Improve ETI frame detection - eti-stuff is even better

[andimik]

A user has posted a very weak raw file (please rename it to raw) which is indeed a ghost frequency of 12C on 44 MHz, but the strange thing is, when you compare eti-stuff and qt-dab in parallel (see screenshot), you will see 11 services found in eti-frontend, whereas Qt-DAB finds just 6 services.

AlbrechtL/welle.io#29 (comment)

[JvanKatwijk]

There are architectural differences between eti-frontend and qt-dab. To balance the load on the rpi (and to prevent an overload of one of the cores), some computations were partitioned in different threads. So, two threads are involved in computing the synchronization and it might well be the case that som triggering information arrives late and is discarded. There is no doubt that handling the whole synchronixation in a single thread is better (and easier to implement), this is a living proof. jan 2017-03-23 19:51 GMT+01:00 andimik <[email protected]>:

…

A user has posted a very weak raw file (please rename it to raw) which is indeed a ghost frequency of 12C on 44 MHz, but the strange thing is, when you compare eti-stuff and qt-dab in parallel (see screenshot), you will see 11 services found in eti-frontend, whereas Qt-DAB finds just 6 services. [image: rte_dab] <https://cloud.githubusercontent.com/assets/24510556/24264617/4ae84b8c-1001-11e7-8704-0f3a00ddccc7.png> AlbrechtL/welle.io#29 (comment) <AlbrechtL/welle.io#29 (comment)> — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#29>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AITzwIEt68KelxmGFI0DHgFG8pmwMCOdks5ror8XgaJpZM4MnNEM> .

-- Jan van Katwijk +31 (0)15 3698980 +31 (0) 628260355