From:
                                                            5/15/97 11:00 PM
Subject: MAX Digest - 14 May 1997 to 15 May
1997To: Recipients of MAX digests 

There are 9 messages totalling 345 lines in this issue.

Topics of the day:

  1. sprintf
  2. Phonemes / speak / native? (2)
  3. snd
  4. Getting rid of the Status window menu item
  5. Efficiency
  6. Is it possible to sell your standalone Max stuff? (2)
  7. phonemes

----------------------------------------------------------------------

Date:    Wed, 14 May 1997 23:25:32 -0500
From:    Steve Salani 
Subject: sprintf

Hi MAXers,

The sprintf object supplied with MAX2.5.2 and 3.0 supports a hex format
specifier (%x) but it's got a bug -- some input values are not correctly
converted to hex, but instead a question mark apears in the output message.

Is there a bug-free version available on an ftp server somewhere? If not,
is there another simple way to convert a number into a hex symbol (the
number box will display hex, but it will not generate the corresponding
text symbol).

Thanks,

Steve Salani
Forte Music
Austin, TX

------------------------------

Date:    Thu, 15 May 1997 08:31:50 +0200
From:    "Dr. K@rlheinz Essl" 
Subject: Re: Phonemes / speak / native?

A year ago I also tried to find sources for (english) phonemes, and the
result of this search was disappointing. The only thing I found were
synthesized phonemes of inferior quality (8bit, u-law) which are completely
useless for artistic work.

The best thing I came across was the TIMIT database on CD-ROM. However, you
have to buy it, and it is not cheap... But the quality is excellent. You
will find the project on the net, but you won't find samples there.

Joseph Hyde  wrote

> Does anyone know of a way to really 'get inside' the Speech Manager or
> PowerTalk, in a more sophisticated way than the 'speak' or 'talk' objects;
> so that one could actually change the pitching of the voices, or even
their
> timbral characteristics

I am also very interested in this question and I would love to have a MAX
object that is capable of doing these things. Outside MAX, there is some
software existing which lets you access the more advanced features of the
SpeechManager, such as "Speaker" by Eric Shieh, "Speecher" by P.H.Pedersen,
and "SING!" by Kurt Kaufman.

> Has anyone done any interesting work using speech with Max?

Check out the "Lexicon-Oracle" at:

        http://www.ping.at/users/essl/workslexicon-oracle.html

This patch is also included into the newest release of my "Realtime
Composition Library for MAX", available under:

        http://www.ping.at/users/essl/works/rtc.html

---

Does anybody know about a native version of a MAX speaker object? At the
moment all existing externals (talk, speak, ispeak) are only 68k-coded, and
won't run with the native version of MAX3.5!

Richard Dudas was nice enough trying to recode Michael Klingbeil's "speak",
but he did not succeed yet...

Greetings,

Dr. K@rlheinz Essl
SAMT - Studio for Advanced Music & Media Technology
Bruckner-Konservatorium Linz / Austria
www:   http://www.ping.at/users/essl/index.html

________________________________________________________________________
     Visit the MindShipMind - an interactive multimedial Web Oracle
          http://www.ping.at/users/essl/works/mindshipmind.html
________________________________________________________________________

------------------------------

Date:    Thu, 15 May 1997 08:35:34 +0200
From:    "Dr. K@rlheinz Essl" 
Subject: Re: snd

Jem Finer  asked:

> I don't seem to be able to get snd to hold more than one sound at once . .
> whats the trick for giving each sound its voice no. - I've tried a la
> manual but no luck so far - am I missing some vital info ?

You need to do some sort of voice allocation - check out my "Amazing Maze"
to see how I solved this problem and how the snd object can access more
than one sound a time:

        http://www.ping.at/users/essl/works/amazing.html

Happy MAXin'

Dr. K@rlheinz Essl
SAMT - Studio for Advanced Music & Media Technology
Bruckner-Konservatorium Linz / Austria
www:   http://www.ping.at/users/essl/index.html

________________________________________________________________________
     Visit the MindShipMind - an interactive multimedial Web Oracle
          http://www.ping.at/users/essl/works/mindshipmind.html
________________________________________________________________________

------------------------------

Date:    Wed, 14 May 1997 23:57:22 -0700
From:    David Zicarelli 
Subject: Getting rid of the Status window menu item

I previously wrote that the feature to exclude the status window
from the Windows menu in Maxplay or standalone applications was
a feature of the application installer. I was wrong. It's not.
It is however, a message you can send to max in version 3.5:

"; max maxinwmenu 1" shows the Status menu item.
"; max maxinwmenu 0" does not show it.

Loadbang this sucker and the Status menu will be gone.

My apologies to K-) Peter for leading him astray and then blaming the
victim.

David Z.

------------------------------

Date:    Thu, 15 May 1997 03:08:21 -0500
From:    Greg Kuhn <0003939581@MCIMAIL.COM>
Subject: Re: Efficiency

I looked for an old posting by DZ regarding efficiency from when I was
asking about it some time ago... couldn't find it. But what I deduced
was this:

Generally speaking, two aspects relate to the efficiency of a patch or
an object: how much memory it takes up and what or how much work the
object/patch has to perform.

For examples, respectively:

To store two integers use a _pack instead of a _table.

To add two numbers use a _+ not an _expr. (All the effort to develop
specialized _()gates is a big clue in this category.)

I would also consider how much time is involved. As I recollect DZ
said nothing about overhead for objects which occur over time... my
presumption. For instance, to cause a right outlet to occur after a
left outlet don't _delay it, _swap it.

Also, you have to figure that screen updates demand extra clock
cycles.

Thus, to answer your query, I would presume that _bpatchers take up
more memory and clock cycles than _(sub)patchers.

For a more subtle example, if in a section of a patch where I need to
make decisions when not much else is happening I'll stack _if's and
_expression objects as needed because you can get them to do more
complicated procedures within single objects --they take up less
memory.  But, in other places when much needs to happen quickly I'd
avoid _if's and and _expr's and use multiple objects or preconfigured
tables or _coll's, etc.. because even though they may take up more
memory in your patch, they may execute with less steps to figure out
what to do --less CPU demand.

At times I've gone so far as to use the Max window to measure the size
of a 100 of a given object to compare memory usage between objects or
patches that do the same thing. Unreliable, silly, but informative?
_message 1 seems to take up less memory than _int 1 (by a very small
amount), but I suspect it takes more code to execute because it is
capable of doing so much more.

I don't know wether or not built-in objects make less demands than
objects from the "patches" or "externals" folders... I suspect in
essence --assuming their size and weight are the same --they are
equivalent in their instances within your patcher.

But, ultimately this is all piecemeal.  If your patch runs slow,
double check your programming --look for redundancies, overkills (e.g.
_speedlims before a display may simply add more processing than just
displaying it,) and processes you can simplify-- and if that doesn't
help cut something out.  I'd start with graphics.

All of this doesn't include reference to the almighty _defer option,
which only functions in Overdrive. Beware deferring messages though,
they can stack up quick and take a long time to play out.

What might be hip is if the trace mode displayed some kind of
efficiency info. (What is the "Clock executing" thing at the end of
Traces?)

But personally, I'd rather a window zoom feature when in editing mode,
and !!!! the ability to click-select an object that is buried under
patch cord connections.

(Just another journal entry from a Max abductee on the Y-files.)

gregory

re:>>>>>>>>>

Date:    Fri, 9 May 1997 09:20:28 +0100 From:    "K-) Peter"
 Subject: Efficiency

> >>        In terms of programming efficiency, are there any major
pitfalls in Max'ing; i.e. are Patchers more efficient processor wise ?
Anything to avoid or use in preference ? Hmmn. Did you read the
"Efficiency" section of the manual? David Z.

> Sorry but I've only just finished the tutorial manual. The "two
pages" I've now found in the reference manual doesn't really answer my
question anyway. 1) Can I assume that the less objects the better even
if they are more complex ? Where is this list of object efficiency
glossed over in the two pages ? 2) Is it safe to use any number of sub
patchers. Presumably Bpatchers have the same overhead ? Yes I've also
seen the two pages on Encapsulation ;-)

------------------------------

Date:    Thu, 15 May 1997 13:50:08 +0200
From:    arjen van der Schoot 
Subject: Is it possible to sell your standalone Max stuff?

Hi,

As I Gather from the recent discussion, Peter seems to be working on a
standalone app that shows little signs of Max, which reminds me of a
question I wanted to ask some time now.
The Max Reference is quite clear: You may give away (but not sell!) max
collectives.
Is there a licensing contract possibility or something similar, or is this
sentence outdated?

The 02R projectmanager and localiser stuff from www.zeep.com are examples
of commercially available max collectives.
Can I sell my standalone max stuff or not?

In my view, Max would become a serious development platform if you could.
It is increasingly fast and stable and is for many projects at least
competitive with Director and C++. The one thing that could render  it
useless as a development tool for commercial apps, is this small clause
that I stumbled on in the reference.

Anybody?

Arjen

------------------------------

Date:    Thu, 15 May 1997 14:19:54 -0400
From:    Charles Cohen 
Subject: Re: phonemes

> whereabouts of the complete set of
> >phonemes for the speaking of English

some here:
http://sunsite.unc.edu/pub/multimedia/sun-sounds/phonemes/

            ****      What's Charles up to?      ****
                 http://www.voicenet.com/~ccohen

------------------------------

Date:    Thu, 15 May 1997 16:01:50 -0400
From:    Stephen Kay 
Subject: Is it possible to sell your standalone Max stuff?

>Can I sell my standalone max stuff or not?

>In my view, Max would become a serious development platform if you could=
=2E
>It is increasingly fast and stable and is for many projects at least
>competitive with Director and C++. The one thing that could render  it
>useless as a development tool for commercial apps, is this small clause
>that I stumbled on in the reference.

AFAIK, it's entirely legal to sell completed MAX applications which have
been created with the application installer.  I believe the clause you
speak of refers to distributing copies of "MaxPLAY" along with collective=
s,
which used to be the only previous option available before the applicatio=
n
installer.

However, don't get all excited about it being the perfect development
platform.  Maybe for very small simple applications, but MAX is bloated
(consider that 100 "+" objects replicate that same code 100 times, where =
in
a coded application it might be a single function call).  Note that a
completed MAX application which does nothing but route one midi channel t=
o
another ends up occupying around 600k of disk space and still takes 2mb o=
f
RAM - I venture to say that hard-coded in C or C++ it would be way smalle=
r
(because there's all this other MAX stuff in there that doesn't get used)=
=2E =

Stephen Kay

------------------------------

Date:    Thu, 15 May 1997 16:01:55 -0400
From:    Stephen Kay 
Subject: Re: Phonemes / speak / native?

Karlheinz Essl:
>The best thing I came across was the TIMIT database on CD-ROM. However,
you
>have to buy it, and it is not cheap... But the quality is excellent. You=

>will find the project on the net, but you won't find samples there.

address, please?

Thanks,
Stephen Kay

------------------------------

End of MAX Digest - 14 May 1997 to 15 May 1997
**********************************************