Taperssection.com
Gear / Technical Help => Post-Processing, Computer / Streaming / Internet Devices & Related Activity => Topic started by: NewTaper on June 21, 2018, 11:16:17 AM
-
I will PayPal in advance anyone who can put together a video that shows exactly how to time align to audio tracks in Adobe Audition due to digital drift.
Soundboard and Audience mix or two audience tapes etc.
I have read a few tutorials but I just can't get it. I am a visual learner so if I can see this done once I can do it and I am sure there are many who would benefit from a video like this!
Please Help.
Thanks,
NT
-
Sorry if this should be in classifieds?
I will PayPal anyone who can put together a video that shows exactly how to time align to audio tracks in Adobe Audition due to digital drift.
Soundboard and Audience mix or two audience tapes etc.
I have read a few tutorials but I just can't get it. I am a visual learner so if I can see this done once I can do it and I am sure there are many who would benefit from a video like this!
I have never used Audition, but the task you need is two alignment sync spots, one near the start and one near the end of the files. When you find these and mark them relative to the track you're not planning to change, you will also need to mark the start of the moved track at each point, so you can measure the difference (how far you moved it.)
After finding the points, I use a spreadsheet to calculate the math (stretch or squash to sync) to make them the same length for the overlapped section.
Once one has been adjusted in length, then simply slide it to match the other (reference) at the first sync point, then check to see if the math worked because the second point should be correct now as well
Then you can mix them.
Sorry it's not a video, but since I don't use Adobe anyhow, my Audacity movie might not help you. ¯\_(ツ)_/¯
That's about as short as I can describe my process. Hope it helps.
-
https://www.youtube.com/watch?v=MOz2GUdL_vw&feature=youtu.be
I did this some years ago.
I can provide you with the excel sheet I wrote that does the math calculations.
About halfway through, you can hear my cats fighting in the background. I left that in for laughs.
Hope this helps!
Uses combination of AA 3.0 and AA-CC. When they rebuilt the engine in AA 5 and on, the new Time Stretch module just isn't as accurate as the time stretch in AA 3.0. So I do everything in AA-CC, and then do the actual stretch in AA 3.
-
download audacity and then follow Morst's instructions here:
http://taperssection.com/index.php?topic=103031.msg1373822#msg1373822
works like a charm.
-
download audacity and then follow Morst's instructions here:
http://taperssection.com/index.php?topic=103031.msg1373822#msg1373822
works like a charm.
I thought I posted my handy spreadsheet link on that thread so folks could enjoy it but I can't seem to find it.
Here is my spreadsheet for calculating the speed change:
https://docs.google.com/spreadsheets/d/1pQGfYwPgBFFzcY5m6aRj-Zbu9HsRumLy-tJB1d8Eufg/edit#gid=583050244
If you can't use it live in a browser, just download it and use it at home.
-
I've found that doing really close alignment will need more than one point. If it's a fairly simple iem/ald/sbd and AUD, then you can probably get away with doing alignment markers every hour. Mixing multiple line sources reveal misalignments quickly, and you'll need to align every 10-15 minutes.
-
I've found that doing really close alignment will need more than one point. If it's a fairly simple iem/ald/sbd and AUD, then you can probably get away with doing alignment markers every hour. Mixing multiple line sources reveal misalignments quickly, and you'll need to align every 10-15 minutes.
I'm talking about stretching (or squashing) one source to match, rather than doing a cut-and-paste pastiche of parts of sources from different clocks.
-
Exactly what I'm talking about to. But digital clock speed don't run at the same speed. So even if you do a time stretch over a two hour period, and its aligned at beginning and end, it will be off a bit in the middle.
-
Exactly what I'm talking about to. But digital clock speed don't run at the same speed. So even if you do a time stretch over a two hour period, and its aligned at beginning and end, it will be off a bit in the middle.
It won't be off in the middle enough to hear, in my experience, unless you're missing samples or parts. But I'm not working with Nomad JB sources!? :-P
If the difference between two sources is less than 10 ms (with the closest source such as the SBD happening first in case they aren't perfect), I find that it's not noticeable. 10 ms is 480 samples at 48kHz. In practice I don't mind fudging up to 20 ms but that's pushing it.
I guess if you were running from cold to warm environments and back, that could cause larger variations?
I don't think I'd be able to sync every song. There's not always a proper clean wave to match up. Besides, what if you match the bass guitar at the start of one song, then sync to a vocal noise at the end? They are not necessarily happening at the same times?!
I have now watched your video, and I see that you are syncing two sources which both have click tracks! That is something I've never encountered! And yet, you have enormous differences in percentages in your song-by-song spreadsheet. I can see from that evidence why you are drawing the conclusions you do, but I question the process. That variation is far too large. If you were to do it "my way" with one point near the start and one near the end, and then calculate the overall percent for the whole file set, you could at least use that as a baseline to compare each track percent to see if it's close!? You should have some higher and some lower if your math is right, and they're not all the same!
-
Yeah, even if two clocks are running at different speeds from each other, they'll each be internally consistent, no?
I've never had any problems just setting two points as far apart as reasonably possible and then deriving a single ratio from that. But, obviously, YMMV.
-
Yeah, even if two clocks are running at different speeds from each other, they'll each be internally consistent, no?
A little bit of variance or error is to be expected, but one would hope that the jitter or drift of each clock from a theoretical perfect clock should be somewhat minimal!? ¯\_(ツ)_/¯
-
The audio in the video clip is a U2 IEM recording, which explains the click. It's even the same feed, although different receiver types. Both are to two independent nomad jb3 (long since retired recorders!), which generally had a pretty good time clock from what I found.
When you're working with line recordings (like IEMs), you hear alignment issues much quicker than with audience recordings. I can hear when things are off by a couple samples. That hollow tunnel sound on vocals is unmistakable.
I have found that the internal clocks on decks are NOT consistent, and there is a slight curve. So if you align something from
00:00:00 - 00:30:00 - 99.999975%
00:30:00 - 01:00:00 - 99.999832%
01:00:00 - 01:30:00 - 99.999712%
I'm making up numbers but there is a change in the alignment. Now I don't know any time alignment program that can calculate the derivative, and make a proper alignment based on that. However, next best thing was to do it in intervals.
But like I said--audience recording mixing is much more forgiving. I typically align audience recordings on an hour basis, whereas much more closely on multiple line (iem/ald/sbd) recordings).
In any case, I get the $25 from the OP for providing a video :)
-
I have found that the internal clocks on decks are NOT consistent, and there is a slight curve. So if you align something from
00:00:00 - 00:30:00 - 99.999975%
00:30:00 - 01:00:00 - 99.999832%
01:00:00 - 01:30:00 - 99.999712%
That is bizarre and fascinating. Which decks have you noticed this with?
-
All, once you get precise enough.
-
Thank you to everyone for their help. I look forward to trying this out this weekend.
Thanks,NT
-
That is bizarre and fascinating. Which decks have you noticed this with?
The speed of sound changes with temperature and humidity. So the speed of sound will be different in a club at the beginning of a show when people just walked in, vs later in the evening when people start to sweat and raise the temperature and humidity level.
If you're mixing 2 pairs of audience mics in the same spot, the differences will mainly be due to clock drift between the two recorders. But if you're mixing a video (speed of light) with audience mics, or a soundboard feed (close enough to the speed of light) with audience mics, you will have the issue of clock drift between the recorders and also the drifting speed of sound (which changes subtly) relative to the constant speed of light or electricity. Short answer, the varying speed of sound will make the drift between sources change over time, beyond that from clock drift between recorders alone.
-
That is bizarre and fascinating. Which decks have you noticed this with?
The speed of sound changes with temperature and humidity. So the speed of sound will be different in a club at the beginning of a show when people just walked in, vs later in the evening when people start to sweat and raise the temperature and humidity level.
If you're mixing 2 pairs of audience mics in the same spot, the differences will mainly be due to clock drift between the two recorders. But if you're mixing a video (speed of light) with audience mics, or a soundboard feed (close enough to the speed of light) with audience mics, you will have the issue of clock drift between the recorders and also the drifting speed of sound (which changes subtly) relative to the constant speed of light or electricity. Short answer, the varying speed of sound will make the drift between sources change over time, beyond that from clock drift between recorders alone.
Interesting!
-
Short answer, the varying speed of sound will make the drift between sources change over time, beyond that from clock drift between recorders alone.
Oh very good point.
My favorite handy widget for calculating stuff related to speed of sound is
https://www.apple.com/downloads/dashboard/calculate_convert/soundreference.html
Sound Reference (sorry windows and linux folks)
It shows that the speed of sound at
44 degrees F as 1100 f/s
72 degrees F as 1131 f/s
92 degrees F as 1153 f/s