Login / Register
 
Converting 16-bit to 24-bit
New Reply
Subscribe
2012
Thread Starter
#1
7th March 2006
Old 7th March 2006
  #1
Gear addict
 
Joined: Mar 2005
Posts: 392

Thread Starter
2012 is offline
Lightbulb Converting 16-bit to 24-bit

hello

i have some older sessions in 16-bit 44.1 kHz.
i am going to re-track the vocals for the
sessions using newer equipment which
is capable of 24-bit.

i know the benefit of recording in 24-bit, but
does converting "upward" create any problems?

is there any benefit to converting the 16-bit
audio into 24-bit and work in a new
24-bit session (44.1) for recording vocals?

or should i leave it in 16-bit, and record
the vocals at 16-bit depth?

if so, what is the best way to do this?

1. save a session copy as 24-bit, and it will
convert the audio and create a new 24-bit session

2. use a 24-bit Dither plug-in on a master fader
and bounce the audio to disk for the new 24-bit session

thank you for your wisdom
#2
7th March 2006
Old 7th March 2006
  #2
Lives for gear
 
TheReal7's Avatar
 
Joined: Apr 2004
Location: In The Woods, Canada
Posts: 775

TheReal7 is offline
You'll definitely benefit from doing the vox in 24bit. I think you can start a new session and just do an import session data. Someone correct me if I am wrong.
__________________
Scott Kroeker
Melodic Guitarist
www.scottkroeker.com
www.myspace.com/scottkroekermusic


Quote:
He who asks a question looks foolish for 5 minutes. He who doesn't ask a question remains foolish forever.
#3
7th March 2006
Old 7th March 2006
  #3
Gear maniac
 
Joined: Apr 2004
Location: berkeley, ca
Posts: 251

kenm is offline
What program are you using? Nuendo will do this automatically when you import the files to a new session.

Ken
#4
7th March 2006
Old 7th March 2006
  #4
Audio Alchemist
 
Lagerfeldt's Avatar
 
Joined: Sep 2004
Location: Copenhagen, Denmark
Posts: 5,180

Lagerfeldt is offline
You will gain nothing by changing the 16 bit files to 24 bit files. The 16 bit files are merely padded.

However, you will improve quality by recording in 24 bits to begin with.

Sequencers such as Logic Pro and Cubase can combine files with different bit depths without any problems, it is not necessary to convert anything. Sample rates is a different matter. Only ProTools has this bit problem.
__________________


Online Mastering.dk

I have come here to chew bubblegum and kick ass... and I'm all out of bubblegum
2012
Thread Starter
#5
7th March 2006
Old 7th March 2006
  #5
Gear addict
 
Joined: Mar 2005
Posts: 392

Thread Starter
2012 is offline
Let me clarify the situation a little.

I have an old Fostex 8 Track Hard Disk recorder which records at 16-bit, 44.1 kHz. There are an album's worth of songs on there that are the stereo mixdown of drums/bass/guitar without any vocals. I want to import these songs into Pro Tools and record the vocals within Pro Tools.

So i'm using the S/PDIF OUT on the 8 Track to the S/PDIF IN on the Digi002.

1. Since the Fostex is 16-bit, should i set the Pro Tools session to record
the incoming digital audio at 16-bit or at 24-bit ?
If i choose 24-bit what does that do with the extra 8-bits?

2. If i have already recorded the tracks into 16-bit sessions, should i convert the files to 24-bit and then track the vocals in a 24-bit session?
Where do the extra 8 bits come from?

Should i perform #1 or #2 to see the best results?
Or will the resulting files be identical?

FYI - Since getting the Digi002 and Pro Tools, every project i've done has been in 24/44.1 format.

Thank you again.
2012
Thread Starter
#6
7th March 2006
Old 7th March 2006
  #6
Gear addict
 
Joined: Mar 2005
Posts: 392

Thread Starter
2012 is offline
Converting 16-bit UP to 24-bit


bump
#7
7th March 2006
Old 7th March 2006
  #7
Audio Alchemist
 
Lagerfeldt's Avatar
 
Joined: Sep 2004
Location: Copenhagen, Denmark
Posts: 5,180

Lagerfeldt is offline
Quote:
Originally Posted by 2012
1. Since the Fostex is 16-bit, should i set the Pro Tools session to record
the incoming digital audio at 16-bit or at 24-bit ?
If i choose 24-bit what does that do with the extra 8-bits?

2. If i have already recorded the tracks into 16-bit sessions, should i convert the files to 24-bit and then track the vocals in a 24-bit session?
Where do the extra 8 bits come from?
Like I said, the data just gets padded, no change to the audio, so do record in 24 bits if you plan to do more recording in the same session.
#8
7th March 2006
Old 7th March 2006
  #8
Gear nut
 
Joined: Jan 2006
Posts: 106

apdv1 is offline
Transfer your existing files at 16 bit into a ProTools session at 24 bit. Record your new vocals against your exsisting material, then dither the new vocals down to 16, open a new session at 16 bit and transfer all your new files into that using spot mode to line them all up.
#9
7th March 2006
Old 7th March 2006
  #9
Lives for gear
 
joaquin's Avatar
 
Joined: Sep 2004
Location: CHILE-Miami
Posts: 1,204

joaquin is offline
what a mess!
just hook the spdif OUT of your fostex into the spdif IN in PT. create a 24bit session and record the Spidif input. no problema. if there was a different sample rate...that's another story, or...if you were coming from 24 to 16!?
Like Lagerfeldt mention there's no improvement over the already recorded audio, but the vox you are to record will definately benefit from the extra resolution.
If I'm mistaken...please correct me, but I'd done transfers from ADAT at 20 bit into PT at 24bit thru the lightpipe, with no issues at all!
Good luck!
__________________
My studio, is like a womb...It's so freaking comfortable!!
Cheers......................Joaquin
.
http://youtu.be/BofOKFMfyW0
www.kzkrecords.com
2012
Thread Starter
#10
7th March 2006
Old 7th March 2006
  #10
Gear addict
 
Joined: Mar 2005
Posts: 392

Thread Starter
2012 is offline
THANK YOU all, especially Lagerfeldt and joaquin

#11
8th March 2006
Old 8th March 2006
  #11
Gear Guru
 
theblue1's Avatar
 
Joined: Mar 2005
Location: Long Beach, CA
Posts: 21,154

theblue1 is offline
Quote:
Originally Posted by Lagerfeldt
You will gain nothing by changing the 16 bit files to 24 bit files. The 16 bit files are merely padded.

However, you will improve quality by recording in 24 bits to begin with.

Sequencers such as Logic Pro and Cubase can combine files with different bit depths without any problems, it is not necessary to convert anything. Sample rates is a different matter. Only ProTools has this bit problem.
I'm sure Laderfeldt didn't mean to suggest that there aren't benefits to working with your old files in a new 24 bit project-- if you're doing further processing, mixing, etc, to the tracks.


What he's getting at, I'm sure, is just that there's no immediate advantage to converting the 16 bit files to 24. You're tacking on extra 'zeros' behind the decimal so you'll have the potential to capture more of the detail generated by further processing.

But if all you do to a 16 bit file is convert it to 24 bits, it should sound exactly the same.


[Where's that beating-dead-horse icon I ought to be using in good-natured self-criticism here? ]
2012
Thread Starter
#12
8th March 2006
Old 8th March 2006
  #12
Gear addict
 
Joined: Mar 2005
Posts: 392

Thread Starter
2012 is offline
16-bit to 24-bit

thank you theblue1 for clarifying a little bit further!
#13
8th March 2006
Old 8th March 2006
  #13
Audio Alchemist
 
Lagerfeldt's Avatar
 
Joined: Sep 2004
Location: Copenhagen, Denmark
Posts: 5,180

Lagerfeldt is offline
Quote:
Originally Posted by theblue1
I'm sure Lagerfeldt didn't mean to suggest that there aren't benefits to working with your old files in a new 24 bit project-- if you're doing further processing, mixing, etc, to the tracks.
Actually the only benefit is to the newly recorded files. Let me try to explain why.

Whether you are using 16 bit files in a session or 16 bit files recorded to 24 bit files (padded) in a session will make no difference to the actual outcome on those files when processed in realtime.

The reason is that once the file is processed in any way (even a gain change of 0.1dB) the DAW always expands to its working bit depth (such as 32 bit float in LE or 48 bit fixed in TDM systems), so the 16 bit file is expanded to the maximum if it is processed, and the 24 bit file (padded) is treated the exact same way, with no advantages or disadvantages for that matter. The result will be identical (try doing a null test where you flip the phase 180 degrees). You can also use a bitstream viewer to confirm how the bit depth expands, it is very interesting stuff.

So there is no limitation on the actual processing based on the original file bit depth, it will always expand to the DAW's bit depth.

The benefit is only to files actually recorded originally in 24 bits.

Hope that explained it. :-)
2012
Thread Starter
#14
9th March 2006
Old 9th March 2006
  #14
Gear addict
 
Joined: Mar 2005
Posts: 392

Thread Starter
2012 is offline
Converting 16-bit to 24-bit

i was told, and so it is my understanding that once the 16-bit files
are in a 24-bit session, and you apply processing, such as EQ or
Comp plug-ins, that then the entire 24-bit depth is used...

is this correct?
#15
9th March 2006
Old 9th March 2006
  #15
Lives for gear
 
joaquin's Avatar
 
Joined: Sep 2004
Location: CHILE-Miami
Posts: 1,204

joaquin is offline
Quote:
Originally Posted by 2012
i was told, and so it is my understanding that once the 16-bit files
are in a 24-bit session, and you apply processing, such as EQ or
Comp plug-ins, that then the entire 24-bit depth is used...

is this correct?
...Yes.
#16
9th March 2006
Old 9th March 2006
  #16
Gear maniac
 
Kalite Marka's Avatar
 
Joined: Apr 2004
Posts: 217

Kalite Marka is offline
My suggestion would be to run most of these tracks through your favorite mic pres or through analog compressors (especially the drum tracks and bass) at really low ratios and record all back on PT 24-bit.
Have done this with 2 albums that got re-released… believe me it does wonders.
2012
Thread Starter
#17
9th March 2006
Old 9th March 2006
  #17
Gear addict
 
Joined: Mar 2005
Posts: 392

Thread Starter
2012 is offline
Converting 16-bit to 24-bit

good idea kalite, i will try that.

thank you everybody!
#18
9th March 2006
Old 9th March 2006
  #18
Gear Guru
 
theblue1's Avatar
 
Joined: Mar 2005
Location: Long Beach, CA
Posts: 21,154

theblue1 is offline
Quote:
Originally Posted by Lagerfeldt
Actually the only benefit is to the newly recorded files. Let me try to explain why.

Whether you are using 16 bit files in a session or 16 bit files recorded to 24 bit files (padded) in a session will make no difference to the actual outcome on those files when processed in realtime.

The reason is that once the file is processed in any way (even a gain change of 0.1dB) the DAW always expands to its working bit depth (such as 32 bit float in LE or 48 bit fixed in TDM systems), so the 16 bit file is expanded to the maximum if it is processed, and the 24 bit file (padded) is treated the exact same way, with no advantages or disadvantages for that matter. The result will be identical (try doing a null test where you flip the phase 180 degrees). You can also use a bitstream viewer to confirm how the bit depth expands, it is very interesting stuff.

So there is no limitation on the actual processing based on the original file bit depth, it will always expand to the DAW's bit depth.

The benefit is only to files actually recorded originally in 24 bits.

Hope that explained it. :-)

I think we're pretty much in agreement here. I think there was more a matter of wording.

If I understand what you're saying, you're saying that the DAW will be operating at its own internal resolution no matter what the input and output resolutions will be. Which is probably the case with most if not all DAWs.

So, if I get your reasoning, the 16 bit file goes into the DAW, is padded out to 24 bits or 32, 48, 64 or whatever the internal resolution of the DAW is and then output to whatever resolution the output format is. So 'converting' the 16 bit file to a higher resolution is simply part of the process anyway.

And that makes plenty of sense.
#19
9th March 2006
Old 9th March 2006
  #19
Gear Guru
 
theblue1's Avatar
 
Joined: Mar 2005
Location: Long Beach, CA
Posts: 21,154

theblue1 is offline
Quote:
Originally Posted by Kalite Marka
My suggestion would be to run most of these tracks through your favorite mic pres or through analog compressors (especially the drum tracks and bass) at really low ratios and record all back on PT 24-bit.
Have done this with 2 albums that got re-released… believe me it does wonders.

That may get you a sound you like better -- or it might not. But it is going to add a new 'layer' of A/D and D/A as well as whatever artifacts, positive or negative, that are induced in your outboard processing. (Sounds a lot like what ME's who use an analog chain do, of course.)


But the advantage I think he's looking for is in working the recorded tracks at 24 bit depth, which is kind of an associated but peripheral issue.
2012
Thread Starter
#20
9th March 2006
Old 9th March 2006
  #20
Gear addict
 
Joined: Mar 2005
Posts: 392

Thread Starter
2012 is offline
Converting 16-bit to 24-bit

initially, i was just curious whether i should transfer the sessions to a 24-bit session to do vocals, or if it would be better to leave them in 16-bit for doing the vocals. i was afraid that converting 16-bit to 24-bit would adversely affect the audio. Since converting upward doesn't really do anything to change the signal at all, it makes sense that doing the vox in 24-bit would be the best route to take.

what i did do was create a 24-bit session and then recorded the 16-bit data from the S/PDIF RCA in on the Digi002, with the clock set to S/PDIF, onto a stereo track.

worked without a hitch.
#21
9th March 2006
Old 9th March 2006
  #21
Lives for gear
 
joaquin's Avatar
 
Joined: Sep 2004
Location: CHILE-Miami
Posts: 1,204

joaquin is offline
Quote:
Originally Posted by Lagerfeldt
...So there is no limitation on the actual processing based on the original file bit depth, it will always expand to the DAW's bit depth.

The benefit is only to files actually recorded originally in 24 bits.

Hope that explained it. :-)
Thanks man, that was an enlightening read thumbsup
so actually all files wil be equal in the session and end up with a bunch of zeros at the bottom, only, the 16bit files will have eight more of them...
There, must be a benefit at some point in the process for the 16bit files to be read as 24...!?
#22
9th March 2006
Old 9th March 2006
  #22
Gear Guru
 
theblue1's Avatar
 
Joined: Mar 2005
Location: Long Beach, CA
Posts: 21,154

theblue1 is offline
As I understand it, a 20 bit converter is probably sufficient for capturing most conventional analog signals, at least if we use the Nyquist Theorem as a guideline. But it apparently was just about as easy to make a 24 bit converter as a 20 bit converter at the time when the big switchover began in the late 90s. No one wants to get left behind in a numbers war, so...


Hmmm. I just reread your last post. I'm not sure I actually addressed what you were getting at. I think I"m lost.

Is this the AOL single's club?

#23
9th March 2006
Old 9th March 2006
  #23
Audio Alchemist
 
Lagerfeldt's Avatar
 
Joined: Sep 2004
Location: Copenhagen, Denmark
Posts: 5,180

Lagerfeldt is offline
Quote:
Originally Posted by joaquin
Thanks man, that was an enlightening read thumbsup
so actually all files wil be equal in the session and end up with a bunch of zeros at the bottom, only, the 16bit files will have eight more of them...
There, must be a benefit at some point in the process for the 16bit files to be read as 24...!?
No, there is no benefit in a realtime situation. You can't magically expand on the information present, and since all signals are treated (expanded) equally, there is no advantage.

However, the reason why I mention realtime as opposed to destructive (offline), is that there has been some discussion as to whether some DAWs treat offline processing differently in terms of bits. And in those situations a 16 bit file padded to 24 bits might actually have an advantage.

But be clear about how the process works, adding any kind of sound changing conversion, dithering or other process while going from 16 to 24 bits will most likely be detrimental to the sound quality, and not outweight the benefits later on IMO.

But in this case I'm guessing offline processing is not relevant, so I didn't originally elaborate on the issue.
#24
9th March 2006
Old 9th March 2006
  #24
Gear Guru
 
theblue1's Avatar
 
Joined: Mar 2005
Location: Long Beach, CA
Posts: 21,154

theblue1 is offline
Quote:
Originally Posted by joaquin
Nyquist theorem is related to frequency not dynamics!
I believe you're right, actually. I did some reading after my post. (Oh? You're supposed to find out what you're talking about before posting? Now you tell me.)

As I get it now from my admittedly rushed re-scan of familiar yet faded material, the Sampling Theorem (aka Nyquist, aka Nyquist–Shannon, aka Whittaker–Shannon Theorem) is apparently restricted to talking about sampling frequency. Dynamic resolution issues are addressed elsewhere.

My bad.

I really should have known better.
#25
9th March 2006
Old 9th March 2006
  #25
Lives for gear
 
octatonic's Avatar
 
Joined: Jun 2004
Location: An Australian in London
Posts: 5,693

octatonic is offline
Quote:
Originally Posted by Lagerfeldt
You will gain nothing by changing the 16 bit files to 24 bit files. The 16 bit files are merely padded.

However, you will improve quality by recording in 24 bits to begin with.

Sequencers such as Logic Pro and Cubase can combine files with different bit depths without any problems, it is not necessary to convert anything. Sample rates is a different matter. Only ProTools has this bit problem.
I use Logic TDM (Digideisgn HD hardware) and I have had to convert all the sample CD's to 24bit in order to be used.
It took weeks to do even with batch converting (500GB).
__________________
"I don't go to mythical places with strange men." Douglas Adams
#26
9th March 2006
Old 9th March 2006
  #26
Lives for gear
 
joaquin's Avatar
 
Joined: Sep 2004
Location: CHILE-Miami
Posts: 1,204

joaquin is offline
Quote:
Originally Posted by Lagerfeldt
However, the reason why I mention realtime as opposed to destructive (offline), is that there has been some discussion as to whether some DAWs treat offline processing differently in terms of bits. And in those situations a 16 bit file padded to 24 bits might actually have an advantage.
Interesting...could you tell us what's the difference...is it, that, since is non-realtime processing with out posible surprises, there's no need for the extra resolution and math goes straight to least amount of posible numbers...I'm just guessing

Quote:
Originally Posted by richmondjames
I use Logic TDM (Digideisgn HD hardware) and I have had to convert all the sample CD's to 24bit in order to be used.
It took weeks to do even with batch converting (500GB).
that's probably different, cause your making the transformation within the same format in a NON-Realtime process, avoiding converters or cables...!?
Did you have to UP-Sample (rate) the sound samples as well?? I wondered my self about this incredible orchestral libraries that are in 44.1KHz...
Thanks.
New Reply Submit Thread to Facebook Facebook  Submit Thread to Twitter Twitter  Submit Thread to LinkedIn LinkedIn  Submit Thread to Google+ Google+ 
 
Thread Tools
Search this Thread
Search this Thread:

Advanced Search

Forum Jump

SEO by vBSEO ©2011, Crawlability, Inc.