Sync Test: CX, H4n, H2, irivers at DVinfo.net

Tim Akin · February 6th, 2010, 11:18 AM

This is not a scientific test by any means, but it does show how each recording device syncs up with the camera audio. All devices were set up as close together as possible. I let the camera (FX1000) and recorders record for 30 minutes with a clap at the beginning and another at the end. The CX, H4n and H2 were all set to PCM 48kHz/16 and the irivers (790 and 795) set to there best 44.1kHz/160KBPS. I did convert the 790 mp3 to 48kHz wav file using Vegas, but the mp3 and wav were exactly the same.

Results: 1st - Yamaha Pocketrak CX - ½ frame to fast
2nd - H4n - 1 fast
3rd - 790 - 1 ½ slow
4th - H2 - 6 ¾ slow
5th - 795 - off the chart

Richard Crowley · February 6th, 2010, 11:53 AM

Those results confirm what I had concluded:

1) Asynchronous (not interconnected for sync) modern devices are pretty good and drift only a few frames over 10-20-30 minutes. Certainly good enough for most any normal production.
The numbers reported in this test are general indications. Of course a different set of exactly the same models would have yielded slightly different results, but all in the same ballpark.

2) Any kind of compressed recording is not really useful for serious production.

Steve House · February 6th, 2010, 12:16 PM

Quote:

1) Asynchronous (not interconnected for sync) modern devices are pretty good and drift only a few frames over 10-20-30 minutes. Certainly good enough for most any normal production.
The numbers reported in this test are general indications. Of course a different set of exactly the same models would have yielded slightly different results, but all in the same ballpark.

2) Any kind of compressed recording is not really useful for serious production.

Certainly agree with your latter point but not so sure about the former. Network specs allow for zero tolerance for sound to lead picture and a maximum of 1/4 frame for sound to lag picture for the entire duration of the program. For a show made up film-style of a series of short takes it's not too hard to achieve but for long-form concert footage etc the reported results would make for major headaches in post. Even 1 frame drift in 60 minutes exceeds allowable tolerance by a factor of 4 and would require manual correction. (source - PBS Producer's Redbook technical specs).

Richard Crowley · February 6th, 2010, 01:01 PM

Steve, we are talking about production recording here, not final edited product. You are absolutely correct that even 1 whole frame error is not acceptable in the final product. However nobody would create a production with an unbroken 10-minute shot of anything, at least nothing that was commercially viable. So, as you say, any place where you make any kind of video edit is an excellent opportunity to "pull-up" whatever sync error has accumulated. Furthermore, long shots, etc. are less critical for exact frame-sync. Not to mention that you can often get away with diddling the video to match the audio. Repeating or dropping single frames is not really detectable by most viewers.

Renton Maclachlan · February 7th, 2010, 01:23 AM

I have a number of iriver 890's and some sync pretty good and a couple don't. Have also experienced speed difference throughout a track...

February 6th, 2010, 11:18 AM	#1
Tim Akin Major Player Join Date: Nov 2008 Location: Newbern, TN Posts: 414	Sync Test: CX, H4n, H2, irivers This is not a scientific test by any means, but it does show how each recording device syncs up with the camera audio. All devices were set up as close together as possible. I let the camera (FX1000) and recorders record for 30 minutes with a clap at the beginning and another at the end. The CX, H4n and H2 were all set to PCM 48kHz/16 and the irivers (790 and 795) set to there best 44.1kHz/160KBPS. I did convert the 790 mp3 to 48kHz wav file using Vegas, but the mp3 and wav were exactly the same. Results: 1st - Yamaha Pocketrak CX - ½ frame to fast 2nd - H4n - 1 fast 3rd - 790 - 1 ½ slow 4th - H2 - 6 ¾ slow 5th - 795 - off the chart Attached Thumbnails __________________ Tim

February 6th, 2010, 11:53 AM	#2
Richard Crowley Trustee Join Date: Aug 2007 Location: Portland, Oregon Posts: 1,238	Thanks for the test report Those results confirm what I had concluded: 1) Asynchronous (not interconnected for sync) modern devices are pretty good and drift only a few frames over 10-20-30 minutes. Certainly good enough for most any normal production. The numbers reported in this test are general indications. Of course a different set of exactly the same models would have yielded slightly different results, but all in the same ballpark. 2) Any kind of compressed recording is not really useful for serious production.

February 6th, 2010, 01:01 PM	#4
Richard Crowley Trustee Join Date: Aug 2007 Location: Portland, Oregon Posts: 1,238	Steve, we are talking about production recording here, not final edited product. You are absolutely correct that even 1 whole frame error is not acceptable in the final product. However nobody would create a production with an unbroken 10-minute shot of anything, at least nothing that was commercially viable. So, as you say, any place where you make any kind of video edit is an excellent opportunity to "pull-up" whatever sync error has accumulated. Furthermore, long shots, etc. are less critical for exact frame-sync. Not to mention that you can often get away with diddling the video to match the audio. Repeating or dropping single frames is not really detectable by most viewers.

February 7th, 2010, 01:23 AM	#5
Renton Maclachlan Trustee Join Date: Jul 2007 Location: New Zealand Posts: 1,180	I have a number of iriver 890's and some sync pretty good and a couple don't. Have also experienced speed difference throughout a track...