MVerb - A Free Open Source Reverb

VST, AU, AAX, CLAP, etc. Plugin Virtual Effects Discussion
Post Reply New Topic
RELATED
PRODUCTS

Post

i have delayed the 'release' of these for a few days now, as i always have this doubt in the back of my head, that there might be some bugs left, or that there's some other things i've forgotten or overlooked.. but, decided to jump into the deep water, and hope the best..

so, here's some alternative versions you can play around with.

windows:
mverb-vst.dll (233k)
mverb-vst-nogui.dll (95.5k)
mverb-ladspa.dll (86k)
mverb.exe (199k)

linux:
mverb-vst.so (209.6k)
mverb-vst-nogui.so (45.6k)
mverb-ladspa.so (41.6k)
mverb (154k)

(links are temporary, until i find some time to fix the ccernn.audio site, for all the new stuff that's coming, or we update the axonlib site to have binaries for download. we'll see)

some notes:
they're built with axonlib (not using vstgui), all compiled from the same (unmodified) source file. the linux gui version (vst/exe) requires XRender (for alpha/transparency). let me know if there's any problems with that, and i'll try to finish a Xlib-only version. the standalone versions doesn't do much (missing audio in/out), but is included for testing. all (cross-)compiled from within linux (gcc v4.4.3). i might boot into windows and re-compile the windows versions there, but only if i really have to (for the obvious reasons).

please report any bugs, problems or other issues, so we can fix them.
and, as always, all source code and more info can be found here: http://axonlib.googlecode.com/

- ccernn (with massive contributions from neolit123)

Post

just a little update from me. i installed java, microsoft redistrabutable package, and the freshly compiled files from tor.helge.skei. none of these made mverb work on my amd machine. :[

anyway, i will stop cluttering this thread now and use mverb at least when i'm on the road, on my lappie. cheers!
Bedroom Producers Blog << Free VST Plugins!

Post

Very interesting, is there a possibility to build a LV2 or DSSI version ?
You can't always get what you waaaant...

Post

i'd like to say big ups to caco for the great effort on mverb. it sounds absolutely great and is quite efficient ! cheers. :-)
bpblog wrote:just a little update from me. i installed java, microsoft redistrabutable package, and the freshly compiled files from tor.helge.skei. none of these made mverb work on my amd machine. :[

anyway, i will stop cluttering this thread now and use mverb at least when i'm on the road, on my lappie. cheers!
hey bpblog, i would like to encourage you to test a bit more.
i think the MSJAVA.DLL and java runtime or sdk dependencies are not needed at all in this case.

i suspect that there might be something else wrong with this particular system.

for example, the only dependency error that will be reported for axonlib binaries will be related to DWMAPI.DLL which is a vista dll afaik.
it is important in the case of IE7 installed on the system (and certain version of IEFRAME.DLL) and should be ignored as any calls will be most likely "delayed" and in fact never made.

the ladspa win32 builds should have pretty much the same dependencies.

i'm curious about your system.
what os and cpu arch exactly ?
are there any crash logs ?
stanlea wrote: Very interesting, is there a possibility to build a LV2 or DSSI version ?
both are on the way.. :)

we currently multi-format-compile in the lines of a single parameter such as -vst or -ladspa.

i guess dssi will be the first one to be added probably and we are going to wrap it around the ladspa format as thats what the dssi format represents essentially.

[edit]: but also the ladspa format should work with dssi supporting hosts since the interface should be exposed.

lubo


---
linux/win32 framework for plugins and exacutables:
holos.googlecode.com

Post

tor.helge.skei wrote: so, here's some alternative versions you can play around with.

windows:
mverb-vst.dll (233k)
mverb-vst-nogui.dll (95.5k)
...
please report any bugs, problems or other issues, so we can fix them.
and, as always, all source code and more info can be found here: http://axonlib.googlecode.com/
a quick report :
host Chainer, windows xp+sp3+mslib, amd athlon 2200
mverb-vst.dll + mverb-vst-nogui.dll
scanned by the host but not running.
original mverb and GUi version working.
Last edited by novaflash on Fri Jun 18, 2010 7:35 am, edited 1 time in total.
Image

Post

Awesome, thanks for the Linux builds ccern :)

The windows versions are working okay here, although I noticed a minor bug in eXT1.4. If you open the GUI and click on the drop down box for selecting presets then the preset name doesn't show up. It is just a blank space but you can click it to select the preset and it loads okay so there is just a little bug in how you report program names to the host.

Post

Thank you for the Linux versions !

I tested the LADSPA and VST (with GUI) in the JOST host, both worked very nicely. The reverb itself is excellent, and its controls are smooth and responsive.

Well done, sir. :) I look forward to the public release.

Best,

dp

Post

novaflash wrote:
a quick report :
host Chainer, windows xp+sp3+mslib, amd athlon 2200
mverb-vst.dll + mverb-vst-nogui.dll
scanned by the host but not running.
original mverb and GUi version working.
thanks for the report.
fix coming soon, with help from thomas [at] xlutop
Last edited by lubomir.ivanov on Fri Jun 18, 2010 1:16 pm, edited 1 time in total.
linux/win32 framework for plugins and exacutables:
holos.googlecode.com

Post

Caco wrote:The windows versions are working okay here, although I noticed a minor bug in eXT1.4. If you open the GUI and click on the drop down box for selecting presets then the preset name doesn't show up. It is just a blank space but you can click it to select the preset and it loads okay so there is just a little bug in how you report program names to the host.
i've seen the same behaviour in energy xt2 too, both the linux and windows version, but no other hosts.. will investigate...
- ccernn

Post

at first glance, it sounds similar to DaSample Glaceverb, that is a nice one too.

thanks for it.

Post

fixed axonlib related issues:
- energyxt blank items in preset menu
- fixed compatibility with chainer and host that use pre vstsdk 2.4

http://stashbox.org/931131/mverb_ax_bui ... 061330.zip
(tmp link / updated)

caco, mingw 3.4.2 is a bit picky for typecastings, log:
http://stashbox.org/930865/mverb_mingw_ ... asting.log

btw you are using msvc for the windows-vst builds i presume?

i get the following performance differences from less to more cpu usage on my 10y/o pc:
msvc (original Mverb.dll?) < mingw-3.4.2-official < mingw-4.4.1-tdm < mingw-4.5.0-tdm

the mingw order is a bit in contradiction with what the gcc improvements and changes in standard optimization flags suggest.

interesting to investigate a bit more into this...
Last edited by lubomir.ivanov on Sat Jun 19, 2010 10:48 am, edited 2 times in total.
linux/win32 framework for plugins and exacutables:
holos.googlecode.com

Post

lubomir.ivanov wrote:fixed axonlib related issues:
- energyxt blank items in preset menu
- fixed compatibility with chainer and host that use pre vstsdk 2.4

http://stashbox.org/930866/mverb_ax_bui ... 060250.zip
(tmp link)

caco, mingw 3.4.2 is a bit picky for typecastings, log:
http://stashbox.org/930865/mverb_mingw_ ... asting.log

btw you are using msvc for the windows-vst builds i presume?

i get the following performance differences from less to more cpu usage on my 10y/o pc:
msvc (original Mverb.dll?) < mingw-3.4.2-official < mingw-4.4.1-tdm < mingw-4.5.0-tdm

the mingw order is a bit in contradiction with what the gcc improvements and changes in standard optimization flags suggest.

interesting to investigate a bit more into this...
I am a little confused. I apologize in advance. Is the Windows link you are providing the link to the latest and newest version? Should we be downloading this version as the bugfixed version?

Thanks,

--Sean

Post

audiojunkie wrote: I am a little confused. I apologize in advance. Is the Windows link you are providing the link to the latest and newest version? Should we be downloading this version as the bugfixed version?

Thanks,

--Sean
hi, the windows-vst and linux-ladspa versions in the zip above a duplicates to the versions already posted in the thread. if you are a windows (vst) or mac (au) user, you should check the link on the first page.

we add a windows-ladspa and linux-vst versions that were previously missing and the bugfixes are only for the library code that was used to make these.
linux/win32 framework for plugins and exacutables:
holos.googlecode.com

Post

lubomir.ivanov wrote:fixed axonlib related issues:
- energyxt blank items in preset menu
- fixed compatibility with chainer and host that use pre vstsdk 2.4

http://stashbox.org/931131/mverb_ax_bui ... 061330.zip
(tmp link / updated)

caco, mingw 3.4.2 is a bit picky for typecastings, log:
http://stashbox.org/930865/mverb_mingw_ ... asting.log

btw you are using msvc for the windows-vst builds i presume?

i get the following performance differences from less to more cpu usage on my 10y/o pc:
msvc (original Mverb.dll?) < mingw-3.4.2-official < mingw-4.4.1-tdm < mingw-4.5.0-tdm

the mingw order is a bit in contradiction with what the gcc improvements and changes in standard optimization flags suggest.

interesting to investigate a bit more into this...
All development was done using mingw to ensure cross platform compatable code but the final build is done in msvc as it is so much better at optimizing the code. My mingw 3.4.2 version used 50% more CPU than the msvc version and the mingw 4.4.1 used 5% more than the mingw 3.4.2 version! GCC on windows is slow and seems to be getting slower :? I have heard it is slow on OSX as well though.

Thanks for the log, will look at trying to remove the warnings when I get chance.

Would you mind if I hosted the Linux versions so that all different formats can be found in one place? It should then be easier for people to find the different versions.

Post

Caco wrote: My mingw 3.4.2 version used 50% more CPU than the msvc version and the mingw 4.4.1 used 5% more than the mingw 3.4.2 version! GCC on windows is slow and seems to be getting slower :? I have heard it is slow on OSX as well though.
the difference there seems quite a lot, so i just had to test a little (in reaper, win7). ten instances of the plugin on a track:

your original version: 23.3%
my first version (cross-compiled from inside linux): 28.7%
lubomir's second version (compiled in windows, i guess?): 25.7%

EDIT: just noticed the samnplerate were set to 192khz :-) but it doesn't matter that much, it's the difference that is interesting here..

there's certainly a difference there, but not the 50% + that you mentioned?
some of it is probably because axonlib have some extra layers for platform and plugin-format independence, and is filled with virtual functions and whatnot (and not much major 'manual' optimizations have been done yet.. will soon..). also, if i'm not utterly mistaken (lubomir, correct me if i'm wrong), we're using only some basic compiler flags (opt="-mfpmath=387 -O3 -Os"), so it could be interesting to try out some of the possible exotic flags...

what kind of compiler flags & features are your original version compiled for/with? sse? what minimum cpu?

- ccernn

and, yes, just host the files. permission granted :)

Post Reply

Return to “Effects”