Fine tuning - checking individual phrases - compared to listening to whole chapter

I’m fairly new using the SAB. I try to build SA for whole NTs in Tanzania and adding audio files to the text and having the text read highlighted gives me quite some headache.

After creating the timing files and going to fine tune, it seems to work fine when listening and checking from the beginning of the chapter. The timing of the highlight of the phrase being read is fairly good. But starting the finetuning in the middle of the chapter, or just listening to a random phrase the timing is off. The longer the chapter the worse it gets towards the end. The highlighting is up to 3 and more seconds behind the audio. For an example, when listening thru John 6 from the beginning the audio and highlighting goes together nicely. But clicking (highlighting) on verse 60b in the fine tuning the words being read are the words of the middle of verse 60c.
Therefore the finetuning is a very tedious task and takes a long time.

What do I do wrong? I use MP3 files, (22050Hz, 32bit (float), 32kbps bit rate) and I’m using aeneas for synchronizing the audio and text. Is there something I miss?

Thanks for your help.


I am assuming your MP3 files are Constant Bit rate if you are using SAB 8.0. SAB 8.0 checks if MP3 files are constant bit rate. But earlier versions did not.

The other thing that may cause issues is the 32bit (float). Most people would only use 16bit. It may be that the libraries used are expecting a 16bit (or perhaps 24bit).

Since your other specs are low 22050Hz or very low 32kbps it looks very odd to have the highest spec sampling.