Taperssection.com
Gear / Technical Help => Ask The Tapers => Topic started by: svenkid on November 22, 2006, 03:01:37 AM
-
TS-
Ive started to get into with the MT recording at 24bit, but also running the jb3 to get a 16bit.
I only have 24bit files of certain recent stuff. My questions are:
1. what program do I need to use to be able to convert from 24>16
2. How do I do this and still be able to keep the 24bit files too.
Thanks TS!
-
<1> Freeware Audacity will do the trick.
<2> Simply edit / export the 16-bit WAVs, and keep a copy of the original 24-bit WAVs
This Audacity Workflow post may help: http://taperssection.com/index.php?topic=71191.0
-
I also like voxengo r8brain http://www.voxengo.com/product/r8brain/ (http://www.voxengo.com/product/r8brain/)
-
what software do you have Carl?
lots of topics here on proper workflow - you want to do any editing 1st then resample and then dither as the final step
-
what software do you have Carl?
lots of topics here on proper workflow - you want to do any editing 1st then resample and then dither as the final step
All I use is soundforge 4.5 for level boosting and fades. Im going to try to get my audacity up and running.
-
lots of topics here on proper workflow - you want to do any editing 1st then resample and then dither as the final step
True, plenty of operation order workflow threads. But each application has its own unique configuration and workflow. One must first understand how the software performs under each configuration. For example: does the software perform internal calculations at 16-bit? 24-bit? 32bfp? Is it selectable, or fixed? If operating in 32bfp for internal calculations, does it automatically dither to 24-bit upon saving the WAV? Or must one export the WAV to achieve dithering? Or must one dither first, and then export? And if one's performing multiple calculations over multiple sessions, does the user need to select 32bfp as the format for the saved, in-progress file, or will the application default to the proper bit-depth upon saving? Only once the user understands the software-specific configurations and workflow may one then configure the software and use the proper workflow to achieve the desired result.
Two reasons I bring all this up: <1> I did not find the Adobe Audition and Audacity software-specific configurations and workflows particularly intuitive, so I spent the time to understand them and document the info for others, and <2> I've seen plenty of examples of people not using their software correctly, and I have no doubt plenty of others have not spent the time to learn their software's configuration and workflow and therefore are not utilizing their software properly.
Hence, the Audacity workflow thread I pointed out. There's a quasi-workflow thread for Cool Edit Pro / Audition as well. I don't know of one posted for Wavelab or Soundforge or Samplitude (may get to the latter eventually).
-
very true Brian - i sent carl a PM about soundforge since thats what I use - one of these days i could probably put together a more detailed work flow specific for Soundforge for the Archives here - just not much time to get into it now ;D
-
sweet! I used wavelabio 5.01 and it did the trick!
-
There's a quasi-workflow thread for Cool Edit Pro / Audition as well.
Brian, would you post a link to the quasi-workflow for Cool Edit Pro / Audition? Thanks!
-
Brian, would you post a link to the quasi-workflow for Cool Edit Pro / Audition? Thanks!
Here's the most relevant post, courtesy of David Klein who sorted it out long before I did and provided invaluable feedback:
http://taperssection.com/index.php?topic=51323.msg667890#msg667890
There's still some useful information after that, but David's post covers the bulk of discussion.
-
What sample rate are you using? That's important b/c you need to resample to 44.1 first, then dither the 24-bit next, then truncate to 16-bit as your last step. SF is capable of doing this easily.
FWIW, I record in 24/44.1, do non-destructive editing in a 24-bit Vegas 6 project file, use the Waves L3 or IDR plugin as the last fx on my master bus with dithering set at 16-bit with noiseshaping engaged, and then just render out to a 16-bit file. Once you get it down in Vegas it's pretty easy (and I believe Vegas is running at 32-bit or higher floating point internally).
-
there should be no truncation required if the dithering was performed properly.
-
there should be no truncation required if the dithering was performed properly.
That's not true. Truncation is literally how you get from 24-bit to 16-bit, it's just that if you dither the 24-bit first, then when you truncate to 16-bit, you get better results. You have to truncate to go from 24-bits to 16-bits, you have to throw away those extra bits by definition, and that's called truncating -- it's just that you can choose to dither or not dither before you perform the truncation. Now some software will dither and truncate in one operation, so you might not realize that is what is going on, but rest assured it is.
-
there should be no truncation required if the dithering was performed properly.
Depends on the internal workflow of the application used for processing. In CEP/Audition, the user process is:
- dither from 32bfp to 16-bit - after dithering, the data is 16-bit, but it's still stored within a 32bfp file, the remaining 16-bits are just padded zeroes
- truncate to 16-bit - this removes the padded zeroes
- Save As 16-bit, this obviously saves the 16-bit data within a 16-bit file
Some applications perform the truncation in step 2 as part of the dithering or savings process and don't require the user to do it independently. However, CEP/Audition does require the user to do it manually.
-
Ha! We answered the same thing within a minute of each other. LOL!
-
I stand corrected... I use WaveLab which results in a 16bit file as a direct result of the dithering process.
-
Colour me n00b but back up a second and 'splain please Lucy...
Now I have the R09 I'm starting to do 24bit. I've only just started playing with these files and for the most part have been leaving them as 24bit but I want go to 16 I'd like to know the steps that i should be taking for optimum results... (I ignorantly assumed it would just be a case of "save as" and the software did it for me. However, we all know what they say when you "assume"... ;)
Anyway, cart before horse: can somebody explain (and/or point me to links explaining) what rendering and dithering are and (more importantly) how they affect final sound?
Ta ever so boys.
-
Anyway, cart before horse: can somebody explain (and/or point me to links explaining) what rendering and dithering are and (more importantly) how they affect final sound?
Rendering is just the language WaveLab uses to describe the process of applying edits to a file.
As for dither, first it's important to understand analog-to-digital conversion and vice versa (http://taperssection.com/index.php?topic=1167.0).
Dithering is a process by which one converts a file from 24-bit to 16-bit. There's a decent blurb on Wikipedia that explains audio dither pretty plainly (http://en.wikipedia.org/wiki/Dither#Digital_audio).
Different editing applications may also employ noise shaping as part of their dither funcions. Noise shaping is a way of shifting the (basically) random noise produce by dither to certain frequencies the listener finds more desirable, usually higher into the frequency range.
Bottom line is a 16-bit file properly dithered from 24-bit will sound better than a 16-bit file truncated from 24-bit. And different dither algorithms sound better to some people than others. I performed a dither comparison (http://taperssection.com/index.php?topic=51476.0) for my own purposes, and shared the results with others. You may not find it super useful, though, since if I recall correctly, divamum, you're often recording unamplified voice. IME, selecting different dither algorithms will have more of an effect on your recordings than someone recording from a PA. The reason is because your recordings will have more dynamic range. The (good) noise introduced by dither therefore will be more noticeable since your signal at times will drop to far softer than one produced recording a PA. This will make the dither you use, and the noise shaping options you select, more noticeable.
That said, probably none of the dither options in my comp will sound -bad-, but you may find you prefer one over another. I'd first start out getting familiar with the overall process of converting 24-bit files to 16-bit, and then start playing around with dither.
What software do you plan to use? While the general process steps are the same regardless of software - edit first (compression, gain changes, EQ, fading, etc.), then resample, then dither - different software uses different workflow to accomplish these actions. I've written a 24-bit workflow for Audacity (http://taperssection.com/index.php?topic=71191.0). I'm also familiar with Cool Edit Pro / Adobe Audition, and David Klein provided some really good informal feedback about its workflow (http://taperssection.com/index.php?topic=51323.msg667890#msg667890) when I first switched to 24-bit. If you let us know what s/w you're using, someone can probably provide info on proper workflow within the s/w.
-
MANY thanks Brian. Of course, this gets into the scary <gasp> mathematical side of things (I'm pretty geekly technologywise as long as i don't have to deal with NUMBERS - give me calcuations and I retire into a corner of "Me Dumb Singer" quivering jelly....) ;)
In any case, much to read and digest here for which THANKS!
I use Wavelab mostly although I also have Soundforge and for superquick down-and-dirty edits will sometimes use Nero Wav (just cause it's faster).
I'll read all those links you provided and once I've managed to grasp it will hopefully have a better idea what I'm doing (or at least what questions to ask).
You guys really are great in here - it's like a perpetually morphing taping encyclopedia :) Thanks again!
-
I use Wavelab mostly although I also have Soundforge and for superquick down-and-dirty edits will sometimes use Nero Wav (just cause it's faster).
I'll read all those links you provided and once I've managed to grasp it will hopefully have a better idea what I'm doing (or at least what questions to ask).
If you're not a math geek, don't sweat the math. It's the concepts that are important. And most important is knowing when and how to dither. Unfortunately, I'm only vaguely familiar with WaveLab, so I'm not much help for its specific workflow.