update to beta 2
ethan
there tried this extensively?) and found that i was having some
allocation/deallocation problems and the amount of free memory for
loading samples was significantly reduced after a few wav loads. i've
made a couple small changes to improve things. not perfect but
better. i've uploaded the relevant eboots here:
http://dspmusic.org/psp/PSPSeq210-beta3.zip
note it's just the eboots, none of the support files. keep everything
else from beta 2. i'm going to test this over the weekend (but am in
a big rush at the moment), wanted to get these out there ASAP so
others could test them too. let me know if you run into any
problems. thanks!
ethan
loading wav files it is too easy to hit the maximum memory ceiling
(1.3MBytes). the issue is that PSPSeq doesn't remove the old wav file
before trying to load the new one. PSPSeq can load up to 1.3MBytes of
WAV data. however, if you load 1 700k WAV file and try and swap it
with another 700k wav file you'll get an error (because it tried to
balance both simultaneously). what i need to do is deallocate the
first wav file, and then see if the second can fit; if so load wav #2
otherwise load wav #1 back in. you don't really see this problem if
you only load small wav files, but if you load big ones it's easy to
get locked out well before the limit is actually reached. i'll try
and fix this this weekend. FYI the pbp i just uploaded fixed a more
basic and serious problem with wav file handling, so there's still
some value to giving it a go.
sorry if this too much info, just thinking aloud here!
Mr. Brew
though, I rebuilt the 1.0 one and it works OK on OE.
I've not used the WAV functionality yet, that's why you haven't heard
me whining about loading problems BTW ;)
Thanks for the background info, it's interesting. Why not just read
the length of the WAV data from inside the RIFF header and compare
that to free-RAM plus size-of-sample-to-be-replaced, before loading
all the data into RAM?
ethan
> though, I rebuilt the 1.0 one and it works OK on OE.
oops was in a big rush out the door. i will try and tackle this
problem for real over the next couple days.
> I've not used the WAV functionality yet, that's why you haven't heard
> me whining about loading problems BTW ;)
bad beta tester! ;p it's funny, getting the wav stuff to work was a
pretty big pain, but i don't know if anybody's using it. i know that
the wacky synths are more interesting, but i think the wav file
playback is pretty powerful and flexible too. it's fun to load in a
loop (one loop, zimba_ku, is in the samples), trigger at step 0 and
add in a bunch of no-retrigger hits, and then test out the various FX
in pseudo-realtime.
> Thanks for the background info, it's interesting. Why not just read
> the length of the WAV data from inside the RIFF header and compare
> that to free-RAM plus size-of-sample-to-be-replaced, before loading
> all the data into RAM?
because i don't know -exactly- how much memory is available,
especially after mallocing and freeing a bunch of files previous. the
only way to be sure it can load is to actually try and load it. it's
really not that bad, just not particularly elegant.
Mr. Brew
I'll be using it extensively from here on - it's great! The FX in
particular sound the business.
One small bug: BPMSYNC doesn't seem to work correctly for tempos other
than 120bpm. When you turn up the tempo, the playback speed goes down
and vice versa. Something backwards in the maths?
It would be nice to be able to change STARTLOC and ENDLOC without them
affecting playback rate when using BPM sync. For example, if I try to
emulate ReCycle by triggering different hits within a loop, I have to
change RDSPEED accordingly. This isn't too bad for triggering at
quarter notes, but say I want to trigger at 7/8 or 3/16, the mental
arithmetic soon becomes overwhelming for my puny brain and I have to
break out the calculator.
Also, (totally unreasonable feature request) how about having two
separate modes for wav playback (a la FREQMODE for the synths). One
could provide pitched playback (I'm really struggling with the old 440
x 2^(x-12) - more calculator fun :) and the other mode oriented
towards loops. While I'm fantasising, it would be awesome to be able
to set the frequency of the SVF with note values for the same reason.
BTW had a crash when going SONG>SEQUENCE (I think while stopped in
SONG mode). Just prior to that I had removed the last few entries in
the list via the triangle button. I haven't been able to reproduce it
so far.
Another small glitch I only noticed today: when deleting synths, the
parameters for any synths playing at the time get audibly messed up.
The actual parameters all seem to be OK, and a save and reload fixes
it, so not a big problem.
Peter Swimm
Peter Swimm
www.toilville.com
Peter Swimm
ethan
> I'll be using it extensively from here on - it's great! The FX in
> particular sound the business.
cool, glad you like. i think there's some possibilities for some
interesting non-repeating dnb by loading in a few loops and setting
some probabilities on retriggering between 0-100, along with messing
with the fx and wav retrigger points on "important" steps.
> One small bug: BPMSYNC doesn't seem to work correctly for tempos other
> than 120bpm. When you turn up the tempo, the playback speed goes down
> and vice versa. Something backwards in the maths?
hey you're right - obviously i haven't used WAV that much either. :)
it's fixed.
> It would be nice to be able to change STARTLOC and ENDLOC without them
> affecting playback rate when using BPM sync. For example, if I try to
> emulate ReCycle by triggering different hits within a loop, I have to
> change RDSPEED accordingly. This isn't too bad for triggering at
> quarter notes, but say I want to trigger at 7/8 or 3/16, the mental
> arithmetic soon becomes overwhelming for my puny brain and I have to
> break out the calculator.
i wasn't sure the best way to handle this. my thought was that
BPMSYNC should overrule everything else, so if it is on and RDSPEED is
1.0, it will always play a loop in BPMSTEP sequencer steps, end of
story. i haven't tried too many other configurations, but it seemed
to me that this way of handling things was reasonable enough. i guess
the only thing you can do is turn the BPMSYNC off and hope that you
don't change the BPM half way through writing your song. :\ either
way there's going to be limitations. i will probably revisit all of
this for version 3.00.
> Also, (totally unreasonable feature request) how about having two
> separate modes for wav playback (a la FREQMODE for the synths). One
> could provide pitched playback (I'm really struggling with the old 440
> x 2^(x-12) - more calculator fun :) and the other mode oriented
> towards loops. While I'm fantasising, it would be awesome to be able
> to set the frequency of the SVF with note values for the same reason.
yeah this is on the future agenda. adding this now would break .seq
compatibility (plus be a fair bit of work to implement) so for 2.10
it's out.
> BTW had a crash when going SONG>SEQUENCE (I think while stopped in
> SONG mode). Just prior to that I had removed the last few entries in
> the list via the triangle button. I haven't been able to reproduce it
> so far.
yike! that's no good. so you had the sequencer stopped, made some
changes to SONG->SEQUENCE, returned to the sequencer, started playback
again and then it crashed? did you reset the sequencer in the
transport (X on PLAY/STOP)?
> Another small glitch I only noticed today: when deleting synths, the
> parameters for any synths playing at the time get audibly messed up.
> The actual parameters all seem to be OK, and a save and reload fixes
> it, so not a big problem.
really? is this reproducable? if so, can you explain how you did
it? don't believe i've seen this one before.
thanks!
Mr. Brew
> BPMSYNC should overrule everything else, so if it is on and RDSPEED is
> 1.0, it will always play a loop in BPMSTEP sequencer steps, end of
> > BTW had a crash when going SONG>SEQUENCE (I think while stopped in
> > SONG mode). Just prior to that I had removed the last few entries in
> > the list via the triangle button. I haven't been able to reproduce it
> > so far.
>
> yike! that's no good. so you had the sequencer stopped, made some
> changes to SONG->SEQUENCE, returned to the sequencer, started playback
> again and then it crashed? did you reset the sequencer in the
> transport (X on PLAY/STOP)?
The crash actually happened as I was in the menu. I selected SONG (d-
pad right), then SEQUENCE (d-pad right) and it locked up completely
before displaying the sequence list. I still haven't been able to
reproduce this, sorry :(
> > Another small glitch I only noticed today: when deleting synths, the
> > parameters for any synths playing at the time get audibly messed up.
> > The actual parameters all seem to be OK, and a save and reload fixes
> > it, so not a big problem.
>
> really? is this reproducable? if so, can you explain how you did
> it? don't believe i've seen this one before.
I can't seem to make it do the glaringly obvious glitching it was
before, but try this (which demonstrates the STEREOIZE paramters
aren't being updated properly):
* see PlaybackUpdateBug-[MrBrew].zip in the files area
* go to loop 04
* listen to ROTSVF0 while deleting NBD0
this is quite subtle and probably only noticable on headphones. I'll
see if I can work out how to make the more serious glitching happen
again.
I also discovered another bug today which results in a fatal crash:
* see ReverseWAVBug-[MrBrew].zip in the files area (I've included the
WAVs)
* go to loop 01
* move the cursor to X:33 / Y:3 (this note has a READSPEED of -1.0,
which worked fine at first, not sure what happened)
ethan
> pad right), then SEQUENCE (d-pad right) and it locked up completely
> before displaying the sequence list. I still haven't been able to
> reproduce this, sorry :(
well that's a good clue. did anything appear (like a messed up menu
box) on the screen when you pressed d-pad right on SEQUENCE? i'll
poke around and see what i can see.
> I can't seem to make it do the glaringly obvious glitching it was
> before, but try this (which demonstrates the STEREOIZE paramters
> aren't being updated properly):
> * see PlaybackUpdateBug-[MrBrew].zip in the files area
> * go to loop 04
> * listen to ROTSVF0 while deleting NBD0
> this is quite subtle and probably only noticable on headphones. I'll
> see if I can work out how to make the more serious glitching happen
> again.
no headphones at the moment, but i can see the problem with stereoize;
the parameter isn't shifted up when deleting tracks. believe it's now
fixed & will experiment with deleting tracks to see if i can recreate
the glitching & make sure everything is copied up.
> I also discovered another bug today which results in a fatal crash:
> * see ReverseWAVBug-[MrBrew].zip in the files area (I've included the
> WAVs)
> * go to loop 01
> * move the cursor to X:33 / Y:3 (this note has a READSPEED of -1.0,
> which worked fine at first, not sure what happened)
good catch - this is actually a save/load bug where negative values
aren't saved (or loaded!) properly. should be fixed, though it
definitely needs some additional testing.
lots of little bugs are popping up. good to catch them now though i
must say i am a bit surprised. amazing what can lurk in the
shadows...
ethan
playing around with the windows build of lgpt for a few minutes and
couldn't get any audio out (not familiar w/ LSDJ so the interface is a
little baffling)... thanks!
Mr. Brew
> box) on the screen when you pressed d-pad right on SEQUENCE? i'll
> poke around and see what i can see.
No, it died the moment after I pressed the d-pad - there was no screen
update before it froze up.
> no headphones at the moment
Surely not! If I may make a recommendation: Sennheiser HD201s. They
are cheap as chips and sound brilliant. They've the best
price:performance ratio of any cans I've ever owned. Let me buy you a
pair as a thank you for PSPSeq?
> lots of little bugs are popping up. good to catch them now though i
> must say i am a bit surprised. amazing what can lurk in the
> shadows...
Just point me where the testing needs to happen. I seem to be a bit of
a bug magnet :(
Mr. Brew
> playing around with the windows build of lgpt for a few minutes and
> couldn't get any audio out (not familiar w/ LSDJ so the interface is a
> little baffling)... thanks!
This is basically just looping a portion of a sample as an oscillator,
in much the same way as the synth waveforms in PSPSeq are just
wavetable data. It tunes the loop to appropriate length so it's in
tune. Little difference from the 'chiptunes' you hear made with
tracker software.
LGPT's pretty nice if you're doing sample based stuff - I really like
the flexibility of the sequencer. But... I don't have a GP32/GP2X and
am unlikely to buy one (PSP does nearly everything better once modded)
- and PSPSeq's unique synths rock in a way sample trackers never
can :)
ethan
> update before it froze up.
ok i just tried to make this happen and i couldn't - and actually i've
never seen anything like it in the past. if you do run into it again,
try and let me know any details please... thanks!
(at least the crash happens when trying to use the menus. it would be
worse if it happened just while playing back a sequence.)
> > no headphones at the moment
>
> Surely not! If I may make a recommendation: Sennheiser HD201s. They
> are cheap as chips and sound brilliant. They've the best
> price:performance ratio of any cans I've ever owned. Let me buy you a
> pair as a thank you for PSPSeq?
hah! actually i have a few pairs of headphones (giant sony for
PSPSeq, crappy koss for the gym, broken grado sr80s); i just didn't
have any on me at the moment of my reply. ;) you're near london,
right? i might be in cambridge during the summer. instead of
headphones i'd gladly meet up and have a pint in their place. :)
oh and as for headphones, generally not a huge sennheiser fan though i
did like the HD555s quite a bit. not sure i've ever heard the 201s
but i'll keep them in mind in case i need another pair.
> Just point me where the testing needs to happen. I seem to be a bit of
> a bug magnet :(
no kidding, but it's appreciated. i just did a pretty big rework and
optimized how wav files are loaded to get around that problem when you
can get "locked out" of swapping wavs because i don't deallocate the
old wav first. i have a lot of testing to do, but that's probably
where the most investigation should happen when a new beta is ready
(most likely over the weekend).
Mr. Brew
> right? i might be in cambridge during the summer. instead of
> headphones i'd gladly meet up and have a pint in their place. :)
Yeah - how did you know? I'll assume it's down to my misuse of the
English language. I'm in Stevenage, which is about halfway between
Central London and Cambridge by train. Taking a break from hackerdom
to run a pub at the moment (believe it or not) so as many pints as you
can handle shouldn't be a problem :)
> oh and as for headphones, generally not a huge sennheiser fan though i
> did like the HD555s quite a bit. not sure i've ever heard the 201s
> but i'll keep them in mind in case i need another pair.
My favourites are Beyerdynamic as it happens - the 201s are nice for
music on the move because being so cheap I don't mind if they get
broken / rained on / stolen :)
ethan
> Yeah - how did you know? I'll assume it's down to my misuse of the
> English language. I'm in Stevenage, which is about halfway between
> Central London and Cambridge by train. Taking a break from hackerdom
> to run a pub at the moment (believe it or not) so as many pints as you
> can handle shouldn't be a problem :)
can't remember how i figured it out, from your email address?
definitely not picking up on the britishisms from the written word. i
think i once had to take a taxi from stevenage to cambridge (made a
big mistake at king's cross). that sucked. my wife and i both love
the uk and i might go over (again) for a short-term work transfer
sometime in the summer. tbd. running a pub, eh? as many pints as i
can handle eh? heh heh.
blah blah blah if this thread continues much further i might have to
ban myself from this forum. ;)