Become a Site Supporter and Never see Ads again!

Author Topic: To shrink or stretch?  (Read 10009 times)

0 Members and 1 Guest are viewing this topic.

Offline 2manyrocks

  • Trade Count: (12)
  • Taperssection All-Star
  • ****
  • Posts: 1664
Re: To shrink or stretch?
« Reply #15 on: July 25, 2014, 10:05:08 AM »
I gave shrinking a shot last night in Reaper to align two tracks from two different recorders by pulling the end of one track backwards to line up the peaks in the other track.  Although I was able to sync the two tracks this way, it was very "processed" sounding, and must not be the right way to do it as the result was awful sounding.   :(

Online Gutbucket

  • record > listen > revise technique
  • Trade Count: (16)
  • Needs to get out more...
  • *****
  • Posts: 15700
  • Gender: Male
  • "Better to love music than respect it" ~Stravinsky
Re: To shrink or stretch?
« Reply #16 on: July 25, 2014, 12:06:58 PM »
Quality of results depends on the algorithm. Here's a couple snippets from the section of the  Samplitude 11 user manual (the version previous to the current release) appropriate to what we are discussing.  It also includes other time-streching algorithm modes more appropriate for things like drums, loops, monophonic samples, and more extreme stretch factors which I've not included below.  Note that it mentions shrinking (length-reduction / time-compression) is less artifact prone than stretching.

I use the Universal HQ mode.

Resampling
Samplers and PCM synthesizers transpose samples using this procedure. Time factor and pitch are dependent upon each other: the shorter the audio material, the higher the pitch and vice versa. The effect is comparable with changing the playing speed of record players or tape recorders. The effect is relatively loss-free, the sound loss is smaller than with all Timestretching/Pitchshifting procedures. If a pitch change is wanted or justified by changing the speed, use this algorithm.
 
Pitchshifting/Timestretching
With all other algorithms, the time factor and pitch are independently adjustable. These algorithms may create artifacts, which can be balanced out using the anti-aliasing filter. The preset timestretching/pitchshifting algorithm can be saved via Timestretch / Pitchshift Patcher (view page 621) ("Tools" menu).
 
Standard
Here, an algorithm is used that usually delivers very good results with factors from 0.9 to 1.1, that works in phase-locked mode, and thereby keeps the room effect of stereo signals. This algorithm is only partially suitable for drum loops or other "beat heavy" audio material as it can change the groove.

Time compression (sample length is reduced) is more successful with this algorithm than timestretching, i.e. it is better to reduce the longer sample than vice versa. This algorithm is especially suited to complex audio material. The CPU load is light. At extreme settings, anti-aliasing effects may occur. An additional anti-aliasing filter can come to your aid. This filter is also available for all other algorithms that work internally with resampling.

Universal HQ
This algorithm serves as a high-quality timestretching/pitchshifting method and offers good audio quality with almost any audio material. Especially when it comes to complex audio recordings like orchestral recordings, this algorithm delivers especially good results. The stereo properties remain intact. The CPU load can become very high
« Last Edit: July 25, 2014, 12:08:45 PM by Gutbucket »
musical volition > vibrations > voltages > numeric values > voltages > vibrations> virtual teleportation time-machine experience
Better recording made easy - >>Improved PAS table<< | Made excellent- >>click here to download the Oddball Microphone Technique illustrated PDF booklet<< (note: This is a 1st draft, now several years old and in need of revision!  Stay tuned)

Offline jagraham

  • Trade Count: (20)
  • Needs to get out more...
  • *****
  • Posts: 2996
  • Gender: Male
Re: To shrink or stretch?
« Reply #17 on: August 03, 2014, 07:42:21 PM »
For all the trouble in aligning between different recorders, I'm sort of thinking of selling my Dr-2d and buying another M10.  Wondering if two M10s will record closely enough to avoid any alignment issues in recordings under an hour and a half?

If the M-10 is anything like the Tascam units, no. It's possible they are made with better aligned internal clocks but any Tascam units I have compared in the past have had .5-1 second drift per hour resulting in a huge echo. Anyone that's attempted this with 2 M-10s that could provide input here?

Just curious - Why wouldn't you just keep and use the DR-2d if you are doing 4ch with miniplug mics? That's exactly how I use it and I find it to be perfect for the function. Not to mention other mics it can be used with and SBD matrices. It makes 4ch audience recording very simple, no need to worry about alignment in post.

I'd alter the source I'd be using the least of in the mix....

The chop and align method is flawed because the sources are slightly out of tune...avoid.

^ Agreed with all this. There's no way it matters if you shrink or stretch, just make sure to edit the source less prominent (especially in a matrix where one source has an issue you are trying to cover up). I wouldn't even think about chop and align, as I would second guess the whole thing to the point of not being able to enjoy the recording. What about songs that are longer than others, they are just supposed to be out of alignment making the recording inconsistent? Perhaps there's an appropriate threshold of time before it can be noticed audibly but I just would avoid it period.

The only issue with the "math and stretch" method that I've encountered was once the files would not align due to some kind of digital issue. Used the same units other times and aligned successfully. Not sure how frequent this is but it happened, couldn't get that recording to align no matter what.
Mics: Nak CM-300s, Nak CM-100s, CP-1s, CP-2s, AT-853s(Cards, Hypers, Omnis) CA-14s(Cards, Omnis)
Pres: CA STC-9200, CA-UBB
Recorders: Tascam DR-70D, DR-2D, Edirol R-09

ISO: 1 Teac ME-120, CP-3 Caps, AT-853 Subcard Caps

Offline 2manyrocks

  • Trade Count: (12)
  • Taperssection All-Star
  • ****
  • Posts: 1664
Re: To shrink or stretch?
« Reply #18 on: August 04, 2014, 11:26:25 AM »
To answer your logical question why not to keep the Dr2d, based on the comparison testing I've done, it seems that the M10 runs at a speed that is closest to my Canon HF M500 video camera, the battery life on the M10 is longer, and I am more comfortable with the M10 controls.  However, a single DR2d would be easier to manage than two M10s and I already have the DR2d. 

For me to go back to using the DR2d, I'd have to get better at managing its battery life and need to understand how to best match up the DR2d audio to my camera audio without making it sound terrible in the process. 

short answer:  I need to learn how to shrink/stretch as I appear to be going at it the wrong way.   

 

Offline Scooter123

  • "I am not an alcoholic. I am a drunk. Drunks don't go to meetings."
  • Trade Count: (9)
  • Needs to get out more...
  • *****
  • Posts: 3774
Re: To shrink or stretch?
« Reply #19 on: August 06, 2014, 11:28:16 AM »
The way I was taught--

It doesn't matter whether you shrink or stretch.  But you do need to pick the recorder which has the best time clock as your base.

Line up the files near the beginning with a sharp note, bass notes don't work so well.

Go out 30-45 minutes and select another sharp note and drop another marker on the  two files, separately. 

Get out of the digital time mode and display samples.

Subtract the difference in the samples between the two points and get the percentage, whether it is .999875 or 100,00123  Doesn't matter.  Stretch or Shrink

Time digital display is not accurate enough

I find working with most DAWs for this project to be annoying because when you place a marker, it does across all the files, which is not useful for time alignment.  So I use a two channel waveform editor to place the markers and go back and forth between the waveform editor and the DAW multi-track.  Adobe Audition either cc or 3 has some advantages in this regard in that it combines a two channel waveform editor with a multi-track and you can seamlessly pop between the two. 

Repeat the marker drill every 30-45 minutes. 

Problems may surface when the recorder malfunctions dropping a section even ever so small, and finding those write errors is, well, I just don't have the skill set to do that. 



Regards,
Scooter123

mk41 > N Box  > Sony M-10
mk4 > N Box > Sony M-10

Offline Sloan Simpson

  • Trade Count: (1)
  • Needs to get out more...
  • *****
  • Posts: 4013
  • Gender: Male
    • Southern Shelter
Re: To shrink or stretch?
« Reply #20 on: August 06, 2014, 01:37:36 PM »
ProTools' Elastic Audio feature makes this really easy. Markers are automatically inserted at all transients during the recordings. You line up the first transient, "lock" it into place, stretch or shrink the final transient as needed, and proceed with rendering. PT makes sure all the transients remain aligned throughout. I hope something like this will makes its way to Reaper eventually.
« Last Edit: August 06, 2014, 05:20:22 PM by Sloan Simpson »

Offline Scooter123

  • "I am not an alcoholic. I am a drunk. Drunks don't go to meetings."
  • Trade Count: (9)
  • Needs to get out more...
  • *****
  • Posts: 3774
Re: To shrink or stretch?
« Reply #21 on: August 08, 2014, 11:06:35 AM »
Question on Pro Tools Method

From my review of the videos on this feature, ProTools automatically places the transient markers on the two tracks.  Do you know if that is by beat or areas in which ProTools spots similar distinct notes.  I assume it by the beat. 

Also stretching seems to be done manually, is that correct?

Regards,
Scooter123

mk41 > N Box  > Sony M-10
mk4 > N Box > Sony M-10

Offline Sloan Simpson

  • Trade Count: (1)
  • Needs to get out more...
  • *****
  • Posts: 4013
  • Gender: Male
    • Southern Shelter
Re: To shrink or stretch?
« Reply #22 on: August 09, 2014, 10:00:55 PM »
Question on Pro Tools Method

From my review of the videos on this feature, ProTools automatically places the transient markers on the two tracks.  Do you know if that is by beat or areas in which ProTools spots similar distinct notes.  I assume it by the beat. 

Also stretching seems to be done manually, is that correct?

By the beat, or whatever is producing the transient. The process of placing the transient markers is done separately for each track.

Then you move around whichever track necessary to align the markers for the first transient in the recording. You then "lock" that transient for the track you're going to shrink/stretch, so that it becomes an "end point" so to speak.  Then you go to the end of the file and manually drag the last transient so that it aligns with the other track. The process aligns the transients throughout the recording (though I couldn't explain the algorithm for this).

Offline Scooter123

  • "I am not an alcoholic. I am a drunk. Drunks don't go to meetings."
  • Trade Count: (9)
  • Needs to get out more...
  • *****
  • Posts: 3774
Re: To shrink or stretch?
« Reply #23 on: August 12, 2014, 11:10:40 AM »
I doubt any automated placement of markers by the beat would be accurate enough for me.  The most accurate method to place markers is, in my judgment, by visual inspection of the waveform.   
Regards,
Scooter123

mk41 > N Box  > Sony M-10
mk4 > N Box > Sony M-10

Offline Sloan Simpson

  • Trade Count: (1)
  • Needs to get out more...
  • *****
  • Posts: 4013
  • Gender: Male
    • Southern Shelter
Re: To shrink or stretch?
« Reply #24 on: August 12, 2014, 04:47:30 PM »
I doubt any automated placement of markers by the beat would be accurate enough for me.  The most accurate method to place markers is, in my judgment, by visual inspection of the waveform.   

I'm probably not explaining it very well but it has worked flawlessly for me. I personally wouldn't sync two separate sources any other way.

Offline montanamon

  • Trade Count: (0)
  • Taperssection Newbie
  • *
  • Posts: 11
  • NeumannKM-184's (matched pair), EdirolR-4Pro (Oade
Re: To shrink or stretch?
« Reply #25 on: August 12, 2014, 06:48:23 PM »
I doubt any automated placement of markers by the beat would be accurate enough for me.  The most accurate method to place markers is, in my judgment, by visual inspection of the waveform.   

I absolutely agree Scooter123 - after MUCH trial an error with the "math" and automated attempts - visably lining it up is the best recommendation i could also make. I have no doubt there maybe a program (or 10) that works perfectly - i just have yet to use it.... anyhoo great little forum for those trying to sync up multiple sources and not using a 4 channel+ recorder! I recorded many SBDs and AUDs on different recorders to later sync and that is why this forum is valuable personally however, after purchasing my R-4Pro i would never go back and do it that way again - the first Matrix i did with my R-4Pro i had synced perfectly (b/c of the same time clock running all 4 channels) in minutes vs HOURS 'n HOURS! All i have to do is pull the AUD into sync with the SBD due to how much delay i had recording (due to the length away fromt the PA my mics were). Anyhoo great post :) cheers from MT

runonce

  • Guest
  • Trade Count: (0)
Re: To shrink or stretch?
« Reply #26 on: August 12, 2014, 07:47:26 PM »
I doubt any automated placement of markers by the beat would be accurate enough for me.  The most accurate method to place markers is, in my judgment, by visual inspection of the waveform.   

I took his explanation as another approach to correcting for clock drift - and not as a means of making the final time sync - I would think you could still slide the tracks around in time after correcting for drift...

Offline hoserama

  • Trade Count: (1)
  • Taperssection Member
  • ***
  • Posts: 519
  • Gender: Male
Re: To shrink or stretch?
« Reply #27 on: August 13, 2014, 09:52:41 AM »
Scooter123 is using my approach--putting in sync markers between different feeds, then finding the difference between the different markers, then time stretching it with the resultant ratio. No need to slide it around. If you do it right, in theory, they should be completely matched the whole way through.

I would think the automated version would work for a project not needing a lot of precision, but when you have projects that have only a margin of error between +/- 2 samples, then I wouldn't trust it at all. It would probably work okay with a clean aud/sbd, although have to be careful that claps on a drum/transient wouldn't throw it off.
Audio: Countryman B3 + AT853(hypers/cards/subcards) + SBD feeds
Wireless Receivers: Lots of those
Antennas: Lots of those
Cables: Lots of those
Recorders: TE TX-6, Zoom L20R, Zoom F8, (3) Tascam 680, (3) Tascam 2D, Zoom H6, and a graveyard of irivers/nomads/minidiscs.

Offline Sloan Simpson

  • Trade Count: (1)
  • Needs to get out more...
  • *****
  • Posts: 4013
  • Gender: Male
    • Southern Shelter
Re: To shrink or stretch?
« Reply #28 on: August 13, 2014, 05:43:39 PM »
Here's an example screenshot (only one file obviously). Looks pretty thorough to me. If you guys are inserting this many markers manually I'm impressed  ;D


Offline hoserama

  • Trade Count: (1)
  • Taperssection Member
  • ***
  • Posts: 519
  • Gender: Male
Re: To shrink or stretch?
« Reply #29 on: August 13, 2014, 08:08:59 PM »
I would just question the accuracy of those markers. Like I said, it's gotta be within +/- two samples. But who knows--it could do it. I might play around with it.
Audio: Countryman B3 + AT853(hypers/cards/subcards) + SBD feeds
Wireless Receivers: Lots of those
Antennas: Lots of those
Cables: Lots of those
Recorders: TE TX-6, Zoom L20R, Zoom F8, (3) Tascam 680, (3) Tascam 2D, Zoom H6, and a graveyard of irivers/nomads/minidiscs.

 

RSS | Mobile
Page created in 0.129 seconds with 44 queries.
© 2002-2024 Taperssection.com
Powered by SMF