I need to record voice and deliver it in MP3.

The quality needs to be great, so my question is how many kbps is the minimum I should use? 128, 256, higher?? would 128 have noticably degradation from an uncompressed wav file for voice??

thanks

In my experience 128 is a bit low for MP3 but it kind of depends how you create it as well. My Marantz PMD-660's only have one MP3 setting (128) and it is good enough for transcription but not great audio. You haven't said how long the recording would be or how you plan on delivering it so size issues can't be determined. If you are recording a Wav file and then converting to an MP3 you can judge for yourself. 256 will probably do it but it is hard to be sure since the end use is not described. HTH

I never really like the way MP3s sound, so I get the quality as high as possible every time.

I need to record voice and deliver it in MP3.

The quality needs to be great, so my question is how many kbps is the minimum I should use? 128, 256, higher?? would 128 have noticably degradation from an uncompressed wav file for voice??

thanks

320K bps should be sufficient. Folks prefer not to go MP3, but if you must, always use the higher bit rates.

Some people, but not all, claim they notice audible artifiacting at 128Kbps, so going at a greater data rate is always a good idea.

Beside, if you record in PCM, you can then render MP3's at any rate your client desires, or you can deliver copies at different rates and let the client choose.

I need to record voice and deliver it in MP3.

The quality needs to be great, so my question is how many kbps is the minimum I should use? 128, 256, higher?? would 128 have noticably degradation from an uncompressed wav file for voice??

thanks

I'll go with Gene's comments. Do your original recording in wav format, bwf if your recorder is able to do it, and render to mp3 after the fact. That way you get the highest possible quality in the original and leave all your options open for the client to decide whether sound quality or file size is the number one priority.

Can't remember the source, but there were some serious studies done and the conclusion was that at the higher bit rates (ie 256 or 340) nobody could consistently tell the difference between the mp3 and the original CD (wav). Same was true for best quality VBR

Having said that, I'm not really crazy about CD sound. We finally got a CD player with a vacuum tube stage and for listening to classical music it's really nice.

I think the key to having voice sound great lies more in the recording technique/equipment than in the specifics of what mp3 bit rates are used (at least within reason - don't think I'd go below 128kb for anything).

I tend to believe you should start as high as possible. Record at 16 bit/44KHz WAV minimum starting and then edit and encode to MP3. For voice 128kbps should probably be fine. For church sermons I upload I normally use 128 because it's a decent compromise between quality and size for that application.

I need to record voice and deliver it in MP3.

The quality needs to be great, so my question is how many kbps is the minimum I should use? 128, 256, higher?? would 128 have noticably degradation from an uncompressed wav file for voice??

thanks

Actually, for voice, 128mpbs is good enough. If you were recording music or some other dynamic or wide range sounds, you should go higher or not use mp3 at all. If you need to go mp3, 225-256kpbs is just fine. It's hard to notice the difference with 256 and 320kbps anyway. As others have pointed out, you have other options.

I cannot 2nd guess your application, but many and typically, people "want the best." Problem is, other issues crop up like storage space for higher bitrates, or whatever. Not only that, you can have the best recorder, record at ac3 or some other format but if your microphone is lousy, then I'd rather record your voice project with an i-river mp3 player at 92kpbs with my rode videomic. I know I'll get better results with that even if you claim you have a U$1k recorder an at 96khz with a gazzillion bitrate but with a poor microphone. I just have to assume that your recording and mics are up to snuff as well.

And a $6k mic in the wrong place, or in the right place but poorly chosen for the application will make everything sound like crap.

I think in order of importance for "great sound" (voice or otherwise) it would be something like

1) Right type of mics in the right place (80%)

2) Right equipment (15%)

3) Everything else (5%)

Maybe a little tongue in cheek, but I think pretty close.

MP3 is fine if it is the end result with no further processing. If it is to be uncompressed and re-compressed in further stages of production then it is not suitable.

All MP3 is horrible, there is no high quality MP3 at all.

Record in uncompressed bwav format - do any editing and manipulation in that format and if you *have* to supply MP3 use the bast rate possible (320 I think).

Once it's in MP3 format you really cannot do anything else to it or it further degrades the recording.

I don't use MP3 at all - it's too horrible to listen to.

Jimmy - good point re re-processing mp3. IFRC I believe they compress L & R channels separately so losses in the process would not be symmetrical for stereo.

John - do you feel the same way about jpeg?

For better or worse the legacy of digital audio/video, whatever is that we live in a purposely degraded world.

On the other hand, analog reproduction (of sound/images, etc - not people!) is also subject to degradation. I guess if one thinks about it though degradation is the norm - birth being just the beginning of a long slow death.

I'd rather listen to an mp3 of a great performane/recording etc than to a perfect rendering of a crappy performance recorded with poorly placed and trashy equipment.

Sort of along the lines of a quote from Ansel Adams that used to be painted on a building on a Tokyo side street which I paraphrase as "There's nothing worse than a perfect image of a fuzzy concept".

John - do you feel the same way about jpeg?
Yes, actually. A Tif is much better - you can see how a jpeg muddies the detail.


For better or worse the legacy of digital audio/video, whatever is that we live in a purposely degraded world.
Maybe, but you don't have to like it.

I guess that compression will get less and less as the cost of storage drops - at least I hope so.

But what about HDTV? That is pushing quality higher.


I'd rather listen to an mp3 of a great performance/recording etc than to a perfect rendering of a crappy performance recorded with poorly placed and trashy equipment.
I would rather listen to an uncompressed great performance / recording than an MP3 - why spoil a great performance by putting it on MP3?

I would not listen to a crappy performance at all whatever it was on.



Sort of along the lines of a quote from Ansel Adams that used to be painted on a building on a Tokyo side street which I paraphrase as "There's nothing worse than a perfect image of a fuzzy concept".
Except, maybe, a fuzzy image of a perfect concept.

Yes, actually. A Tif is much better - you can see how a jpeg muddies the detail.



Maybe, but you don't have to like it.

I guess that compression will get less and less as the cost of storage drops - at least I hope so.

But what about HDTV? That is pushing quality higher.



I would rather listen to an uncompressed great performance / recording than an MP3 - why spoil a great performance by putting it on MP3?

I would not listen to a crappy performance at all whatever it was on.




Except, maybe, a fuzzy image of a perfect concept.

Compression is a funny thing when dealing with the human senses.As long as it's *imperceptible*, it's as if the compression is transparent.

Thing is - you can't get three people in a room to agree just on who's perception of imperceptible..??

OTOH - if you don't need compression - why use it?

Another point when recording sound for video, recording in MP3 exacerbates any sync drift issues when recording double system sound. MP3's compression does all sorts of weirdness to the sample clock. Shooting double system with a recorder that's capable of both WAV and MP3 might give reasonable shot lengths before drift causes loss of sync when set to WAV (PCM) file format but drift wildly when set to MP3. Ironically, the longer the shot length the more important lack of drift becomes and yet the only situation where it would ever make sense to record in MP3 would be where ultra-long single take recording times are important. Unless you absolutely need to record a single take that runs continuously for several hours I can't imagine any reason at all to ever use MP3 for the master recording, even for stand-alone audio without associated video - CF and other memory cards are cheap as dirt these days.

How much compression you'll "hear" from a 128 mp3 file also varies a lot. mp3 works on the idea that certain frequencies are less important than others, so the more it compresses the more it attacks those frequencies. If you stay within a certain range it may sound fine, but on higher frequencies for instance (a hi hat or even a voice with a naturally high pitch) you'll hear the compression a lot more.

The compressed mp3 file can even stand out more once you start mixing it with other sourced files (say 48k 24 bit uncompressed or better, which should pretty much be your only consideration for serious use). If you convert an mp3 back to wav it of course makes no difference - the compression (or "harm") has already been done