Taperssection.com
Gear / Technical Help => Recording Gear => Topic started by: taperdave on March 30, 2013, 07:04:39 PM
-
I am trying to verify something disturbing about the Edirol, but don't own one
I need someone impartial to record 24/96 or anything identically on a M-10 and edirol.
I believe the Edirol loses about a 10th of a second or less as it makes the file switch on long continuous recordings, while the M-10 is gapless/seamless.
any independent verification appreciated, makes audio matrix with dissimilar devices a.....
thanks
Dave
-
Not sure which 'Edirol' you are referring to. I have used the R-09HR and the R-26 and both in my experience have been completely gapless in runninng from 1 x 2GB file to another.
-
makes audio matrix with dissimilar devices a.....
Well, different devices = different clocks = time adjusting anyways, so... I don't think that 1/10th of a second matters much - unless I'm missing something?
Also, my experience with the Roland R-05 is a perfect gapless split. The few times I have reached the 2GB limit and the gap has fallen in the middle of a song, I haven't noticed anything - but my ears can deceive me of course.
-
Yes, I think also the Edirol makes split quite well. It's hard to find an example because I also rarely tape 3hrs/2gb without stop. I think the missing part can be counted well only if you can compare to an other source without cut on that point. I checked, its really just 1/10 sec:
(http://img850.imageshack.us/img850/3650/edirolr09hrautosplit.jpg)
-
OK thanks guys this confirms what I thought I had figured out in syncing audio for long projects.
so far all of the M-10 sourced were actually gapless and the 09HR has about a 10th of a second gap at each file split so a three hour show in hi res has about three gaps where the matrix sounds really goofy, unless you fix the gaps
Also I haven't had to do any time stretching as long as I correctly align each new 09HR file at the split, way less hassle than trying to fix a drift that gets worse as the show progresses (ie impossible as the stretch amount needs to change after each gap)
glad I am not crazy
Dave
-
Good catch, taperdave.
-
Also I haven't had to do any time stretching as long as I correctly align each new 09HR file at the split, way less hassle than trying to fix a drift that gets worse as the show progresses (ie impossible as the stretch amount needs to change after each gap)
But do your other recording device (soundboard, other portable recorder or whatever) run at the exact same speed than the R09HR? Then I think you've been pretty lucky! :)
-
Yes, I think also the Edirol makes split quite well. It's hard to find an example because I also rarely tape 3hrs/2gb without stop. I think the missing part can be counted well only if you can compare to an other source without cut on that point. I checked, its really just 1/10 sec:
(http://img850.imageshack.us/img850/3650/edirolr09hrautosplit.jpg)
I'm sorry... what am I missing here? There's nothing about those two waveforms that resembles one-another.
-
Yes, I think also the Edirol makes split quite well. It's hard to find an example because I also rarely tape 3hrs/2gb without stop. I think the missing part can be counted well only if you can compare to an other source without cut on that point. I checked, its really just 1/10 sec:
I'm sorry... what am I missing here? There's nothing about those two waveforms that resembles one-another.
Its the same sound. One is aud, one is sbd. Difference is probably because of big crowd noise and maybe the beginning of a song with low vol and even you can see only very very short part on the timeline (and even a badly mixed sbd can look quite different). But those are synched, so I just wanted to show it is really short cut, what's missing when Edirol starts a new file.
taperdave: if all the problem is that 10th sec is missing not a big deal to synch at every new file, I guess. It's worse if the 2 sources are not recorded with the same speed, but in that case no matter if there are cuts or not, needs to use time warp anyway. I need to change speed in case of iriver and Microtrack usually :(
-
taperdave: if all the problem is that 10th sec is missing not a big deal to synch at every new file, I guess. It's worse if the 2 sources are not recorded with the same speed, but in that case no matter if there are cuts or not, needs to use time warp anyway. I need to change speed in case of iriver and Microtrack usually :(
that's basically what I meant in my 2 prev posts
-
I never saw or heard any cut at the file crossover with my R09HR. And I had plenty of sets cross the 2GB mark. If it happened at crowd noise, it might not be noticeable, but if during music it most certainly would have. I don't know what makes you think there is a .10 second gap other than the R09HR needs to be stretched to match the other source. A tenth of a second drift to a second source would not be unusual. I don't know why you aren't joining the files prior to trying to do any alignment.
Sorry, I just don't believe there is a gap at the 2GB split. It would be visually noticeable when the files are joined and audibly as well. I think you need to rethink your approach here if you are getting a gap when you are aligning them with another source.
-
OK here is the sitch
I have done this type of project several times through several iterations of the software, I am slow and a bit OCD, my own recordings have been from different cams as well as MT2 and then the M-10.
The recordings I have gotten from others are a variety of cams and the Edirol and M-10 for audio (normally 24/48 or 24/96).
So here is how the project goes
multicam HD with matrix audio, generally I master the audio separately in Wavelab and then do the assembly and multicam edit in Vegas
everything starts out well but by the middle of the 3 hour show the audio is not still in sync and by isolating each pair of audio tracks against each other I find the Edirol is a tiny bit shorter (originally I thought it was a little bit faster).
I tried to time stretch the Edirol recordings but really I suck at that and wasn't able to get a satisfactory stretch over the 3 hour set that made everything match up and I tried on one particular piece a total of 14 times before giving up that method.
I eventually inserted tiny pieces of silence in the Edirol recording at a few places through the show, but again it was better but not "right".
Honestly it didn't really occur to me until recently to spend more time and look at alignment of each segment against each other on both sides for alignment. I knew that most HD video has small audio drops at the splits and discovered that my Panny drops three audio frames while maintaining the video at each split. Vegas automatically joins up the different files "seamlessly" so it just didn't occur to me that the R09HR didn't behave like the M-10. when listening to a L/R comp of the 2 nicest sources of the most recent project I realized that everything was luscious and yummy till right after the first Edirol split where some nice concussive snare hits were out of sync. BLAMMO problem solved, it dropped about 3.5 frames at each split. I listened to the audio without any other source and I could hear the tiny problem, but asked some family members to listen to the segment and asked them if they heard a problem and they said "no". It's short enough to sound seamless by itself unless you already know there is an issue.
I'll post up some audio evidence later tonight, busy day today
Dave
-
I never saw or heard any cut at the file crossover with my R09HR. And I had plenty of sets cross the 2GB mark. If it happened at crowd noise, it might not be noticeable, but if during music it most certainly would have. I don't know what makes you think there is a .10 second gap other than the R09HR needs to be stretched to match the other source. A tenth of a second drift to a second source would not be unusual. I don't know why you aren't joining the files prior to trying to do any alignment.
Sorry, I just don't believe there is a gap at the 2GB split. It would be visually noticeable when the files are joined and audibly as well. I think you need to rethink your approach here if you are getting a gap when you are aligning them with another source.
^^ this
I also ran loads of shows on an R-09HR with 2gB file splits and never had any gap, 1/10 sec would have been noticeable in the music for sure.
-
Maybe it's a fault on your R09HR?
-
I never saw or heard any cut at the file crossover with my R09HR. And I had plenty of sets cross the 2GB mark. If it happened at crowd noise, it might not be noticeable, but if during music it most certainly would have. I don't know what makes you think there is a .10 second gap other than the R09HR needs to be stretched to match the other source. A tenth of a second drift to a second source would not be unusual. I don't know why you aren't joining the files prior to trying to do any alignment.
Sorry, I just don't believe there is a gap at the 2GB split. It would be visually noticeable when the files are joined and audibly as well. I think you need to rethink your approach here if you are getting a gap when you are aligning them with another source.
^^ this
I also ran loads of shows on an R-09HR with 2gB file splits and never had any gap, 1/10 sec would have been noticeable in the music for sure.
Would it?
I had a bad optical cable once that was dropping samples...my matrix would be pefect and then suddenly lose sync...not drift - sudden loss of sync.
But you would never catch it listening to each source independently.
-
everything starts out well but by the middle of the 3 hour show the audio is not still in sync and by isolating each pair of audio tracks against each other I find the Edirol is a tiny bit shorter (originally I thought it was a little bit faster).
I tried to time stretch the Edirol recordings but really I suck at that and wasn't able to get a satisfactory stretch over the 3 hour set that made everything match up and I tried on one particular piece a total of 14 times before giving up that method.
Have you checked not only the beginning but the end of the first file of Edirol? Is it synched well? If yes, no stretch needs, only needs to resynch the 2nd Edirol file and you'll have a little missing part, but your whole stuff shall be ok. If its out of synch even before the first cut, then needs to change speed.
-
Someone correct me if I'm wrong, but I don't think any two separate recorders will ever record the exact same file length. They may be very close, but never exact.
Is that just what you're wondering about? If you use separate recorders you will generally have to do some time stretching/shrinking of one of the files to line up separate sources of the same show. Some people have claimed to be lucky and only had minor drift after a hour or more, but that has never been my experience. Especially after a three hour show, I find it hard to believe two separate recorders would be recording in sync by the end, with only a dropout to create a new file causing the latency.
-
I own an R-4, an R-09, and an R-09HR. and have never had anything but seamless file transitions with any of them.
-
Someone correct me if I'm wrong, but I don't think any two separate recorders will ever record the exact same file length. They may be very close, but never exact.
Is that just what you're wondering about? If you use separate recorders you will generally have to do some time stretching/shrinking of one of the files to line up separate sources of the same show. Some people have claimed to be lucky and only had minor drift after a hour or more, but that has never been my experience. Especially after a three hour show, I find it hard to believe two separate recorders would be recording in sync by the end, with only a dropout to create a new file causing the latency.
Yes, that's what I meant in my previous posts - no one seemed to acknowledge it anyway :P But I think you're very much right, sir, the device's clocks very rarely run at the exact same speed. Speed difference would be HARDLY noticeable during a 10 second or even a song' length, but when the files get longer (in the hours range) I think it's virtually impossible to have the exact same timing - one source or the other will need to be stretched, more or less, but stretched after all.