Re: [foo-devel] Buff Bugs

[Ramon Gonzalez-A. <ar...@ma...>]

Dear Martin,

First thing, is to mention that I think Gerhard is not reading the 
foo-devel mails.  I talked to him the other day and he was certainly 
not getting them, but I thought he was trying to solve the problem, and 
I assume he has not yet been able. So, for the time being, I suggest 
that we just add his personal address to our foo-devel mails.

> ref	A number specifying time in s in the output medium that shall
> 	correspond to time origin of the context (time zero).
>
> off	A number specifying the time offset in s from the contexts time 
> origin
> 	where the execution (with run-task) should start.
>

Yes, I agree, the definitions might be confussing. As a matter of fact, 
you made me doubt at some point, as you know. I talked it over with 
Gerhard, and he was agreeing with my interpretation. So, might be the 
original idea was different, and then it was never done like that. I 
cannot say. What I can tell you is that I have used it for years as I 
told you, and was working.

> summarizing, i am still not sure whether we really discovered the
> bug.  as we said, there is one, since your example doesn't lead to the
> expected result (expected by me, in this case).
>
> i had a look at task.m again, and from what i can see i fixed a thing
> regarding to offset: it did the same before (specifying the beginning
> of rendering in the context), but it was understood to be in samples
> rather than seconds.  given the difference of several orders of
> magnitude between samples and seconds, "offset" had almost no impact,
> if i am right.
>
> so i think first we have to find out the semantics of the parameters
> off and ref.  maybe gerhard can help what the original idea was?


Well. Several questions.

You are right, Offset and Reference seem to be the same. I remember 
vaguely commenting with Gerhard about the usefulness of having two 
parameters. I think he had it more clear than me, and it was related, 
as I mentioned previously, to setting different reference points in the 
file, where offsets would be applied. Thus, you could operate 
algorithmically with relative entry points between events, 
independantly of where in a global context this network of time-related 
events would be inserted.
Certainly, even from a perspective of conceptual clarity, 
algorithmically one could deal with this aspect, without having two 
different parameters. Anyway, I think much was thought becasue of the 
incremental mixing, and the possibility of undoing a relative time 
insertion. Probably you may want to correct the whole structure of 
events in its insertion point, independantly of theur relationships, or 
modify only one, independantly of its global insertion in the context.

I can swear you. You have to beleive me, that I have been using Offset 
(and not Reference) in the way I mentioned before (i.e. as a mere 
offset from the begining of the soundfile) fro many years, and foo was 
doing the proper thing. It is impossible to mix up seconds and samples 
experiencewise. This is true of my last running Foo, which as I said 
was  some Linux version, which I used until you came here. As I said, i 
have not used Reference at all, or almost at all, so I don't know about 
Reference, But about Offset I am preety sure, I can saw you all my 
scores.

(I guess the file with which you are comparing must be a later task.m, 
and the error was introduced there. I cannot explain it myself 
otherwise....)

I agree again with you that, we have to settle up now what do we 
understand by offset and reference. I have so much work done under this 
basis that I would tend to think we should keep it as it was, but on 
the other hand it could be one more addendum to the Next-compatibility 
file. Let's hear what Gerhard might have to say about it.

All the best,

Ramon