PDA

View Full Version : Overlapping content when splitting



caldernet
06-05-2010, 02:08 PM
I split the radio shows I record in 2 minute long files.

After every split, the next file contains about 20 seconds of the end of the previous file.

Is there a way to eliminate this or change the length of overlap?

Thanks.

Cheryl Wester
06-06-2010, 04:54 PM
This is due to buffering. The only solution is to use audio capture. The amount of buffering is dependent on your Internet connection and the server on the other end.

JonQ
06-18-2010, 07:53 AM
I have to assume that this technically possible, but just not a capability of the current version of the software. The input stream is kept in buffer until written out. If chunks of that stream were popped off to a secondary write buffer for each output file, then there's no reason that portions of that content should be duplicated across 2 files.

Eg:
Input Buffer 1-second Chunks: abcdefghijk

Output File 01.snd buffer: abcde

When the split occurs, a second buffer is created
Output File 02.snd buffer: fghijk

When the next split occurs you just re-use the old buffer.

This should work fine. You could then add the option to allow a pre-determined amount of overlap if someone actually wanted that. In that case you could fork n-seconds of chunks into both buffers:

Output File 01.snd buffer: abcDE
Output File 02.snd buffer: DEfghijk

I propose this as an enhancement for a future version. I really need to split multi-hour recordings, but the overlap makes this impractical since I can play the split files back-to-back in a seamless fashion. I don't want to do audio capture due to the loss of quality.

Cheryl Wester
06-18-2010, 10:13 AM
Thanks for the feedback. Our developers do read the forums all the time so perhaps.

stream-recorder.com
06-18-2010, 10:43 PM
I have to assume that this technically possible, but just not a capability of the current version of the software. The input stream is kept in buffer until written out. If chunks of that stream were popped off to a secondary write buffer for each output file, then there's no reason that portions of that content should be duplicated across 2 files.

Eg:
Input Buffer 1-second Chunks: abcdefghijk

Output File 01.snd buffer: abcde

When the split occurs, a second buffer is created
Output File 02.snd buffer: fghijk

When the next split occurs you just re-use the old buffer.

This should work fine. You could then add the option to allow a pre-determined amount of overlap if someone actually wanted that. In that case you could fork n-seconds of chunks into both buffers:

Output File 01.snd buffer: abcDE
Output File 02.snd buffer: DEfghijk

I propose this as an enhancement for a future version. I really need to split multi-hour recordings, but the overlap makes this impractical since I can play the split files back-to-back in a seamless fashion. I don't want to do audio capture due to the loss of quality.
If I'm not mistaken, Replay Music makes a split when the volume is too low. Another problem is that splitting will never be perfect in case of doing it automatically. And you don't need to re-record. You can use lossless MP3 splitting software instead.

If anybody else needs auto splitting, please post here. As Cheryl mentioned, the developers do read the forums.