Sound Quality
I need a good quality recording from a 45min song file. I record our church sermons with N-Track then I mix them down to a file size under 20megs. The reason I need to mix them down to a small size is because. The web site that hosts the sermons requires the file size to be under 20megs. The only way I can think to drop the file size down is to change the sampling rate. But when that happens the quality of the sermon stinks. Please help.
http://www.sermoncloud.com/glen-bu…-powers
www.gbepc.org
thx
Scott
My Webpage
Are you posting them as mp3 or .wav or 
D
I would crunch them to mp3 but even then, I don’t know if 45 min of audio is going to go below 20 megs. Mine are just about 1min = 1meg at 128/14400
cliff
Chop them into parts I and II. Or I and II and III.
Also, make them mono - no need for stereo for a sermon.
I am posting them as mp3. That is what is required my the hosting site. What if I do the master recording as a lower sampling rate?
For just voice, you might try recording at 16 bits at 22,050hz. The voice quality should still be OK. The conversion to mp3 shouldn’t kill it either. Worth a try…
D
EDIT: According to a couple of sources, normal voice range is about 500 Hz to 2 kHz. Thus, according to Nyquist’s theorem, 22,050 sampling rate has you covered up to ~11,000 hz. Should work great for just voice.
open Windows explorer - locate the folder where you store Ns files - open folder - click on last box (righthand side) of explorer toolbar - click on details - now you can see the size of your tracks -
Dr J
A couple of facts to consider:
• Mono and stereo .mp3s of the same length are the same size. Converting to mono will not save you any space, although spoken voice may actually be clearer in mono at very high compression rates due to reduced stereo “swishing” (sure, it’s a technical term).
• Mp3s of the same length from files of different sampling rates are the same size. So recording your source file at a lower bitrate will not save you anything. Additionally, Bladeenc, which n-Track uses to convert to .mp3 can only convert 32k, 44.1k and 48k .wav files to .mp3 (I don’t know about other codecs). So recording at 22.05k would only save you space temporarily as you’d have to resample up before compressing.
• The rule of thumb is always to record your source at the highest quality level you reasonably can and then compress down from there.
Is that first thing really true, Captain? Mono and stereo mp3 files are the same size?
Newer MP3 encoders allow for both variable bit rates and sampling frequency in the encoded MP3 file. As Captain D points out, the original sampling rate of the wave file does not affect the size of the MP3 file, and therefore the wave file should be recorded at the highest feasible sampling and bit rates. During encoding (i.e. converting from wave to MP3), however, you will have wide lattitude in setting both bit rates and sampling frequencies. Reducing either or both of these will reduce the file size. The LAME encoder can also use a variable bit rate (higher for complex parts lower for simpler parts) but not all MP3 players will work with variable bit rates.
Because you are recording spoken word only (which has limited dynamics) the bitrate can probably be set fairly low (maybe down to 192 kbits/sec). As D says, you can probably set sampling frequency down to 11.025 khz. But you should experiment a bit to verify the quality of the encoded file.
T
Ahhh, 192kb/sec is not low. 128 is CD standard. So for spoken word, 90 should be OK, and will significantly reduce the mp3 filesize. But I wouldn’t drop the sampling freq lower than 22k. (my 2 cents worth).
Record your original wave at a decent quality (at least 16bit 44k). Edit and devide the final wave into about 2 equal parts. Then encode each half at 64kbs/22,050hz using LAME. That should get you very close to a 20meg file for each half. I just tried this with a recent a’cappella recording I did and the final result was very listenable (not audiophile quality, but certainly easily passable).
Good luck,
Mike
| Quote (vanclan @ Nov. 28 2006,12:44) |
| Ahhh, 192kb/sec is not low. 128 is CD standard. So for spoken word, 90 should be OK, and will significantly reduce the mp3 filesize. But I wouldn't drop the sampling freq lower than 22k. (my 2 cents worth). |
Oops. Vanclan is absolutely right about bitrate. 192kb/sec is incorrect... and 90 kb/sec is probably good choice. And do stick to the higher sampling rate if splitting the file is an option. If it is not, you will have to decrease sampling rate below 22k.
| Quote (TomS @ Nov. 28 2006,10:08) |
| Is that first thing really true, Captain? Mono and stereo mp3 files are the same size? |
Yes. It’s easy enough to test for yourself. The issue came up a few months ago when I started doing music and sound effects for the BostonGeek podcast. The guys doing the 'cast had been rendering it in mono, assuming they were saving space and bandwidth. However, this ended up damaging the full glory of the wonderful music I was writing for them so I decided to find out just how much they were saving - and it turned out to be zero. So now they render in stereo.