Timing shift

[Ebuechner]

I'm having a problem with the timing while using version 4.2.2 5. If I'm sequencing and playing the song from beginning to end no problem. But if I select a point to start from in the middle or towards the end of the song I can watch the time in Decatur jump ahead and then I'm off on timing. While doing this it'll even cut the last split second off of the song and not play it. I have tried the audio with 128 and 192 constant bit rate

[Gilrock]

Yeah I've seen strange things like this also now that I'm more heavily into sequencing.  I've noticed when I start in the middle the line likes to do a quick jump within the first second so what I do is start it back a little earlier so that the part I'm sequencing isn't near the jump area.

[Ebuechner]

That's exactly it.  It seems to be proportional to the position that you are in the song. I've tried a couple of different computers  same issue every time.

[rcowan]

I'm seeing this same issue with 4.3.02. I think it becomes more evident with longer songs. My audio file is a medley at just under 12 minutes long. Has anyone found a way around this or is this something that is going to be fixed? My faces look like they're dubbing a foreign film.

[Lights On Fifth]

I had to to try this myself and there is definitely a lag, this is a sequence that I have used with Xlights since 2012 it now shows lag issues in the vis


 update after I re-saved my sequence everything works perfect I can click anywhere on the time line and no lag issues with any mouth movements or any models so try hitting save sequence then re-open and see if that heps

[rcowan]

update after I re-saved my sequence everything works perfect I can click anywhere on the time line and no lag issues with any mouth movements or any models so try hitting save sequence then re-open and see if that heps

Thanks Lights on Fifth,

I'll try again but I tried everything; saving, rendering, closing the app and reopening. Nothing seemed to make a difference.

[rcowan]

Not sure what else to try. I've recreated the file from Audacity, making sure that it is 192 Kbps and CBR. I've saved, rendered, closed the program, opened the sequence again. Nothing fixes the issue. It is definitely exponential. I noticed that towards the end of the sequence it can jump by almost a full second. Whereas closer to the beginning it might be less than 1/2 a second. It's hard to see but if I slow the audio down to 1/4 speed then click on the timeline to start playing the sequence you can see it jump.

[Phrog30]

Not sure what else to try. I've recreated the file from Audacity, making sure that it is 192 Kbps and CBR. I've saved, rendered, closed the program, opened the sequence again. Nothing fixes the issue. It is definitely exponential. I noticed that towards the end of the sequence it can jump by almost a full second. Whereas closer to the beginning it might be less than 1/2 a second. It's hard to see but if I slow the audio down to 1/4 speed then click on the timeline to start playing the sequence you can see it jump.

Try 128kbps. I remember Gil saying he had a song that caused issues and he found 128 to work, but 192 didn't. Not sure if I'm remembering the thread correctly, but worth a shot.

James

Sent from my SM-G900V using Tapatalk

[rcowan]

Not sure what else to try

Actually I found something else to try. I used a different audio file. One that I had laying around but that was about 12 minutes long. This file has not been through Audacity. So, I created a new sequence and played the file. No issues. There were no jumps at any point in the audio. I then took that MP3 file, opened it in Audacity and then exported it. Copied the exported file over the one that my sequence was using and bam, it starts jumping. I tried a couple of different bitrates (all CBR of course) but it didn't make a difference. So, there seems to be an issue with the MP3 encoder that I'm using with Audacity. I'm using LAME 3.99.3.

I'm going to attach the test MP3 file I was using. If someone can just verify what I'm seeing I'd appreciate it.

Just create a new sequence with the MP3 file. Don't set anything else up such as timings, effects, models, etc. Zoom in some on the timeline so that you can see every 2 seconds. Scroll over to about the 8 minute mark and then click on the timeline to start the sequence. It should play smoothly as expected.

Now, take that MP3 file and import it into Audacity. Then turn right around and export it to a MP3. Follow the same steps above for creating the sequence and testing. You should notice that after about a second of it playing that the timing bar jumps about a second.

Thanks in advance to anyone who decides to test this for me.

P.S. The MP3 file is about 27 MB. I'm not sure if it will allow me to attach it. In case it doesn't, it can be downloaded here: http://incompetech.com/music/royalty-free/mp3-royaltyfree/Perspectives.mp3

[rcowan]

Try 128kbps. I remember Gil saying he had a song that caused issues and he found 128 to work, but 192 didn't. Not sure if I'm remembering the thread correctly, but worth a shot.

Thanks James. I'm willing to try anything. Unfortunately, I just tried that and it didn't make a difference.

[Phrog30]

Have you tried another machine?

Sent from my SM-G900V using Tapatalk

[Gilrock]

The question is whether its in the right location after that initial jump.  If so then just start your marker earlier and don't use its position till after the jump.  It's not going to get fixed unless we do a complete revamp and switch audio libraries.

[rcowan]

The question is whether its in the right location after that initial jump.

Unfortunately it's not. If I watch the waveform and compare it to what I'm hearing it's definitely out of sync. However, as I posted a little bit earlier, as a test, I took a long MP3 file that I downloaded and created a new sequence with that. There were no issues. I then ran that file through Audacity and then exported it. I didn't actually do anything with it in Audacity. I just wanted to test a theory. Sure enough, when I ran the Audacified (my new word) MP3 file through xLights it was having the shift issue. So my current theory is that something in the Audacity MP3 encoder (LAME) is messing up the file. You wouldn't know it hear it. I was hoping that someone would take the file that I posted and go through the same steps to see if it happens for them or not. If it can't be reproduced then I can look at environmental issues on my side.

In the meantime, I'm splitting my large 12 minute sequence file into smaller sequences. I have 3 main songs in my show and then some filler in between them. I figured sequencing those individually would probably be better in the long run so that I can easily rearrange the show if I want to.

By the way, I'm sure you've been able to glean from my explanations what I'm talking about but for anyone else who wants to see what I'm talking about, see this video that Sean did for Gil about 10 months ago. https://vimeo.com/121568187 That's exactly what it's doing. However, I'm assuming that this bug was fixed since it doesn't happen at the beginning of the sequence. It's not until you get several minutes in that you start to notice it. Also, it seems to be exponential. The further you get from the beginning of the song the longer the time shift is.

[DMJPixel]

I will try your file later today and report findings.


Sent from my iPhone using Tapatalk

[DMJPixel]

I tried your suggestion this morning on my MAC running xlights v4.3.02.  I followed the steps as you described (exported as 160 CBR) and did not see any issues.  I played the file from about the 7:58-8:02 mark for at least 60 seconds.

I will try on my Win10 PC later today and report.
Dave