Topic: Matching speed AUD + SBD

[travelinbeat]

Hey there folks, I know I've read about people doing this before, so I'm hoping you all may be able to help me out.

I recorded a show last night using both open air mics and a soundboard patch and now I am trying to line the two up.  The problem is that for some reason, my sources are not staying lined-up throughout the entire show.  I recorded the SBD onto my iRiver h120 and the open air using Church Cards into my Sony MZ-RH10 Hi-MD at PCM.  Both recordings are 16Bit, 44Hz, but the MD seems to fall away from the Soundboard recording.  I am using Sony Soundforge 8.0 and I am wondering how I can get my approx. 80min recordings to sync nicely.  Again, I'm able to line them up fine at the beginning, but after just a few minutes very early signs of deviation are already apparent and by the end of the mix it is nearly an incomprehensible jumble.

Thanks for any and all advice!

[lordbelial]

That because of internal clocks on the recorders. Even if you record at 44.1 Khz, it's not accurate, so if the Iriver records at 44.11 Khz and the MD records at 44.09 the show will be desaligned with time.

The only thing you can do, and I've dealt with this in the past, is.

Source one: AUD as single file
Source two: SDB, split into single tracks using CDWAVE

Then you can align them and they can be easily sincronized.

I know this is a pain in the ass but it's the only way i know.

Edit. I suggest you to use Wavelab to matrix, but only 'cause I don't know if Soundforge has this feature...

[stantheman1976]

Easier way is to use Sony Vegas or find someone who has it.  I wrote the following as a way to sync a separate audio track to a video.  You can still follow the main parts.  You just won't have the video track spoken of below.  Don't split the tracks up.  I never have.  With digital recordings the difference you're seeing will be constant so you can edit the waveforms as a whole.

Import both tracks into Vegas.  You'll use your original auido and visual reference.  Bring your new audio track in just below the original audio track.  Go to the beginning of your original audio track and listen for some type of standout event like a cymbal crash stick hits.  If it's just talking that will work too.  With speech the waveform will be quieter and have more nitceable peaks and valleys.  Use a good pair of headphones.  Move the new audio track so it starts to match up with the original.  Find those common events on both audio tracks to help you match them up.  Move the new audio track left or right accordingly until it plays in sync with the original.  When they are both lined up correctly it will sound like one single track when they're played together.  It's really not that hard.  It just takes a little practice to get the hang of things.

One issue you may experience is time drift.  Both sources my be perfectly aligned starting out but will be out of sync later on.  This is caused because the internal mechanisms of different devices might run at slightly different speeds.  So even though you're recording the same thing one device might move a couple milliseconds slower than the other.  With a long single recording it adds up. 

Here's how to remedy that.

First off, go to your options and make sure "Quantize To Frames" is NOT checked.  If this is checked then when you make an adjustment it has to be one complete frame at a time.  If it is unchecked you can make the adjustment as small as you want.

Also, go look in your options and make sure audio resampling is set to highest quality.

Now, go to the end of your new audio track and compare it visually to your original track.  Hopefully you should see some similar points in each and be able to tell how far off they are.  If your original audio track is distorted too bad or AGC was used it may be almost square.  Use your ears in that case.  Play the tracks together and you should be able to tell if your new track is ahead or behind your original.  You may have to adjust the track volume of one or the other to help out.  Just make sure you set the new audio track back to 0 before you render.  Go to the end of the new audio track and hold your mouse cursor right over the end.  Hold down the CTRL key and you'll see the "~" symbol appear.  Keep that CTRL key held down and move your mouse to the left or right.  You are now stretching or shrinking your audio track in real time.  Move it in small increments and listen and compare your two tracks visually each time.  You might have to zoom in and move in extremely small increments to get it right.  Once you're done both tracks should sound like one when played together. 

[nedstruzz]

Do any other editing programs do what you describe above?  Say Pro Tools or Adobe Audition?

[stantheman1976]

I don't think so.  You can adjust audio easily with those programs but you can't do it visually like Vegas can. 

[live2496]

Do any other editing programs do what you describe above?  Say Pro Tools or Adobe Audition?

In samplitude you can do this in the object editor. You put both stereo files onto two separate tracks. Each file is an object in Samplitude. You have to be in the context of a virtual project.

Then line up the beginning of both objects exactly with so that the drums sound synchronized. You have to use a cymbal sound, high hat or some event that stands out.

Then I go to the very end and slide the bottom object in time until some event at the very end lines up exactly. Then I determine how many samples the two files are out of sync. To do this you have to set the display to show samples and determine how many samples the object has shifted to get it to sync up at the end.

Then realign the objects at the beginning again, and finally enter a number into the object editor that tells the software how many samples to stretch the audio. (You do this to only one object and leave the other alone.)

Upon playback, the object with the time stretch value will be resampled to stay in sync with the other object.

I did this on a multitrack project recently and it worked perfectly.

Gordon

[nedstruzz]

If anyone knows how to do this with Audition it would be hugely appreciated. 

[travelinbeat]

Easier way is to use Sony Vegas [...]

You are AWESOME.  I had zero familiarity with this software, and after hours of wrestling with Soundforge, I read your post, got the software, and within 10 minutes the entire job is DONE.  MAJOR +t, you saved the day! 

[stantheman1976]

Glad it worked for you.  I love Vegas for its versatility.  There are just so many things that it can do and there's always new tricks to explore and discover.

[MarkE]

       This is great. I had the exact same thing happen to me this week.. An aud, a board, not matched up.. Heres what I did... I use Magix Audio Cleaning Lab.. I like it alot.. It set it up as 2 tracks. It was time consuming, but I did the every song thing..

      And also listened to it (for the most part). Found places where you could audibly hear it going from perfect match to slight delay in the snare drum. Id split it there ansd re-alighn it manually. I got almost every spot. It took a couple sit downd, ut it came out nice..
Great post +T
MarkE

[travelinbeat]

here's the end result in Kickdown Central-- Kimya Dawson from The Moldy Peaches & the "Juno" soundtrack

http://taperssection.com/index.php/topic,101877.0.html

[anr]

I think Wavelab has an appalling interface, so this is the only occasion I use it.  As described above, open both tracks in Montage.  Assuming both files start and finish at the same point (a simple edit) make sure they are aligned at the beginning (not immediately obvious in Wavelab as they open at the cursor - drag both to the left).  Place the cursor at the end of the longer file, right click in the other file and click "stretch to cursor". 

This assumes the recordings are linear.  When one source is, for example, an old tape that has perhaps been stretched, then I'd say you have to treat each track separately, then stitch them together. 

[stantheman1976]

When dealing with an analog source you will have to use even smaller increments than single songs.  Analog sources are a PITA to sync.

[stantheman1976]

here's the end result in Kickdown Central-- Kimya Dawson from The Moldy Peaches & the "Juno" soundtrack

http://taperssection.com/index.php/topic,101877.0.html

I'll check it out.

If you transfer MD with 1/8" I'd recommend going line in to your sound card instead of mic though.

[travelinbeat]

If you transfer MD with 1/8" I'd recommend going line in to your sound card instead of mic though.

I was thinking that same thing, but I couldn't justify why... so out of curiosity, why would this be preferable?

I ended up just doing mic after considering that the whole source would be turned-down so much anyways.

[batchain]

If anyone knows how to do this with Audition it would be hugely appreciated. 

Here's how i do it in Audition 3.0, but I think it's the same in 2.0 . First make sure you're working with 32bit files to minimize editing artifacts. Go to Effects>Time Pitch>Stretch(process). You can stretch or shorten a file using the Ratio which is a % of the original file length.

To figure the % you need to shorten it by you have to go into the Multitrack section. I line up the two files at or near the beginning of the recording. It's real obvious if you listen to the vocal banter between songs to get the echo out. Then go to the end of the two tracks - you should be able to see how much they are out of sync. Now just highlight and measure that difference which is usually going to be in so many thousandth of seconds, but don't worry about the measurement length look at the begin and end points of your measured difference. This is going to be in H:M:S format which you will need to do a little math to convert the entire two files to seconds. By dividing the larger file into the smaller file you'll get the ratio of the difference between the two.

Now back to the Stretch process select the longer file and apply that number (usually 99.XXXx) to shorten the file making sure to enable maintain pitch. It can still be little bit of trial an error because you are correcting to a point that may not be at the end of the show thus not giving you a fully accurate % amount. So if the two files still don't sync up then experiment with slightly higher number for the "small x value" at the end of the 99.XXXx number and go back and redo the stretch process again from the original length file. I'd be very hesitant to stretch the same stretched file again.

[stantheman1976]

If you transfer MD with 1/8" I'd recommend going line in to your sound card instead of mic though.

I was thinking that same thing, but I couldn't justify why... so out of curiosity, why would this be preferable?

I ended up just doing mic after considering that the whole source would be turned-down so much anyways.

Mic in may boost the signal when it's coming in.  Line in takes the signal exactly as it's being fed and doesn't change anything.  Unless you're using an actual microphone line in is always best to use.