jppi_Stu
The original intent (as per the description) was for IVR phone systems and things like that.

actually I find it very good (I should ask a practising musician anyway). the tone of voice it maybe a bit too "active", but the precision is quite good, that's exactly what's necessary, and what can give a clear rhythmical feeling.

acclivity, if you still want to give it a shot, I think the metrical precision of this example toghether with your neutral expression would work out very well. (but don't let me abuse your good will)

regards,

Joao

acclivity, if you still want to give it a shot, I think the metrical precision of this example toghether with your neutral expression would work out very well. (but don't let me abuse your good will)

quote]

OK, I have had another go. "Numbers_EnglishMale.flac"

Edit 1: Also got my wife's assistance for "Numbers_EnglishFemale.flac"

If you need any more attempts, perhaps give me your email address, rather than loading lots of tries onto FS. Or even if you want it quickly, as there's sometimes a delay at weekends getting files approved.

Mike