Re: [Audacity-devel] HowTo: convert sample rate?
A free multi-track audio editor and recorder
Brought to you by:
aosiniao
From: Lynn A. <l_d...@ad...> - 2005-02-09 15:29:18
|
> The "project rate" is controlled by the pop-up menu in the lower-left > corner of the window. That is the rate that is used for playback, > recording, and exporting. Your tracks can be at a different sample > rate - changing the track's sample rate changes how that track is > interpreted, but leaves the project sample rate alone. THANKS! I completely overlooked that ... I didn't really notice the widget in the status area .. and then thought it was read-only. Does this cause the mDirty flag to be set (WaveTrack and/or WaveClip and/or project) and PushState to happen? Should it? I found and stepped through the AStatus::OnRate11 handler. I was expecting that this would trigger a "Save" flag at exit, but I was allowed to exit without the "Save" dialog. Also curious ... why is the 16000 option greyed out? 8000 is available. > To convert to mono, use the track's pop-up menu to split the stereo > track. Then either set both tracks to mono, or else delete one. I'm unclear ... let's say I have a stereo track with rather extreme stereo separation. The person on the left counts to ten into the left microphone. The person on the right counts from eleven to twenty into the right microphone. I want to merge these into a mono track (counting from one to twenty). Can I do this directly? Indirectly? Does this involved "cut/paste"? > This is definitely confusing, and I'd be happy if somebody wanted to > add a "Change sample rate..." option somewhere. But the UI is harder > than it looks...what would you do if given a project with multiple > mixed sample rates? The current system, while a little harder to grasp > at first, is quite powerful and flexible, and consistent with itself > once you understand it. My naive impression is that "CleanSpeech" can do this rather easily ... since it pretty much assumes mono speech. "Accept no substitutes" <g>. I suppose a menu item could be implemented, but stay greyed out (disabled) if the tracks had different sampling rates. My naive impression is that the majority of recordings would have the same sampling rate, so Audacity might consider allowing this ... a relatively smart menu item would trigger a "blunt instrument approach". Am I missing something? Has this been discussed at length before? I can try to implement this if it would help Audacity. It's kind of a "non issue" for "CleanSpeech". <alert comment="aside"> I am regularly struck by the differences between my use of Audacity and a real audio engineer. It just never occurred to me that different tracks would have different sample rates. My orientation is to facilitate end-users loading a mono track, trimming some material from the start and finish, hitting the "Prepare" button, and being done with it. One down, twenty to go. They are typically busy volunteers who aren't necessarily all that computer literate, and working with older, donated computers. They may not know bit-rate from back-rub. The original sound source may be be practically retched ... recorded with donated hobby microphones and donated cassette decks. The speaker is alternatively shouting or whispering to put a dramatic flair on his/her material, so the actual s/n can be pretty minimal in places. The hum/hiss/noise "baseline" can be at -25db, and that's not worst case, just below average audio quality for some church recordings. Noise at -30db is fairly typical. Many churches seem to use under-recorded settings so the "red light" never comes on, so the typical "local peak" signal is only -10db or less. This might be the extent of their "audio engineer" training: "When the next-to-last song is getting about done, hit the "Record" button. Try to pay enough attention to turn the tape over if the sermon goes longer than 45 minutes. When the closing song starts, hit the "Stop" button. Don't touch the recording level knob." </alert> > > Related question ... when I've saving a .mp3 file, how do I specify I > > want 16 kbps rather than 20, 24, 32, etc.? > > In the preferences, under "File Formats". > > By the way, is it worth saving an MP3 as 11 KHz AND with a small > bitrate? In other words, does 11 KHz, 32 kbps actually save any space > over 44 KHz, 32 kbps, or sound any better? I would think they'd be the > same size, so you may as well stick with the better sample rate if > you're saving as an MP3... <alert comment="newbie non-audio engineer"> I'm unclear on this ... IANAAE ... I am not an audio engineer ... not even close ... only barely aware of what rms means (actually fuzzy) For speech, 16kbps is plenty and keeps the file size down (I think ... may be confused here). My impression is that 11025 sampling rate (sr) is preferred over 16000 sr or 22050 sr, but I don't know why for sure ... and I may be mistaken. It seems like the file sizes end up the same, however, when I've checked. When you specify 16kbps to the mp3 encoder, are you indicating ... "do the best you can, but it has to fit into a file that is only 2KB per second" ... 10 seconds has to fit in 20KB. When you specify 11025, is this indicating that frequencies over (11025 / 2) = 5512.5hz can be ignored? Does specification of 16000 indicate that frequencies up to 8000hz need to be factored in ... but there probably isn't much signal "up there"? Seems like the job of the mp3 encoder is enhanced and it can do a better job with speech if it knows to ignore high frequencies, but I am very ignorant in this respect. I suppose I should investigate on my own and not embarass myself by my ignorance. </alert> |