Hi,
I am looking for a little help to get started using Audacity. My main concern is having recorded a sound track using MP3 files and then needing to modify these tracks, is there a progressive degrading of the file associted with re-recoding/encoding? In other words if I creat a mixdown file in MP3 and then want to modify it and create another soundtrack, does this file degrade each time this is done?
Hope that makes sense.
TC
Short answer - yes.
Every time you re-encode with MP3 compression you will be throwing away more audio data. Ideally you should be doing all your production in WAV or Audacity projects which are lossless (including original capture) - and only export to a compressed file for delivery.
You are of course stuck if the only available copy of your source material is in an already compressed format like MP3. There are some very simple MP3 editors around that let you do stuf like cut bits out - but they are very basic and don’t do all the Audacity stuff.
WC
Hi WC,
Many thanks for the reply you appear to have confirmed my suspicions. Just another thought, if you convert an .mp3 file to a .wav file and use that I suspect taht apart from the original loss the only other loss would be on the final conversion to an .mp3 mixdown file. Any thoughts?
All the best
TC
Yes, that’s my understanding too. You get compression damage when the original MP3 was created - converting to a WAV or Audacity project will retain that original compression damage.
Working in WAV or Audacity project files will add no further compressin damgage (as they are lossless).
Exporting to MP3 for final delivery will then add a second set of compression damage.
WC
Hi WC,
Thanks for the help. Two minds are better than one. I will now have to get down to it and produce some soundtracks.
All the best
TC
“Almost” lossless.
I’m being picky here, and since you are talking about the end format being MP3, the losses that may be caused are going to be so small as to be irrelevant, however for the highest quality processing, Audacity uses 32bit uncompressed audio by default. If you export as 16 bit WAV files, then the lowering of the bit depth will result in a small loss of quality. Sticking to 16 bit formats throughout does not get round the problem as many processes within Audacity are done at 32 bit, and if the file being processed is a 16 bit file, then the processed signal will be dithered.
For “lossless” processing, it is necessary to use 32bit audio throughout, then dither will only be applied during the final output conversion (if exported at a lower bit depth).
For “lossless” processing, it is necessary to use 32bit audio throughout, then dither will only be applied during the final output conversion (if exported at a lower bit depth).stevethefiddle
Hi Steve,
Thanks for the info. As you may gather I am trying to get the best possible quality at final mixdown stage to mp3 which will probably be at 320 kbps. However, from your comments 32 bit is the way to go. Just so I am sure about this, if you rip a cd track using Windows Media Player as a wav file is this 32 bit or 16bit? Also I use an Edirol R09 for voice-overs, should I be recording at maximum 24 bit wav in order to maintain best quality. My understanding was that many sound editing programs could not handle even 24 bit?
As Audacity is set as default to export at 16bit wav does it make any diference if the Default Sample Format is set to match at 16bit or are the losses so small as to be undetectable by the human ear?
Sorry won’t be able to get back to any reply until early next week.
All the best
TC
CDs are 16bit, so when you rip them you will get a 16 bit file.
Audacity can “up-sample” the 16 bit file to 32 bit (up-sampling is lossless).
The (small) gains that you make are during processing. When processing 32bit audio, there are virtually no “rounding off” errors.
The final Export will introduce a small (minute) amount of dither (noise) to smooth out the rounding off errors caused by down sampling back to 16bit.
16bit audio is very much more widely supported than 24 or 32 bit audio, so unless your files are for personal use only you should make your final Export as 16bit. (I think that Lame does the conversion automatically, but I don’t use MP3 very much so I’ve not checked that out.
There has been some discussion about this recently. Apparently quite a lot of sound cards that claim to support 24 bit audio are shipped with Windows drivers that only support 16 bit. However, the difference in sound quality, while arguably better at 24 bit, is unlikely to be very noticeably different from 16 bit. The difference between 16 bit and 32 bit while processing is probably a lot more significant, and unless you are doing a lot of processing even that is unlikely to be very noticeable.
Full scale dynamic range for 16 bit audio is 96dB - that is enough to hear the cat next door sneeze. It’s unlikely that your recording room is that quiet, so the noise floor due to 16 bit recording is likely to get buried beneath all the other noise in the environment (and from the analogue parts of the equipment - microphones frequently have a dynamic range well below 70dB, so one could argue that many of the extra bits in 24 bit recording are largely wasted).
Hi STeve,
Thanks for the info. I guess to summarize I need to record voice-overs at 16 bit and as CD’s are 16 bit I can work on these files and save Project files in .wav format until final mixdown to .mp3.
Many thanks for yuor help
All the best
TC