Taperssection.com

Gear / Technical Help => Ask The Tapers => Topic started by: Elguapo511 on April 15, 2018, 04:34:57 PM

Title: Does digital drift exist? (marco benevento at schubas chicago cont)
Post by: Elguapo511 on April 15, 2018, 04:34:57 PM
i recorded this show w four miss on the stage (dr2d four track recorder)
Then I plugged in a two track recorder (dr07) to the sound board.

no problems. and a totally radical show.
but

I loaded them all into audacity and aligned them, yet there is still drift at the end.  Does digital drift exist.

What tools do are available for making a matrix and do i have cut the tracks and re align every track?

thanks for any help.
Title: Re: Does digital drift exist? (marco benevento at schubas chicago cont)
Post by: kingdong on April 15, 2018, 06:15:06 PM
Two different devices may/will have internal clocks that run at slightly different speed.  Over the length of a whole show you will usually see a slightly different file length.  When I have to fix this I typically use Audacity and stretch the shorter file to be the length of the longer.
http://manual.audacityteam.org/man/change_speed.html

The process I loosely use is to open both waveforms in the same Audacity project and align the beginning of the files.  Then I go the the end of the files and look for an identifiable peak.  I will note the time difference between the two and use that amount to stretch the shorter file.  Since we are typically talking about less than a second over the course of a set I don't worry about the very slight pitch shift that this will induce.  I haven't noticed it to cause phasing issues or other atrifacts, but I don't have the most discerning ear on this forum by a long shot.
Title: Re: Does digital drift exist? (marco benevento at schubas chicago cont)
Post by: if_then_else on April 16, 2018, 01:28:49 AM
Even if the tracks are aligned perfectly there might be some issues with the timestretching / pitch correction algorithm on the resulting mix-down file  - e.g. depending on the genre. E.g. I was having some issues with 2 psychedelic rock and drone music recordings and the default algorithm in Samplitude 10.2. No issues with the same project and the "élastique pro" algorithm.
Title: Re: Does digital drift exist? (marco benevento at schubas chicago cont)
Post by: Elguapo511 on April 16, 2018, 04:31:48 PM
gotcha.  this is what i was looking for.  Ill give it a try.

also,  any real time eq available for mac? and free
Title: Re: Does digital drift exist? (marco benevento at schubas chicago cont)
Post by: Elguapo511 on April 17, 2018, 12:20:29 AM
sooo.... i recorded the show with four mics and two soundboard channels.

is there a way to mark a point in one track, then mark another point in another track.  then match those points up automatically.


I am having trouble aligning the tracks, i can't seem to get them together.   

i used math, and the track move tool.

just wondering if I can do it with less guessing.

thanks
Title: Re: Does digital drift exist? (marco benevento at schubas chicago cont)
Post by: if_then_else on April 17, 2018, 01:57:13 AM
Set a marker at a single cymbal or kick drum hit at the beginning of the show.
Do the same thing for a single event at the end of your first stereo track.
Now lock the individual channels and vary the time stretch factor on the second source until both stereo tracks are aligned correctly.
Listen back to the aligned tracks. If they still sound out of phase, unlock the second source and carefully move ("nudge") the stereo track to the left or right by a few samples. Repeat until you're happy with the result.
With my own recorders and the typical distances at small to medium sized venues you're looking into a ratio of ~1.000004:1 - ~1.000006:1.
Title: Re: Does digital drift exist? (marco benevento at schubas chicago cont)
Post by: twatts (pants are so over-rated...) on April 17, 2018, 02:23:20 AM
https://www.sounddevices.com/tech-notes/what-is-timecode-overview-of-timecode-for-sound-and-picture
Title: Re: Does digital drift exist? (marco benevento at schubas chicago cont)
Post by: nulldogmas on April 17, 2018, 04:16:35 PM
Set a marker at a single cymbal or kick drum hit at the beginning of the show.
Do the same thing for a single event at the end of your first stereo track.
Now lock the individual channels and vary the time stretch factor on the second source until both stereo tracks are aligned correctly.
Listen back to the aligned tracks. If they still sound out of phase, unlock the second source and carefully move ("nudge") the stereo track to the left or right by a few samples. Repeat until you're happy with the result.
With my own recorders and the typical distances at small to medium sized venues you're looking into a ratio of ~1.000004:1 - ~1.000006:1.

^This.

Note that it doesn't have to be the very beginning and very end of the show — so long as you have two good markers a reasonable distance apart, you should have enough spread to get a good measurement of the drift.
Title: Re: Does digital drift exist? (marco benevento at schubas chicago cont)
Post by: wforwumbo on April 18, 2018, 12:13:47 PM
Question... we’re both recorders at the same sample rate? If you had to resample one track, there will be an inherent linear group delay. With linear filters, this will result in minimal (if any audible) distortion, but it will mean that tracks won’t align without time stretching with a phase vocoder.
Title: Re: Does digital drift exist? (marco benevento at schubas chicago cont)
Post by: Elguapo511 on April 18, 2018, 09:15:33 PM
both 16 bit.   but thanks for the tip.

the pics are very different and picking up different frequency is making it hard to find alignment....
Title: Re: Does digital drift exist? (marco benevento at schubas chicago cont)
Post by: nulldogmas on April 18, 2018, 09:37:54 PM
A loud clap, start of PA music, anything? It doesn't have to be perfect alignment, just closer than what you have now.
Title: Re: Does digital drift exist? (marco benevento at schubas chicago cont)
Post by: Elguapo511 on April 18, 2018, 09:59:11 PM
this is pretty impossible.

 :banging head:

it still seems to get nuts at the end.

does it matter if i slow one down  or speed one up?
Title: Re: Does digital drift exist? (marco benevento at schubas chicago cont)
Post by: noahbickart on April 18, 2018, 11:27:05 PM
Question... we’re both recorders at the same sample rate? If you had to resample one track, there will be an inherent linear group delay. With linear filters, this will result in minimal (if any audible) distortion, but it will mean that tracks won’t align without time stretching with a phase vocoder.
Title: Re: Does digital drift exist? (marco benevento at schubas chicago cont)
Post by: nulldogmas on April 19, 2018, 07:30:39 AM

does it matter if i slow one down  or speed one up?

Nope.
Title: Re: Does digital drift exist? (marco benevento at schubas chicago cont)
Post by: capnhook on April 19, 2018, 08:22:25 AM
this is pretty impossible.

 :banging head:

it still seems to get nuts at the end.

does it matter if i slow one down  or speed one up?

Are you using a version of Audacity, other than V2.1.0?
Title: Re: Does digital drift exist? (marco benevento at schubas chicago cont)
Post by: Elguapo511 on April 19, 2018, 02:01:22 PM
Audacity 2.1.3
Title: Re: Does digital drift exist? (marco benevento at schubas chicago cont)
Post by: Elguapo511 on April 19, 2018, 02:03:50 PM
if anyone is interested in aligning these two tracks, it will speed up my mixing process.  this show is a week old.  does anyone even listen to music?

i will pay you back in jams
Title: Re: Does digital drift exist? (marco benevento at schubas chicago cont)
Post by: capnhook on April 19, 2018, 02:12:49 PM
There you go, mystery solved.

In Audacity versions above V2.1.0 there is a problem with "Change Speed".

Values you might enter for the amount of speed change get changed, without any notice to you.....the value's digits get truncated for some damn reason.

Say, if you enter 0.001500652783961 for the factor, the program truncates the value to 0.0015.  WRONG.

Go back to V2.1.0 and I predict you will be pleased.  Why the team broke something that worked properly is anybody's guess.

I never asked anyone why, I'm just standing firm on V2.1.0 until I hear this regression has been addressed.
Title: Re: Does digital drift exist? (marco benevento at schubas chicago cont)
Post by: Elguapo511 on April 20, 2018, 10:37:59 AM
it worked.

all hail capn hook.

After all this, I am wondering if there is a better free software that mac users use for post production?
having to wait to process while eq is making this also very difficult to A/B levels.

all hail capnhook and this hear board
Title: Re: Does digital drift exist? (marco benevento at schubas chicago cont)
Post by: nulldogmas on April 21, 2018, 08:25:40 AM
Yay success!

I've found Sound Studio to be excellent for aligning two sources, but it's not free ($30).
Title: Re: Does digital drift exist? (marco benevento at schubas chicago cont)
Post by: capnhook on April 21, 2018, 08:56:00 PM
it worked.

all hail capnhook.

After all this, I am wondering if there is a better free software that mac users use for post production?
having to wait to process while eq is making this also very difficult to A/B levels.

all hail capnhook and this hear board

Congratulations Elguapo, and here's to making more great tapes..!