Audacity Forum

I'm doing a language documentation project, and it entails a lot of file editing, like batch processing 150 hours or recordings, getting them all to a good volume (some are quiet) and exporting copies as mp3s. I'm using 2.0.5. on windows 7. I then transcribe the files, import the labels into audacity, and export each transcribed sentence as a separate file using export multiple. Audacity has been giving me some problems, the biggest being the following.

The problems are this -First - when I go to export a lot of my files to mp3, it pauses after each and every single file an says that the lame encoder doesn't support the bitrate, would I like to change it**. *Of course I'd like to!* But what I'd really like is if I didn't have to click yes, literally five hundred times, each time separated by one to five minutes. Is there any way that I can say "yes to all future questions"? As it is, it's virtually un-usable when I have to be at the computer with audacity open, clicking yes every five minutes for a day. The second problem is almost the same. When exporting multiple files from single recordings, it forces me to confirm that I indeed want the file to be what the transcription is, and it forces me to do it every single time. I routinely have 300 transcriptions per file, and I have to sit and click ENTER ENTER ENTER ENTER ENTER x 300, 400 times. It needs to be possible to accept a default chain of actions, rather than have to re-affirm the same thing over and over again for hours. By the time I'm done this project I'll have 45,000 transcribed sentences, and maybe five hundred hours of .wav files, which will need batch processing. Either I figure out how to *actually* automate this work in audacity (which I know it can do) or I have to find a different program.

My only other problem is figuring out how to do the equivalent of auto-level with batch processing, but that I suspect is just a matter of figuring out the right combination of normalize and amplify, and doesn't require a new feature.

Chris's Compressor is designed to produce volume corrected works with wildly different volume originals. Try it manually on several very different clips. Last time I checked, you could include Chris in a Chain.

http://theaudacitytopodcast.com/chriss- ... -audacity/

Here's a sample of what happens to a short clip (first wave) with the default compression settings (first number, second wave) and the compression pushed up to 0.77 (third wave).

http://kozco.com/tech/audacity/pix/Chri ... Afters.png

~~

What MP3 compression value was Audacity objecting to? What was the Before value?

Forget Chains and Batches for a minute. Load a simple sound clip and export it as MP3 with 320 quality value.
Now do it again with a different filename and the MP3 quality value should be right where you left it. Is it? Is it still 320?

Delete those two test files.

When exporting multiple files from single recordings, it forces me to confirm that I indeed want the file to be what the transcription is, and it forces me to do it every single time.

That didn't come out English. Try again?

Koz

For the mp3 export, it says "the project sample rate (96000) and bitrate (128kbps) combination is not supported by the MP3 file format. You may resample to one of the rates below." the highest is 48000. I exported at 32000, then exported the same clip with another name, and the same problem comes up, it defaults back to what the original bitrate was, 96000. I tried setting the "project Rate" at the bottom to 32000 before starting the chain, but it didn't stick... It seems the the recorder I'm using is automatically set to that sample rate. If I could choose a default export bitrate for mp3s (like I do when I rip CDs using windows media player, for example), it would be a lot easier.

As to the part that didn't come through in English, I import labels to audacity, and then "export multiple" - select "using label/track name" . The "Edit Metadata" comes up, with the name of the track already chosen, but I have to click to approve the metadata of each and every exported file. What I need to be able to do is to approve all of the metadata from the "export multiple" window.

@ Koz - the maximum sample rate is 48000 Hz for MP3. If the project rate is greater, an "Invalid Sample Rate" dialogue appears where you have to choose the rate.

muskwatch wrote:For the mp3 export, it says "the project sample rate (96000) and bitrate (128kbps) combination is not supported by the MP3 file format. You may resample to one of the rates below." the highest is 48000. I exported at 32000, then exported the same clip with another name, and the same problem comes up, it defaults back to what the original bitrate was, 96000.

I think you mean the original sample rate.

muskwatch wrote: I tried setting the "project Rate" at the bottom to 32000 before starting the chain, but it didn't stick... It seems the the recorder I'm using is automatically set to that sample rate.

Each file acted on in the Chain is effectively a new empty project so if the files are 96000 Hz then the project rate will be set to 96000 Hz for each file.

muskwatch wrote:If I could choose a default export bitrate for mp3s (like I do when I rip CDs using windows media player, for example), it would be a lot easier.

I'll add your vote to assign export parameters for each Chain export command (where that format has export options).

muskwatch wrote:As to the part that didn't come through in English, I import labels to audacity, and then "export multiple" - select "using label/track name" . The "Edit Metadata" comes up, with the name of the track already chosen, but I have to click to approve the metadata of each and every exported file. What I need to be able to do is to approve all of the metadata from the "export multiple" window.

Again I'll add your vote, but meantime there is an unintuitive solution. Enter the metadata first using File > Edit Metadata... , then uncheck "Show Metadata Editor prior to export step" in the Import / Export Preferences. See the big green div near the top of http://manual.audacityteam.org/o/man/me ... ditor.html .


Gale