Helpful ReplyPlayback slower than recorded

Author
DThompson55
Max Output Level: -90 dBFS
  • Total Posts : 2
  • Joined: 2015/03/01 14:52:25
  • Status: offline
2015/03/01 15:23:06 (permalink)

Playback slower than recorded

Hi - This is weird.  We recorded our band a couple weeks ago using Sonar on Windows 7 PC, everything sounded great.  We came back today to do overdubs, and everything is playing back slower and at a lower frequency.  So what used to be a low A at 110hz is playing back at around 101hz, somewhere between G and G#.  Songs that were about 3:00 minutes are stretching out to almost 20 seconds longer. Has anyone run into this before?

Another problem is that the vocals might have been recorded at that slower playback.  I imagine they'll have to be tossed.
#1
slartabartfast
Max Output Level: -22.5 dBFS
  • Total Posts : 5289
  • Joined: 2005/10/30 01:38:34
  • Status: offline
Re: Playback slower than recorded 2015/03/01 16:50:27 (permalink) ☄ Helpfulby DThompson55 2015/03/01 18:21:51
Classic symptom of a sampling rate mismatch, i.e. recording at a different sampling rate than the sampling rate set for the project on playback. It has the same effect as changing the record and playback of tape speed in analog recording. 
 
The vocals can be salvaged by resampling which is pretty easily done externally using something like R8Brain on the audio.
 
http://www.voxengo.com/product/r8brain/
post edited by slartabartfast - 2015/03/01 17:29:08
#2
DThompson55
Max Output Level: -90 dBFS
  • Total Posts : 2
  • Joined: 2015/03/01 14:52:25
  • Status: offline
Re: Playback slower than recorded 2015/03/01 23:22:40 (permalink)
I'd like to believe it is a sampling rate mismatch, but since only one sampling rate is allowed per project and the it appears that the sampling rate is pretty difficult to change after you've started a project, I'm guessing it is something other than sampling rate.  It's set to 44100.  How can I check to be sure that it's a sampling rate problem?  I mean, we all heard the tracks just fine at the end of the first day of recording.
 
post edited by DThompson55 - 2015/03/02 00:01:40
#3
slartabartfast
Max Output Level: -22.5 dBFS
  • Total Posts : 5289
  • Joined: 2005/10/30 01:38:34
  • Status: offline
Re: Playback slower than recorded 2015/03/02 02:02:22 (permalink)
Not all audio interfaces set their own A/D sampling rate based on the rate set in the DAW software. You need to be sure that the interface and the DAW are both using the same SR or this problem can occur even if you do not move the audio from another project or session into the current one. Check the audio interface manual and the settings in the interface if it is an option to set a rate there. The rough pitch/timing differences you mention would suggest a 44.1 vs 48 KHz mismatch which is the most common problem pair, since both are "standard" rates and close enough to be confused.
 
http://www.emusician.com/...le-rate-mismatch/34080
post edited by slartabartfast - 2015/03/02 02:15:13
#4
Anderton
Max Output Level: 0 dBFS
  • Total Posts : 14070
  • Joined: 2003/11/06 14:02:03
  • Status: offline
Re: Playback slower than recorded 2015/03/02 09:31:44 (permalink)
I'd bet on slartabartfast's reply.

The first 3 books in "The Musician's Guide to Home Recording" series are available from Hal Leonard and http://www.reverb.com. Listen to my music on http://www.YouTube.com/thecraiganderton, and visit http://www.craiganderton.com. Thanks!
#5
Cactus Music
Max Output Level: 0 dBFS
  • Total Posts : 8424
  • Joined: 2004/02/09 21:34:04
  • Status: offline
Re: Playback slower than recorded 2015/03/02 10:06:23 (permalink)
I think if you open the Audio folder and check the properties of the wave files you'll find the answer. 

Johnny V  
Cakelab  
Focusrite 6i61st - Tascam us1641. 
3 Desktops and 3 Laptops W7 and W10
 http://www.cactusmusic.ca/
 
 
#6
Sir Les
Max Output Level: -67 dBFS
  • Total Posts : 1182
  • Joined: 2012/07/09 04:56:19
  • Location: MONTREAL, QUEBEC, CANADA,
  • Status: offline
Re: Playback slower than recorded 2015/03/02 10:19:30 (permalink)
yep Star is right...usual cause for the change up in bit rates, is another app using same driver /audio device, playing back different bitrate content, or recording using different settings, and left those setting be, as the one you are seeing the issues with now being mismatched...because of.
 
(using audio device as OS default, and letting windows sounds play through usually the culprit for changing those settings)(or just using as default to play music already mastered, can change the settings of the audio device from  48khz to 44.1khz or other)...opening the project with the device not set right before hand for record and play back, cause the issue you mention....as the audio in the project,has to be converted up or down now to play back through the mismatched settings of the audio device...
 
Just have to know what was set when making the project, and set the audio device app or mixer app setup to match those settings...before opening the project...you can right click a wav from the project, and get properties , find bitrates and other info...and use it to set the audio app to.
 
might work if you use Sonar audio device preferences /open audio device panel, but I think some audio apps get locked when a DAW is using the audio device.(some allow DAWs to changes, some do not and get locked)..So, remember what is set per project, and make sure you open them projects with the audio device setup for them before the project is opened, for normal playback and timing/sync....if you cannot change bitrate through the DAW preferences audio tab, and opening your audio app in that way to change bitrate to match.
post edited by Sir Les - 2015/03/02 10:37:11

1. Intel 5960x 3.5mhz , ASUS x99 deluxe u3.1, Asus Thunderbolt ex II,   G skills f4 3000 Memory 32GB , ADATA ssd 250GB Main Drive, Lots of WD Red 7200 Mechanical Drives with Black Drives, 14x multi optical Drive, LG Multi Blu Drive,  2X Extern WD Mybooks usb 3.0, AMD r7 270 video card, Motu 828x TB , Motu Midi XT.
2.  USING MAC PRO, as win 10 has damaged 2 x99 systems 8.1 is also to blame for the final burnout trying to roll back!
 
3.  Something Wonderful: https://1drv.ms/f/s!AlHkRy9cXBbYpQNvVBCt8r7fQ5PS
#7
THambrecht
Max Output Level: -73 dBFS
  • Total Posts : 867
  • Joined: 2010/12/10 06:42:03
  • Location: Germany
  • Status: offline
Re: Playback slower than recorded 2015/03/02 13:37:42 (permalink)
Sometimes (only sometimes) that will help during recording and playback:
Preferences ---> Sync and Caching ---> Trigger freewheel
 
Check your recorded wav-files. Play them outside of Sonar.

We digitize tapes, vinyl, dat, md ... in broadcast and studio quality for publishers, public institutions and individuals.
4 x Intel Quad-CPU, 4GHz Sonar Platinum (Windows 10 - 64Bit) and 14 computers for recording tapes, vinyl ...

4 x RME Fireface 800, 2 x Roland Octa Capture and 4 x Roland Quad Capture, Focusrite .... Studer A80, RP99, EMT948 ...

(Germany)  http://www.hambrecht.de
#8
slartabartfast
Max Output Level: -22.5 dBFS
  • Total Posts : 5289
  • Joined: 2005/10/30 01:38:34
  • Status: offline
Re: Playback slower than recorded 2015/03/03 11:56:24 (permalink)
In case the OP is still interested, I am finding the post by Sir Les a little confusing. The reason you have a discordance between your audio interface and your Sonar project is likely, as he says, due to something re-setting the sample rate for the audio interface and its driver. But I am unclear what the point of getting the properties from the wave file would be. 
 
Say your audio interface is set to collect samples at a rate of 48K samples per second. In sampling 1 second of real world sound it will deliver 48000 samples to the audio buffer somewhere in computer memory. Sonar has no way of knowing what the sampling rate in the interface is set to, it is just going to the buffer and picking up a series of numbers (amplitudes) from the buffer. Internally it represents this data as 4,8000 different sequential numbers. Now if the project in Sonar is set to 44.1K samples/sec when you listen to play back of the audio track in Sonar, it is going to be delivered to the D/A playback device at a rate of 44,100 samples/sec. At the end of one second of playback at 44100 samples/sec there will be 3,900 samples left to play, so to account for the full 48,000 samples it will take Sonar about 1.088 seconds. The frequency represented by the rate of change of sequential samples will drop as the rate at which the samples are delivered to D/A output device gets longer as well. So you have a recording that plays back lower and longer than the real world performance.
 
When Sonar records the sound it includes the requirement that it is to be played back at 44.1K sample rate in the recorded file. In a wave file this sample rate is in a specific location in the file, and it is that header information that Windows and audio devices look for to tell them how many samples to deliver per second, and what they will report as the sample rate in the properties window. When you go to look at the properties of the file, you will find that it says it was recorded at 44.1K, even though it was sampled at 48K and playback at 44.1K results in the longer and lower error. The samples in this mislabeled 44.1K file are in fact exactly accurate, and if your playback device were to play it back at 48K instead the sound would be recovered in a perfect state. But your playback device reads the erroneous sample rate header and plays it wrong.
 
If Sonar were asked to import a properly identified 48K wave file into a project set for 44.1K it would automatically resample the file to match the 44.1K project sample rate. Resampling is something of a misnomer, since the original sound is no longer available to be sampled. What happens is that a mathematical algorithm calculates how the sound represented by the 48K data should be represented by 44.1K data and makes a substitution. That "resampled" data is not necessarily exactly identical, but given a good enough algorithm and close enough sample rates it is indistinguishable in practice.
 
What you would ideally like to do is to tell Sonar that the file that was sampled at 48K but incorrectly labeled as needing to be interpreted as having been sampled at 44.1K, should be considered to be a 48K. I do not know of any way to do that within Sonar, but given the frequency of this problem, perhaps it should be a feature request. An elegant kludge would be to export the affected tracks to wave files and then use a hex editor to change the bytes in the file that specify the sample rate. Without changing a single sample, the problem is solved, and when Sonar imports this new, correctly identified file into an audio track, automatic resampling would result in a track that would play as expected. There are several wave file editing utilities available that will allow you to do that without delving deeply into the wave file format specification. I have not used them, but this might be a simple option if you plan to go that route. http://www.railjonrogut.com/HeaderInvestigator.htm
 
If you just use a time stretch/shrink process to get the track to the correct length, you will still have the pitch problem. Time and pitch changing algorithms are designed to change the length without changing the pitch or vice versa to enable two recorded performances of slightly different speed or tuning to be fit together. Applying time correction followed by pitch correction is an option but introduces much more opportunity for loss of fidelity.
 
As Sir Les says the best solution is to be diligent in observing the actual settings in Sonar and audio interfaces to be sure they match before the problem occurs.
#9
Jump to:
© 2025 APG vNext Commercial Version 5.1