Roll Eyes
Lips Sealed
123 ->
List results:
Search options:
Use \ before commas in usernames
I just streamed a single segment Final Fantasy X run this weekend, and I recorded the live commentary track in Audacity.  The thing is, the track is over 11 hours long, and when I go to export it as a .wav file, it only saves about 4 1/2 hours of it, sometimes less.

Does anyone know of any way to export an 11 hour wave file in Audacity, or will I have to split the commentary into several files?

Additonally, on a previous 11+ hour run that I encoded with anri-chan, it only processed the first 5 hours 20 seconds of audio, the rest of it had no sound. The original .avi has all the sound, so I'm not sure what's going on here.
Thread title:  
Edit history:
Exo: 2011-09-25 08:58:04 am
The wav file size is limited to 2gb.
I've never used audacity but I'm pretty sure you can encode with it so that would be your only option to save the whole thing as one file.
The 4 1/2 hour file that it saved was 6.7 GB though, so I'm not sure if it is cutting it off because of size or length in this case.
That's odd.. all I'm aware off is the 2gb limit which is cause by the 4 byte header :/
Edit history:
CMiller: 2011-09-25 11:17:49 am
CMiller: 2011-09-25 11:14:38 am
The limit is actually 4GB as per the spec given the 32 bit integer used to record the file size header, but some programs limit it to 2GB.
Fucking Weeaboo
There's no theoretical time limit, just a size limit.  Technically it's suppose to cut off at 4GB, which is some 6+ hrs at CD quality.  You can get the full 11 hours if you put it at a lower quality, like 22.05 kHz and/or 8-bit and/or mono.
ok, I suppose next time I do the run I will record the commentary at lower quality. It is doubtful that I will submit this run anyway, so I don't really need to save the commentary this time.

That just leaves the issue of anri-chan not processing all 11+ hours of game audio. Is that a similar problem, or something different?
Heavy Metal Powered
Have you tried to export it as MP3 320 kbit? You can always convert it back to WAV... The sound quality wont be as good as it was originally but since it's just commentary the difference would be hardly noticeable.
Balls jerky
If it's only spoken word you could go significantly lower than 320 and not notice a difference.
yeah, this is a huge pain in the ass. several workarounds have already been mentioned. the most "fun" comes from how vfw audio is wav internally so you can't normally have an avisynth script with audio longer than 6 hours (anri uses avisynth internally). however you may be able to hack the scripts to output 22050 hz audio to double the barrier to 12 hours. don't quote me on that though.
so here's another problem: even when I encode the video in 3 separate parts, the audio still cuts out at the same spot.  So I guess it's when anri reads the source AVI that it cuts the sound off, not just when I try to encode 11+ hours at once. The source AVI has all the sound though, when I watch it in a media player.

This is definitely a big problem, because I don't really want to record a single segment run in multiple parts and have people confuse it for a segmented run.  It is easy enough to tell that it is single segment since the cursor stays on memory the entire time, not resetting to the initial position, but I can't count on everyone noticing that.
like i said, pain in the ass. anri uses avisource() for avi input files by default which is vfw (afaik). you will have to use directshowsource() (which will probably cause audio desync) or one of the others (e.g. ffvideosource() coupled with a later audiodub(ffaudiosource())). as an alternative you might try segmenting the original file, then telling anri to append the segmented files. if avisynth is using vfw internally for avisource()++avisource() this will do nothing, but if it's using directshow (which it might) then it may solve your problem. good luck.
I suppose next time I record the run I will stop and start the recording at a black screen with no sound, that way there won't be any "suspicious" breaks when I append the videos in anri.  Obviously if I'm doing the run live, it is pretty clear that the run is single segment, but I just don't want there to be any discrepancies.  Such are the trials of running a game this long in one sitting Undecided
you should be able to split the original to lossless new master files and tell anri to append the files with zero frames lost and little to no perceptible break in the audio. however splitting them in the first place may be difficult given the same audio length problem.
Looks like that might work.  I haven't actually gone through with encoding the whole thing yet, but I split the original avi into 3 separate avis in VirtualDub, maintaining the same lossless compression, then appended those 3 avis in anri-chan.  When Vdub opens to check the top/bottom fields, I checked the audio, and it appears to have all the audio now.  So hopefully it will process all the audio when I encode it as well.
No good. Even though it reads all the audio from the source now, when it goes to write the .wav file after encoding the video, it runs into the .wav size problem and only encodes about half of the audio.

Not sure what other options there are now, other than encoding 2 or 3 separate files, which I really don't want to do for a single-segment run, but I may not have any other choice.  At least I wasn't planning on submitting this run, but I guess further attempts are on hold until a solution is found Sad
this is partially my fault for pushing snow to move to wav nmf instead of using bepipe. however bepipe has its own problems (e.g., iirc it is .net).

let me reassure you that i don't believe in things like "you are the only one ever to do this, therefore you shouldn't". i wonder what the prospects would be for patching anri.bat to pipe the script through bepipe (expecting it to be in its program files folder) instead of making wav nmf. then we could make that patch available somewhere less obvious so that people wouldn't accidentally use it. an alternate page in the kb would probably suffice.

a problem is that i'm not sure whether piping an avisynth script through bepipe is vfw or directshow. if it's the former then you will end up with cut off audio or no audio. anyway i'm going to look at this more when i get home tonight.
Yes, a cucco riding the ground.
I wouldn't let this problem stop you from doing attempts. Other single-segment runs have been split into multiple parts due to technical constraints, such as Marshmallow's Jet Force Gemini run. As long as you make it clear that your run is single-segment, there won't be any confusion.
Edit history:
nate: 2011-09-28 04:06:07 pm
ok, make a backup of your existing c:\program files\anrichan_whatever\anri.bat and try this anri.bat instead. turns out bepipe was still in there, just only activated if there was an "error" with wav dump encoding, which wasn't detected in your case. note that bepipe iirc requires .net. you may already have it installed. if not you may see an error when it tries to run. just google .net framework to find the download from microsoft.
Thanks nate, I'll give that a try.  I'm actually testing encoding with anri 3.2 to see if that works, but if I can get it to work in anri 3.3, even better. I'm pretty sure I have .net installed already.
hmm, now im getting this

That was fixed in 3.2 wasn't it?
bleh ... snow will have to weigh in.
Apparantly bepipe doesn't like it either.  This was done in anri 3.2. Only processed 18.5k out of 40k+ seconds of audio.

you can see there was some kind of error there that was covered up by the bepipe % output ... wonder what it was and what's going on ...
Yeah, I missed what the error was; I just came back and it had stopped at 18445 seconds, but the % was still rolling.