Pollinators Secret Mixter Signups
skip
Home » Forums » DIY » Tenor, willing to sing most anything

Tenor, willing to sing most anything

celionati
.
permalink   Sun, Feb 3, 2008 @ 1:23 PM
I sing tenor, and I’m looking for satisfying projects.

I am trained in bel canto, willing to make weird noises, poking into throat singing.

I made a track on a whim yesterday, and someone was silly and/or kind enough to encourage me to post it. It can be found at http://66.92.135.14/Clients... The a cappella vocal tracks are licensed by me under CC attribution sharealike. The text is derived from the old “Taily-Po” folk story, as presented in Uncle Remus, which I believe is public domain now. I mixed with an instrumental by Angus Skrim, which they posted on their Web site http://angusskrim.com/ but without an explicit license.

If this track can be helpful, I’d like to provide it. But I’m not sure how to proceed. The 24-bit raw track seems to be the most valuable, but even with FLAC compression it’s more than double the 10 MB limit. I could split, trim silence, … ?

I’m also very willing to do other stuff as useful. This was a one-take, since I had no particular idea of what it should be. I can polish things up quite a bit. You can find somewhat better polish, but out of date, at http://people.cs.uchicago.e... The texts to “Nina” and “Upon Woman’s Love” are public domain, but “Countin’ Flowers on the Wall” is still under copyright. Performances are all CC attribution sharealike. I can dig out the short sample from which I looped the “Woman’s Love.”

I do my recording on an Edirol R09 at 44.1 KHz 24 bit samples. I have a cheap mono condenser mic, and a cheap dynamic, but so far I’ve preferred the built-in stereo.

So, if anyone cares to suggest how I can be most useful, please fire away. Sometime soon, I’ll pick some way to post the “Taily-Po,” or maybe the short sample from “Upon Woman’s Love,” as an experiment, anyway.

Mike O’Donnell
http://people.cs.uchicago.e...
DJ Rkod
.
permalink   Mon, Feb 4, 2008 @ 12:14 AM
Convert that sucker to 16-bit. 24 is fine but it’s debatable whether the human ear can perceive the difference and the loss of quality should be negligible if it’s there at all.

Trimming silence won’t help you as FLAC uses less space for less volume and at 0db it uses nothing.
 
.
permalink   celionati Mon, Feb 4, 2008 @ 3:32 PM
Thanks for the advice. Alas, I had 10,929,715 bytes of FLAC even at 16-bit monaural. So, I posted an OGG at quality level 10.

If someone would post, or point to (I searched and didn’t find) guidelines on the most useful forms of uploads, that will be very helpful.
 
.
permalink   celionati Mon, Feb 4, 2008 @ 3:44 PM
Quote: DJ RkodConvert that sucker to 16-bit. 24 is fine but it’s debatable whether the human ear can perceive the difference and the loss of quality should be negligible if it’s there at all.
Don’t want to start an argument—-just trying to figure things out. But this sounds fishy. All of the studio work that I know of (my knowledge is not extensive here) uses 24 bits, and often higher sample rates than 44.1 KHz, for all editing until a final cutback to 16 bits. The dynamic range in singing can easily blow away 24 bits, to say nothing of 16. But, I’m really wondering how much OGG compression sacrifices in detail that would be useful to a mixer.

If the answer is just that the mixing community doesn’t go for the highest fi, that don’t bother me at all, but I started out supposing it will be like studio production. And I get rather silly myself with samples, and enjoy finding breathing, heartbeats, all sorts of low-level stuff.

Quote: DJ RkodTrimming silence won’t help you as FLAC uses less space for less volume and at 0db it uses nothing.
I mean silence where the microphone is on, but nobody singing—-maybe I should call it “ambience.” I’ll run a test sometime to see how well FLAC reduces that. If the results are already known, a pointer will save me from repeating them.
 
.
permalink   fourstones Mon, Feb 4, 2008 @ 4:13 PM
In general when the target is the web (mp3s, etc) then 16bit vocals is a norm. For remixing the preference is absolutely zero effects like reverb, even compression to give the producer maximum control when fitting the vocals into the rest of the mix so losing the fidelity is less of an issue.

Zero’ing out the space between the vocals is a grand idea and should reduce the overall size of the pell file.

VS
 
.
permalink   DJ Rkod Tue, Feb 5, 2008 @ 3:38 AM
Quote: celionatiDon’t want to start an argument—-just trying to figure things out. But this sounds fishy. All of the studio work that I know of (my knowledge is not extensive here) uses 24 bits, and often higher sample rates than 44.1 KHz, for all editing until a final cutback to 16 bits.

True. Most professional studio work is done at 24-bit 96khz. But most of us are not working in a studio, and it’s still in doubt whether the human ear can even perceive the difference between 16 and 24 bit.

I’m an audio quality freak myself. I don’t mix MP3s or OGGs because I feel that they are ‘damaged,’ and unusable. But at some point extra audio quality is just not worth the trade-off in space.

Bear with me here. I took one of the mixes I did here, rendered it to 24-bit WAV, duplicated it in an audio editor, converted the duplicate to 16-bit and then computed the difference between the two by inverting one and mixing them together. There was a bit of static, but crucially, I had to amplify the file over 160 db before I could even hear it. 24-bit is essentially not much more quality for double the filesize.

Quote: celionatiThe dynamic range in singing can easily blow away 24 bits, to say nothing of 16. But, I’m really wondering how much OGG compression sacrifices in detail that would be useful to a mixer.

As I said in the last paragraph, I’m an audio quality freak and don’t mix with lossy files. (MP3, OGG, WMA) But I am the exception, not the rule. Most people here will happily use your files, though I would suggest putting up an MP3 because it is more widely supported.

Quote: celionatiI mean silence where the microphone is on, but nobody singing—-maybe I should call it “ambience.” I’ll run a test sometime to see how well FLAC reduces that. If the results are already known, a pointer will save me from repeating them.

Gates are your friend. A noise gate properly set up will take care of background noise automatically.

In conclusion I’d like to say that I think you should just go to the internet archive and put the full WAV tracks up there. You can link to them in the comments of your files here. So much easier that way.
spinmeister
.
permalink   Tue, Feb 5, 2008 @ 7:07 AM
16 bit vs 24 bit has very little to do with the end product since that tends to be in 16bit - even for CD’s.

Recording at 24 bit makes a lot of sense because it gives you a much more forgivenesses for recording at low volumes, because of the extra accuracy of the numbers.

So if your track is at a good volume, 16 bit is quite fine. However if your track is recorded at a low voulme, 16 bit can loose some accuracy.

During mixing 24 bit tends to give better results, too.

Both have to do with rounding. Using 24 bit samples is like having more decimal places compared to 16 bit. At high numbers more decimal places don’t really mean much. But when the numbers are small, more decimal places become more meaningful.

Look the effect of rounding at a low numbers at numbers between 10 and 20, a decimal place means a lot: e.g. 9.6 vs. 10.4 rounded are both 10, but they really are 8% apart. At high numbers between 1000 and 2000 an extra decimal place doesn’t mean much. 1499.6 and 1500.4 are much less than 0.1 % apart.

Mixing (including effects processing) is like doing math with several numbers . It is better to do the rounding AFTER dong the math, rather than before.

To illustrate with a simple example: compare the effects of rounding 1.1 plus 2.2 plus 3.3 by rounding before (1 plus 2 plus 3 = 6) vs. after (1.1 plus 2.2 plus 3.3 = 7).

Good volume tracks at 16 bit are fine. Low volume tracks at 16 bit can be troublesome to work with.
 
.
permalink   Surveillance_Party Mon, Apr 7, 2008 @ 5:56 AM
http://en.wikipedia.org/wik...

Dither is an intentionally applied form of noise, used to randomize quantization error, thereby preventing large-scale patterns such as contouring that are more objectionable than uncorrelated noise. Dither is routinely used in processing of both digital audio and digital video data, and is often one of the last stages of audio production to compact disc.

——-
It’s a very entertaining and simply written read. Dither applies to lots of things in life. I especially like the bit about whacking the aeroplanes to make them perform better!
 
.
permalink   Surveillance_Party Mon, Apr 7, 2008 @ 5:58 AM
Oh btw, don’t forget that if you use a 16 bit dithered vocal source file, and then mix down to 16 bit at the end, thats actually two layers of distortion you’re getting.
 
.
permalink   Surveillance_Party Mon, Apr 7, 2008 @ 5:59 AM
Which probably doesn’t matter, once you mangle it down to an MP3! LAWLS etc.

I’m a stop talking to myself now.