Topic: Normalizing, Dithering, SDB/AUD mix help

[therodge]

Hey Guys,

I'm going to use Soundforge 7.0 to normalize and change the sample rate of a NBJ3 recording I did recently...

1) when going from 48khz > 44.1khz, should I dither the recording and which dither setting is best?

2) when normalizing, should I normalize the volume to the -6dB mark or can I should I go all the way up to just a hair under the clipping mark?  I read in the help file that normalizing over -6dB can start damaging the integrety of the sound. What about using the "loudness" setting?

3) When making a sdb/aud mix, is there a rule of thumb to go by?  example: 2 parts SDB and 1 part AUD.  Any tips or preferences from your expierence on mixing sdb and aud? 

Thanks in advance! 

[John Kelly]

Hey Guys,

I'm going to use Soundforge 7.0 to normalize and change the bit rate of a NBJ3 recording I did recently...

1) when going from 48khz > 44.1khz, should I dither the recording and which dither setting is best?

2) when normalizing, should I normalize the volume to the -6dB mark or can I should I go all the way up to just a hair under the clipping mark?  I read in the help file that normalizing over -6dB can start damaging the integrety of the sound. What about using the "loudness" setting?

3) When making a sdb/aud mix, is there a rule of thumb to go by?  example: 2 parts SDB and 1 part AUD.  Any tips or preferences from your expierence on mixing sdb and aud? 

Thanks in advance! 

1)  You have to dither, and I would always dither at the highest quality setting.  I'm unfamiliar with Soundforge so I can't really get specific.

2)  The best rule of thumb is to run your gear as hot as you can at the show.  Then you wont need to do many adjustments when you get home.  The very few times I've normalized I've always done it to 0db.

3)  It really depends on your ear.  I like the 60 SBD 40 AUD.  Other people like different settings...  Just try out a few and see what you like the best.

[dklein]

Sample rate conversion - highest quality settings yes, I'd be tempted to do it in 32 bit floating point format for your multitrack mix and dither the whole thing down after it's mixed.

Normalizing - If you are going to mix the sources, don't normalize anything before hand.  It's just unnecessary, destructive processing.  Instead, do whatever you need to do when you're in the multitrack environment.

SBD/AUD mix - whatever your ear likes.  It also depends on the sonic characteristics of each source.  Be forewarned...because your sources were recorded on 2 different recorders, with the timing variations that always exist in these devices you're about to get into synchronization issues between the 2 sources.  Definitely one of the more challenging tasks to do well.

[pfife]

Hey Guys,

I'm going to use Soundforge 7.0 to normalize and change the bit rate of a NBJ3 recording I did recently...

1) when going from 48khz > 44.1khz, should I dither the recording and which dither setting is best?

2) when normalizing, should I normalize the volume to the -6dB mark or can I should I go all the way up to just a hair under the clipping mark?  I read in the help file that normalizing over -6dB can start damaging the integrety of the sound. What about using the "loudness" setting?

3) When making a sdb/aud mix, is there a rule of thumb to go by?  example: 2 parts SDB and 1 part AUD.  Any tips or preferences from your expierence on mixing sdb and aud? 

Thanks in advance! 

Dithering - isn't that used for changing bit rates, and not sample rates?  I am pretty sure that the resample in SF (havent used it in a while...) has 4 levels of accuracy, and they all take the same amount of time to run, so you might as well be the most accurate.  I have always used 4.

Normalizing- SF has Peak normalize and RMS - if you are going peak, the idea is to go up to 0db.  If you are using RMS, I think you are going to get nasty results if you go above -6db.  But, as someone else stated, you should try to run your gear as hot as possible to capture dynamic range- you can't recreate that using the computer.

HTH - and please someone correct me if I am wrong here.

[SPLASTiK]

YOU ONLY DITHER WHEN CHANGING BIT RATES.

ala going from 24bit to 16bit.  Dither doesn't have anything to do with sample rates and it should be the last thing you do to the audio file.
edit: didnt see you said you were changing bit depth as well. But that should be the last thing you do to the audio files.

Normalizing- I've always been told to normalize to 98% or -0.2db by Audio Engineers

[therodge]

Thanks guys for all the input!

Ya, I'm a dummy.  I'm not changing the Bit Rate, I'm changing the Sample Rate.  So, I'm going to take Splastik's advice and not do it.

I might normalize the volume but it's already pretty good on volume so I probably won't.

On the SBD/AUD mix, It's gonna be awhile before I can get the other taper's AUD mix... I got the SDB, was going to try the mixing later for kicks.


pfife,
GREAT AVITAR!  the recording that we are discussing is O.A.R. at the Ryman Auditorium in Nashville, TN... Home of the Titans! haha.  I got tix to the Kansas City Monday Night Football game at home this year, I'm stoked!

[pfife]

pfife,
GREAT AVITAR!  the recording that we are discussing is O.A.R. at the Ryman Auditorium in Nashville, TN... Home of the Titans! haha.  I got tix to the Kansas City Monday Night Football game at home this year, I'm stoked!

Going a little OT here, but thanks man- I like the Titans (I have mad respect for Steve McNair, and I am sad that they got rid of George and Kearse) but I completely despise Kellen Winslow II - So, I figured it would be appropriate to see his ass getting smashed every time I log onto TS.com.   

+t  good luck on that recording

[hexyjones]

I've heard it said that you cannot matrix resampled material...time resolution sacraficed in resampling will cause a delay once you get a few songs beyond the intial alignment point...

Both sources must be mastered at the same rate to matrix...

I dont know if this is true...

[leegeddy]

I've heard it said that you cannot matrix resampled material...time resolution sacraficed in resampling will cause a delay once you get a few songs beyond the intial alignment point...

Both sources must be mastered at the same rate to matrix...

I dont know if this is true...

not true. you can matrix multiple amount of sources. i've matrixed 2 analog sources (sbd + audience) before. the difficulty level goes up as expected, but it can be done.

EDIT:  perhaps i misread the post above.  you can not matrix a 44.1khz and a 48khz source. both must be in the same sampling freq., either in 44.1k or 48k.

marc

[aberg]

Is there any clear cut way of mixing two sources and syncing them up properly? Any good software to do this with.. I always run into a timing issue... like it will be in sync near the point I sync it up but later in the recording, the delay becomes apparent.. and without deleting small bits, it will never line up... any way to fix that?

[hexyjones]

Is there any clear cut way of mixing two sources and syncing them up properly? Any good software to do this with.. I always run into a timing issue... like it will be in sync near the point I sync it up but later in the recording, the delay becomes apparent.. and without deleting small bits, it will never line up... any way to fix that?

I've heard others say the exact same thing...sync it at the beginning, but 5 songs into the track...the timing will changes and the later tracks wont sync...

What kind of sources are you using...?

I havent experimented yet...but I have a friend who had AUD and SBD of the same show - (but on MD) - he remastered them...and they would not sync...

I wonder if this is limited to MD? or do full PCM sources suffer the same time resolution problems...?

[leegeddy]

Is there any clear cut way of mixing two sources and syncing them up properly? Any good software to do this with.. I always run into a timing issue... like it will be in sync near the point I sync it up but later in the recording, the delay becomes apparent.. and without deleting small bits, it will never line up... any way to fix that?

I've heard others say the exact same thing...sync it at the beginning, but 5 songs into the track...the timing will changes and the later tracks wont sync...

What kind of sources are you using...?

I havent experimented yet...but I have a friend who had AUD and SBD of the same show - (but on MD) - he remastered them...and they would not sync...

I wonder if this is limited to MD? or do full PCM sources suffer the same time resolution problems...?

when i record local tribute bands, i would use my m1 for the aud and the d100 for the sbd. i transfer the 2 sources in the same manner and start working on the sync.  i've YET to have 2 sources sync perfectly without manipulations.

no 2 crystals are identical, hence the difference in the recording lengths, albeit very small. but, when you multiply the tiny offset by few hours, you will get a time difference that's very noticeable.

we all know that sony portables don't have an external word clock input, so post production work will have to do. 

hope this helps,
marc

ps. there was a thread several months back discussiing sync issues in detail. i can't find it now, but i know it's here somewhere.

[Uncle Jimmy]

OK, here's my deal. I had a video of a band (not the White Stripes) at a venue (not the Fillmore) that I also had a seperate (not stealth) recording of. I wnated to strip the audio from the video file and replace it. What i ended up having to do was use the original camera audio as a time template, if you will.

For you, choose one of your sources to be the "master track". Overlay the other track on top and see where you are at. For me, the best solution was to split the "secondary" track into individual songs (or managable chunks). I then compressed each song a little bit (LITTLE) and then match it up (drum strikes work well, starts of verses, something like that.) Then, do the same for the next song. make sure that any overlap or gap occurs during crowd noise where it is easy to mask. (e.g., take a 1/4 second of noise before the gap, reverse it, paste it over the gap. the levels match on one side perfectly.)

  This is the only way i could do it, as the result had to be the exact same duration as the original, or it would not have worked. Result was good and the only synch problems with the video were there to begin with.

YMMV,
-UJ

[dklein]

I find it easiest to extract the original audio from the videotape, then load it into a multitrack environment with the good audio and line them up (which is dead easy in multitrack).  Then pull out the good audio, mux it back into the clips and it's perfectly synched.  If the drift is too much, I've found the stretch feature in Wavelab (and only wavelab) still sounds good and can be used for fine tuning.

[therodge]

Thanks for all the help fellaz!

Here is a link to my first Matrix!: http://bt.etree.org/details.php?id=12205

It was of Robinella & the CC Stringband.  Tell me what you think!