Enhancer Bug thread

As per the documentation, Getting the SDTA_[Left|Right]Sample attributes when streaming sound results in a null response (that's from V44 autodocs). The A-Eon datatypes stream by default.

MPEGA library is not included because of licencing issues.

I believe (but am happy to be corrected) those are only supposed to return NULL if the DataType doesn't support extracting the data (ie. only streams like the MPEGA DT) or - based on the wording you've given, but not something I'm familiar with - if the DataType is currently streaming, ie. SDTM_PLAY has been called.

In all other cases the sample data should be returned.

When I complained that the new sound.datatype was playing mono sounds through a single channel instead of both stereo channels like the OS4 dound.datatype I was told it was not possible with the new sound.datatype. I have my doubts about that too.

The API compatibility isn't broken - it's the same as it's always been since 3.9 and I try to take great care (sometimes fail) with the sound.datatype to make it so.  What you're experiencing is a NULL return value for the SDTA_Sample (or the left/right equivalent) which is clearly described in the OS3.9 and OS4 sound_dtc autodoc:

  "SDTA_Sample [...] For streaming sound playback, SDTA_LeftSample,
SDTA_RightSample, and SDTA_Sample will all be NULL (V44)"

...and...

  "SDTA_Continuous [...] Default for this tag is FALSE"

These statements hold true for the new sound.datatype - the SDTA_Continuous flag is FALSE but if it is set, the sample attributes will return NULL for a whole variety of reasons - the main one being that there is no single buffer.  I was about to argue that if you perform the SDTA_Sample when loading 8SVX or AIFF samples, you SHOULD get the buffer as these are the stock OS4 datatypes that don't stream.  As pointed out by "Xenic" on amigans.net, unfortunately it turns out that didn't work either - all my fault - but I have now fixed this scenario and shall have sent the new version to Matthew for his approval.

Now, the new A-EON datatypes for wav, mod and mpega are self-streaming - that is, the subclasses instruct the sound.datatype to set SDTA_Continuous and expect a stream.  It doesn't make sense for audio files, especially compressed, to persist in memory in one large buffer - MP3 files can expand quite significantly and consume 50+MB of memory for several minutes when you only need a few kilobytes per second.  This is especially pertinent under A1XE/SE with lower memory resources or even OS3.9 (if the datatypes ever make an appearance there) along with slow longer times on those machines!

Let's not forget that the sound.datatype is 25 years old and was designed primarily for small one-shot samples of that era which could be loaded into memory - that's not where we are today.  I will, however, investigate how easy it would be to load smaller files into memory and switch to streaming above some threshold - no guarantees on that though.

If you're really wanting to read all of the data from the file, there is the SDTM_FETCH method (which takes sdtFetch msg) where you can read data, regardless of whether it is streaming or not.  You can request the number of samples or number of bytes by way of setting a flag.  You can even change the format by supplying appropriate attributes.  Just make sure your buffer is big enough to take the data you want and you may have to make several calls to get everything.  This method was created for this very purpose (and another reason I won't go into).

Unfortunately, it's still the same: SDTA_Sample returns NULL (others attributes are OK). No sound played.

I use an AmigaOne XE G4. Same result on a X1000

Do you have the same wav.datatype and sound.datatype as me ?

Please be reassured that we respond very quickly to any queries that are made about Enhancer Software components and if any problems are present, they are fixed as rapidly as possible.

In general the cursor in MultiEdit is unusual in that it disappears if you click outside the editing area.

I think the cursor in NotePad remains as long as the window is active.

The cursor in TTX is just ghosted when the window is inactive so you still know where the cursor is; which is what I prefer.

Does the "Help" key (Scroll Lock on the keyboard) work with MultiEdit for you? Here Multiviewer opens with an empty window and the system freezes.

I'm using a new X5000 so I don't know yet if it's a problem with the program or the OS.

The difference in MultiEdit is that it doesn't automaticaly reactivate the editor gadget until you move the move over the gadget.

Does TTX use the texteditor.gadget?

Do you get a crashlog on serial? Should be able to work it out from that.