Thread: [Audacity-devel] Patch that adds WASAPI loopback support
A free multi-track audio editor and recorder
Brought to you by:
aosiniao
From: Leland <le...@au...> - 2012-12-28 23:44:45
|
Attached is a patch that adds loopback support to the WASAPI portaudio host along with some pics of it actually working. In the second pic, the devices with "(loopback)" after the name are just duplicates of the playback devices, but they'll be handled as if they were capture devices. There is still one issue that I know of which needs to be resolved. If the device expects to capture stereo data, then you'd better record in stereo else you'll crash. Haven't tracked this one down yet. Leland |
From: Leland L. <ll...@ho...> - 2012-12-29 07:36:23
Attachments:
wasapi-loopback2.patch
|
On 12/28/2012 5:44 PM, Leland wrote: > Attached is a patch that adds loopback support to the WASAPI portaudio > host along with some pics of it actually working. In the second pic, > the devices with "(loopback)" after the name are just duplicates of the > playback devices, but they'll be handled as if they were capture devices. > > There is still one issue that I know of which needs to be resolved. If > the device expects to capture stereo data, then you'd better record in > stereo else you'll crash. Haven't tracked this one down yet. > The attached patch corrects the stereo<->mono issue and should now be fully functional. Well, it works here on an RME Babyface (USB) and the internal SB Recon3D. Leland |
From: Gale A. <ga...@au...> - 2012-12-29 22:10:06
|
| From Leland Lucius <ll...@ho...> | Sat, 29 Dec 2012 01:36:23 -0600 | Subject: [Audacity-devel] Patch that adds WASAPI loopback support > On 12/28/2012 5:44 PM, Leland wrote: > > Attached is a patch that adds loopback support to the WASAPI portaudio > > host along with some pics of it actually working. In the second pic, > > the devices with "(loopback)" after the name are just duplicates of the > > playback devices, but they'll be handled as if they were capture devices. > > > > There is still one issue that I know of which needs to be resolved. If > > the device expects to capture stereo data, then you'd better record in > > stereo else you'll crash. Haven't tracked this one down yet. > > > The attached patch corrects the stereo<->mono issue and should now be > fully functional. Well, it works here on an RME Babyface (USB) and the > internal SB Recon3D. Thanks for this, Leland. So this V2 loopback patch goes on top of the three WASAPI patches listed here: http://audacity.238276.n2.nabble.com/Fwd-Would-you-guys-be-interested-in-td7556901.html#a7556908 ? Gale |
From: Leland L. <ll...@ho...> - 2012-12-29 22:26:51
|
On 12/29/2012 4:09 PM, Gale Andrews wrote: > > | From Leland Lucius <ll...@ho...> > | Sat, 29 Dec 2012 01:36:23 -0600 > | Subject: [Audacity-devel] Patch that adds WASAPI loopback support >> On 12/28/2012 5:44 PM, Leland wrote: >>> Attached is a patch that adds loopback support to the WASAPI portaudio >>> host along with some pics of it actually working. In the second pic, >>> the devices with "(loopback)" after the name are just duplicates of the >>> playback devices, but they'll be handled as if they were capture devices. >>> >>> There is still one issue that I know of which needs to be resolved. If >>> the device expects to capture stereo data, then you'd better record in >>> stereo else you'll crash. Haven't tracked this one down yet. >>> >> The attached patch corrects the stereo<->mono issue and should now be >> fully functional. Well, it works here on an RME Babyface (USB) and the >> internal SB Recon3D. > > Thanks for this, Leland. > > So this V2 loopback patch goes on top of the three WASAPI patches > listed here: > http://audacity.238276.n2.nabble.com/Fwd-Would-you-guys-be-interested-in-td7556901.html#a7556908 ? > Yes. Maybe I should rework them all and make one patch??? Leland |
From: Martyn S. <mar...@gm...> - 2012-12-31 21:20:34
|
Hi Leland On 29/12/2012 22:26, Leland Lucius wrote: > On 12/29/2012 4:09 PM, Gale Andrews wrote: >> >> | From Leland Lucius <ll...@ho...> >> | Sat, 29 Dec 2012 01:36:23 -0600 >> | Subject: [Audacity-devel] Patch that adds WASAPI loopback support >>> On 12/28/2012 5:44 PM, Leland wrote: >>>> Attached is a patch that adds loopback support to the WASAPI portaudio >>>> host along with some pics of it actually working. In the second pic, >>>> the devices with "(loopback)" after the name are just duplicates of the >>>> playback devices, but they'll be handled as if they were capture devices. >>>> >>>> There is still one issue that I know of which needs to be resolved. If >>>> the device expects to capture stereo data, then you'd better record in >>>> stereo else you'll crash. Haven't tracked this one down yet. >>>> >>> The attached patch corrects the stereo<->mono issue and should now be >>> fully functional. Well, it works here on an RME Babyface (USB) and the >>> internal SB Recon3D. >> >> Thanks for this, Leland. >> >> So this V2 loopback patch goes on top of the three WASAPI patches >> listed here: >> http://audacity.238276.n2.nabble.com/Fwd-Would-you-guys-be-interested-in-td7556901.html#a7556908 ? >> > Yes. Maybe I should rework them all and make one patch??? I followed those instructions but got these linker errors: 8>portmixer.lib(px_mixer.obj) : error LNK2019: unresolved external symbol _Pa_GetStreamHostApiType referenced in function _Px_OpenMixer 8>portmixer.lib(px_win_ds.obj) : error LNK2019: unresolved external symbol _PaWinDS_GetStreamOutputGUID referenced in function _OpenMixer_Win_DirectSound 8>portmixer.lib(px_win_ds.obj) : error LNK2019: unresolved external symbol _PaWinDS_GetStreamInputGUID referenced in function _OpenMixer_Win_DirectSound any ideas? (VS2008 Express) Thanks Martyn > Leland > > > > ------------------------------------------------------------------------------ > Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS, > MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current > with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft > MVPs and experts. SALE $99.99 this month only -- learn more at: > http://p.sf.net/sfu/learnmore_122912 > _______________________________________________ > audacity-devel mailing list > aud...@li... > https://lists.sourceforge.net/lists/listinfo/audacity-devel > |
From: Leland L. <ll...@ho...> - 2012-12-31 21:57:39
|
Yea, yea, Gale told me I was stoopid too. :-) I was just walking through everything again because Gale was hitting the same problem you did and I realized what was going on. Just apply the portaudio.patch found in lib-src/portmixer. That should get y'all going. A new patch will be coming shortly that will add JACK support as well. On 12/31/2012 3:20 PM, Martyn Shaw wrote: > Hi Leland > > On 29/12/2012 22:26, Leland Lucius wrote: >> On 12/29/2012 4:09 PM, Gale Andrews wrote: >>> | From Leland Lucius <ll...@ho...> >>> | Sat, 29 Dec 2012 01:36:23 -0600 >>> | Subject: [Audacity-devel] Patch that adds WASAPI loopback support >>>> On 12/28/2012 5:44 PM, Leland wrote: >>>>> Attached is a patch that adds loopback support to the WASAPI portaudio >>>>> host along with some pics of it actually working. In the second pic, >>>>> the devices with "(loopback)" after the name are just duplicates of the >>>>> playback devices, but they'll be handled as if they were capture devices. >>>>> >>>>> There is still one issue that I know of which needs to be resolved. If >>>>> the device expects to capture stereo data, then you'd better record in >>>>> stereo else you'll crash. Haven't tracked this one down yet. >>>>> >>>> The attached patch corrects the stereo<->mono issue and should now be >>>> fully functional. Well, it works here on an RME Babyface (USB) and the >>>> internal SB Recon3D. >>> Thanks for this, Leland. >>> >>> So this V2 loopback patch goes on top of the three WASAPI patches >>> listed here: >>> http://audacity.238276.n2.nabble.com/Fwd-Would-you-guys-be-interested-in-td7556901.html#a7556908 ? >>> >> Yes. Maybe I should rework them all and make one patch??? > I followed those instructions but got these linker errors: > 8>portmixer.lib(px_mixer.obj) : error LNK2019: unresolved external > symbol _Pa_GetStreamHostApiType referenced in function _Px_OpenMixer > 8>portmixer.lib(px_win_ds.obj) : error LNK2019: unresolved external > symbol _PaWinDS_GetStreamOutputGUID referenced in function > _OpenMixer_Win_DirectSound > 8>portmixer.lib(px_win_ds.obj) : error LNK2019: unresolved external > symbol _PaWinDS_GetStreamInputGUID referenced in function > _OpenMixer_Win_DirectSound > > any ideas? (VS2008 Express) > > Thanks > Martyn > >> Leland >> >> >> >> ------------------------------------------------------------------------------ >> Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS, >> MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current >> with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft >> MVPs and experts. SALE $99.99 this month only -- learn more at: >> http://p.sf.net/sfu/learnmore_122912 >> _______________________________________________ >> audacity-devel mailing list >> aud...@li... >> https://lists.sourceforge.net/lists/listinfo/audacity-devel >> > ------------------------------------------------------------------------------ > Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS, > MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current > with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft > MVPs and experts. SALE $99.99 this month only -- learn more at: > http://p.sf.net/sfu/learnmore_122412 > _______________________________________________ > audacity-devel mailing list > aud...@li... > https://lists.sourceforge.net/lists/listinfo/audacity-devel > |
From: Gale A. <ga...@au...> - 2013-01-01 04:05:37
|
| From Leland Lucius <ll...@ho...> | Mon, 31 Dec 2012 15:57:46 -0600 | Subject: [Audacity-devel] Patch that adds WASAPI loopback support > Yea, yea, Gale told me I was stoopid too. :-) I was just walking > through everything again because Gale was hitting the same problem you > did and I realized what was going on. > > Just apply the portaudio.patch found in lib-src/portmixer. That should > get y'all going. > > A new patch will be coming shortly that will add JACK support as well. The below are all findings for Unicode Release build. I had applied the portaudio patch in lib-src/portmixer (I had to do it manually which was tedious). That all built and linked fine, and the resulting EXE (without loopback recording) seems to record and play OK for the four hosts (I didn't enable ASIO). A few points. * The EXE takes ~25 seconds to launch (2.4 GHz, dual core 6 GB RAM, Windows 7 x64) as opposed to ~10 seconds for HEAD. * Quitting the EXE still leaves the Audacity process running, which has to be killed in Task Manager. * WDM-KS renames the sound device for the three built-in inputs from "IDT High Def" to "MuxedIn1" or "MuxedIn2". * WDM-KS doesn't see the stereo mix or line-in of my USB sound card. This is "Trust Audio External" with Windows Vista drivers. These are the latest drivers for my particular model and work fine under the other three hosts. Product link: http://tinyurl.com/axefphz . Then I applied the "wasapi-loopback2.patch" on top. That patched, built and linked apparently OK. However when I launch it, nothing happens for about 30 seconds, then the process quits of its own accord. I guess some of this could be if I patched for portmixer incorrectly. Also if I am running PC Tools "Spyware Doctor with AntiVirus", launching the loopback-enabled build of Audacity makes PC Tools ask if I should allow Audacity to connect to the internet. But that is most likely a problem with PC Tools. I'll try "wasapi-jack-portmixer.patch" next but I'll wait a bit in case of any reaction to the above. Thanks, Leland! Gale > On 12/31/2012 3:20 PM, Martyn Shaw wrote: > > Hi Leland > > > > On 29/12/2012 22:26, Leland Lucius wrote: > >> On 12/29/2012 4:09 PM, Gale Andrews wrote: > >>> | From Leland Lucius <ll...@ho...> > >>> | Sat, 29 Dec 2012 01:36:23 -0600 > >>> | Subject: [Audacity-devel] Patch that adds WASAPI loopback support > >>>> On 12/28/2012 5:44 PM, Leland wrote: > >>>>> Attached is a patch that adds loopback support to the WASAPI portaudio > >>>>> host along with some pics of it actually working. In the second pic, > >>>>> the devices with "(loopback)" after the name are just duplicates of the > >>>>> playback devices, but they'll be handled as if they were capture devices. > >>>>> > >>>>> There is still one issue that I know of which needs to be resolved. If > >>>>> the device expects to capture stereo data, then you'd better record in > >>>>> stereo else you'll crash. Haven't tracked this one down yet. > >>>>> > >>>> The attached patch corrects the stereo<->mono issue and should now be > >>>> fully functional. Well, it works here on an RME Babyface (USB) and the > >>>> internal SB Recon3D. > >>> Thanks for this, Leland. > >>> > >>> So this V2 loopback patch goes on top of the three WASAPI patches > >>> listed here: > >>> http://audacity.238276.n2.nabble.com/Fwd-Would-you-guys-be-interested-in-td7556901.html#a7556908 ? > >>> > >> Yes. Maybe I should rework them all and make one patch??? > > I followed those instructions but got these linker errors: > > 8>portmixer.lib(px_mixer.obj) : error LNK2019: unresolved external > > symbol _Pa_GetStreamHostApiType referenced in function _Px_OpenMixer > > 8>portmixer.lib(px_win_ds.obj) : error LNK2019: unresolved external > > symbol _PaWinDS_GetStreamOutputGUID referenced in function > > _OpenMixer_Win_DirectSound > > 8>portmixer.lib(px_win_ds.obj) : error LNK2019: unresolved external > > symbol _PaWinDS_GetStreamInputGUID referenced in function > > _OpenMixer_Win_DirectSound > > > > any ideas? (VS2008 Express) > > > > Thanks > > Martyn > > > >> Leland |
From: Leland L. <ll...@ho...> - 2013-01-01 05:57:06
|
On 12/31/2012 10:05 PM, Gale Andrews wrote: > > | From Leland Lucius <ll...@ho...> > | Mon, 31 Dec 2012 15:57:46 -0600 > | Subject: [Audacity-devel] Patch that adds WASAPI loopback support >> Yea, yea, Gale told me I was stoopid too. :-) I was just walking >> through everything again because Gale was hitting the same problem you >> did and I realized what was going on. >> >> Just apply the portaudio.patch found in lib-src/portmixer. That should >> get y'all going. >> >> A new patch will be coming shortly that will add JACK support as well. > > The below are all findings for Unicode Release build. > > I had applied the portaudio patch in lib-src/portmixer (I had to do it > manually which was tedious). > > That all built and linked fine, and the resulting EXE (without loopback > recording) seems to record and play OK for the four hosts (I didn't > enable ASIO). > > A few points. > > * The EXE takes ~25 seconds to launch (2.4 GHz, dual core 6 GB > RAM, Windows 7 x64) as opposed to ~10 seconds for HEAD. > > * Quitting the EXE still leaves the Audacity process running, which > has to be killed in Task Manager. > > * WDM-KS renames the sound device for the three built-in inputs > from "IDT High Def" to "MuxedIn1" or "MuxedIn2". > > * WDM-KS doesn't see the stereo mix or line-in of my USB sound > card. This is "Trust Audio External" with Windows Vista drivers. > These are the latest drivers for my particular model and work > fine under the other three hosts. Product link: > http://tinyurl.com/axefphz . > > Then I applied the "wasapi-loopback2.patch" on top. That patched, > built and linked apparently OK. > > However when I launch it, nothing happens for about 30 seconds, > then the process quits of its own accord. > > I guess some of this could be if I patched for portmixer incorrectly. > > Also if I am running PC Tools "Spyware Doctor with AntiVirus", > launching the loopback-enabled build of Audacity makes PC Tools > ask if I should allow Audacity to connect to the internet. But that > is most likely a problem with PC Tools. > > I'll try "wasapi-jack-portmixer.patch" next but I'll wait a bit in > case of any reaction to the above. > Definitely try the wasapi-jack-portmixer.patch since most of what you describe shouldn't be happening at all, especially the Spyware Doctor thing (unless you applied the JACK patch as well). The WDM-KS renaming of your devices thing happens here too on my internal SB Recon3Di (but not my RME): MME name WDM-KS name -------------------------- ---------------------------------- SPDIF-Out (SB Recon3Di) SPDIF Out (Malcolm DigitalOutWave) Speakers (SB Recon3Di) Speakers (Malcolm Lineout/Headphone) Microphone (SB Recon3Di) Microphone (Malcolm Mic1Wave) What U Hear (SB Recon3Di) What U Hear (Malcolm WUHInWave) Not sure what the deal with that is, but I can see the WDM-KS name in the .inf file for my device. Didn't see your names in the .inf for your device, but if you run DxDiag and then click "Save All Information" at the bottom, you'll get a report that shows the same names as you've mentioned above. Could be they're some sort of product development names or where the hardware vendor sourced the drivers from or, well, who knows really. A couple of quick googles didn't produce any answers. Now, in regards to your normal 10 second startup. We have slow startups here too. Have y'all an idea of why this happens? I've looked into it a little and it seems to be related to i18n, but I didn't dig very hard yet. Leland |
From: Gale A. <ga...@au...> - 2013-01-02 05:22:12
|
Hi Leland, I enjoyed my excursion into Audacity loopback recording. Most of this feedback will probably only interest Leland, so just to summarise that the only bad issue I found so far is that I always get "error opening" under WASAPI after I change "Default Format" in Windows. Details below. On the more general question: > Now, in regards to your normal 10 second startup. We have slow startups > here too. Have y'all an idea of why this happens? I've looked into it > a little and it seems to be related to i18n, but I didn't dig very hard yet. On Ubuntu 12.10 on my laptop, Audacity HEAD starts in two seconds, and it's almost as quick on OS X (Mac Mini), so other things being equal it must be Windows-specific. It is also related to machine speed. On slow Windows machines you can wait up to 45 seconds now, so I guess that will be extended further by this patch. | From Leland Lucius <ll...@ho...> | Mon, 31 Dec 2012 23:57:12 -0600 | Subject: [Audacity-devel] Patch that adds WASAPI loopback support > On 12/31/2012 10:05 PM, Gale Andrews wrote: > > > > | From Leland Lucius <ll...@ho...> > > | Mon, 31 Dec 2012 15:57:46 -0600 > > | Subject: [Audacity-devel] Patch that adds WASAPI loopback support > >> Yea, yea, Gale told me I was stoopid too. :-) I was just walking > >> through everything again because Gale was hitting the same problem you > >> did and I realized what was going on. > >> > >> Just apply the portaudio.patch found in lib-src/portmixer. That should > >> get y'all going. > >> > >> A new patch will be coming shortly that will add JACK support as well. > > > > The below are all findings for Unicode Release build. > > > > I had applied the portaudio patch in lib-src/portmixer (I had to do it > > manually which was tedious). > > > > That all built and linked fine, and the resulting EXE (without loopback > > recording) seems to record and play OK for the four hosts (I didn't > > enable ASIO). > > > > A few points. > > > > * The EXE takes ~25 seconds to launch (2.4 GHz, dual core 6 GB > > RAM, Windows 7 x64) as opposed to ~10 seconds for HEAD. > > > > * Quitting the EXE still leaves the Audacity process running, which > > has to be killed in Task Manager. > > > > * WDM-KS renames the sound device for the three built-in inputs > > from "IDT High Def" to "MuxedIn1" or "MuxedIn2". > > > > * WDM-KS doesn't see the stereo mix or line-in of my USB sound > > card. This is "Trust Audio External" with Windows Vista drivers. > > These are the latest drivers for my particular model and work > > fine under the other three hosts. Product link: > > http://tinyurl.com/axefphz . > > > > Then I applied the "wasapi-loopback2.patch" on top. That patched, > > built and linked apparently OK. > > > > However when I launch it, nothing happens for about 30 seconds, > > then the process quits of its own accord. > > > > I guess some of this could be if I patched for portmixer incorrectly. > > > > Also if I am running PC Tools "Spyware Doctor with AntiVirus", > > launching the loopback-enabled build of Audacity makes PC Tools > > ask if I should allow Audacity to connect to the internet. But that > > is most likely a problem with PC Tools. > > > > I'll try "wasapi-jack-portmixer.patch" next but I'll wait a bit in > > case of any reaction to the above. > > > Definitely try the wasapi-jack-portmixer.patch since most of what you > describe shouldn't be happening at all, especially the Spyware Doctor > thing (unless you applied the JACK patch as well). OK, I applied the "wasapi-jack-portmixer.patch" to a fresh HEAD checkout, but only enabled DirectSound of the optional choices. I will add JACK later and try that. First off, I don't get the refusal to quit/refusal to launch problems (though Spyware Doctor still asked about internet access for Audacity first time I launched the newly built Audacity). I note you have a new patch "saved hostname was being used for all interface types". I'm not sure what that symptom is, and have not tried that patch yet. I have a problem not noticed before (unrelated to your patch) with my inbuilt IDT sound device. Almost whenever I change Host, the input and output reverts to its default (so I can never retain the inbuilt stereo mix input when I change Host). This is a recent OEM computer so the IDT drivers are correct, but this was not a problem with the previous Windows 7 computer. Any ideas, while you are working on this? Another issue that arose with Michael's Device Toolbar changes in 1.3.13 is that on initialised cfg, the MME Input and Output device is no longer Sound Mapper but a specific input (usually internal mic and speakers respectively). This is an issue for VI users. If the Windows defaults are set to other than the devices Audacity launches with, VI users nay be confused that they hear no audio input or output in Audacity. Do you fancy trying to set the defaults back to Sound Mapper? Otherwise I think it should go on Bugzilla. > For the WASAPI loopback examples, I had to record in stereo since > the SB Recon3D didn't seem to want to do mono. When using the RME, > I'm able to choose mono or stereo, so I'm not sure what's going on here. On first launch of the EXE with initialised .cfg I had the opposite issue. The built-in device would not start loopback recording in stereo; the red recording cursor just hung at zero. It wasn't a sample rate or other issue that I can identify, and did not happen again on further launches even on initialised .cfg. I noticed two characteristics of loopback recording (both with the USB and built-in device), maybe not very important. * When first used in an Audacity session, loopback recording hangs the recording cursor if no audio stream is playing. Once you have recorded some loopback audio then it will record low level noise if there is no stream. Sound device "stereo mix" always records low level noise if there is no stream. * Loopback recording from an Audacity track always hangs the recording cursor unless overdub is on. Am I correct Windows DirectSound is still the only host that guarantees no resampling (if set "Exclusive" mode in Windows)? > Much better results can be had from the WDM-KS host and even some > improvement for ASIO by adjusting the "Latency->Audio to buffer" value > in preferences. I can even get ASIO and KS to produce almost the same > latency if I specify a 2ms buffer in preferences...no clicks/stuttering/popping/ > etc. Yes very impressive, none of WASAPI, MME or Windows Direct Sound will record at that buffer setting. I did find one repeatable issue specific to WASAPI (both with built-in and USB device). If I change the "Default Format" in Windows for a particular input while under WASAPI Host, I always get "error opening input device" when recording from that input. This is so even if the Audacity rate matches Default Format when not in "Exclusive Mode". The error can be fixed by refreshing the device list (Rescan Audio Devices, or open a drop-down in Device Toolbar and click on an entry, even if you do not change it). Similarly if I change Default Format for an output device, that device errors until the device list is refreshed. Gale |
From: Leland L. <ll...@ho...> - 2013-01-03 07:10:10
|
Hi Gale, On 1/1/2013 11:21 PM, Gale Andrews wrote: >> Now, in regards to your normal 10 second startup. We have slow startups >> here too. Have y'all an idea of why this happens? I've looked into it >> a little and it seems to be related to i18n, but I didn't dig very hard yet. > > On Ubuntu 12.10 on my laptop, Audacity HEAD starts in two seconds, > and it's almost as quick on OS X (Mac Mini), so other things being > equal it must be Windows-specific. > > It is also related to machine speed. On slow Windows machines you > can wait up to 45 seconds now, so I guess that will be extended > further by this patch. > I'll get to your other questions and observations on the portaudio changes in a bit, but this slow starting thing is driving me bonkers. Can you check something on one of your machines where the startup is slow? Take a look at your audacity.cfg and tell me what the Locale/Language is set to? If it is, change it to "en" and try starting Audacity to see if there's a difference. If it starts faster, wait for a few minutes and try again. Also, does that machine have many VST and/or Ladspa effects installed? Thanks, Leland |
From: Gale A. <ga...@au...> - 2013-01-08 13:18:46
|
| From Leland Lucius <ll...@ho...> | Thu, 03 Jan 2013 01:10:22 -0600 | Subject: [Audacity-devel] Patch that adds WASAPI loopback support > Hi Gale, > > On 1/1/2013 11:21 PM, Gale Andrews wrote: > >> Now, in regards to your normal 10 second startup. We have slow startups > >> here too. Have y'all an idea of why this happens? I've looked into it > >> a little and it seems to be related to i18n, but I didn't dig very hard yet. > > > > On Ubuntu 12.10 on my laptop, Audacity HEAD starts in two seconds, > > and it's almost as quick on OS X (Mac Mini), so other things being > > equal it must be Windows-specific. > > > > It is also related to machine speed. On slow Windows machines you > > can wait up to 45 seconds now, so I guess that will be extended > > further by this patch. > > > I'll get to your other questions and observations on the portaudio > changes in a bit, but this slow starting thing is driving me bonkers. [...] > Also, does that machine have many VST and/or Ladspa effects installed? I did a bit of experimenting on that (the slowest machine I have now is the netbook with Win 7 Starter). The figures I've been quoting for start times are with minimal VST or LADSPA effects. The quick answer is no discernible difference trying about 300 LADSPA DLL's. 300 VST's adds about 2 seconds to the Desktop launch and 5 seconds to the netbook launch (assuming of course I am not rescanning). I'm not quite clear from: http://bugzilla.audacityteam.org/show_bug.cgi?id=368 whether we are loading VST's on launch or not when we don't scan. I'm figuring we are loading, because people with unlicensed plug-ins always get to see the nag screens every Audacity launch, scan or not. Gale |
From: Leland L. <ll...@ho...> - 2013-01-08 14:52:55
|
On 1/8/2013 7:18 AM, Gale Andrews wrote: > > | From Leland Lucius <ll...@ho...> > | Thu, 03 Jan 2013 01:10:22 -0600 > | Subject: [Audacity-devel] Patch that adds WASAPI loopback support >> Hi Gale, >> >> On 1/1/2013 11:21 PM, Gale Andrews wrote: >>>> Now, in regards to your normal 10 second startup. We have slow startups >>>> here too. Have y'all an idea of why this happens? I've looked into it >>>> a little and it seems to be related to i18n, but I didn't dig very hard yet. >>> >>> On Ubuntu 12.10 on my laptop, Audacity HEAD starts in two seconds, >>> and it's almost as quick on OS X (Mac Mini), so other things being >>> equal it must be Windows-specific. >>> >>> It is also related to machine speed. On slow Windows machines you >>> can wait up to 45 seconds now, so I guess that will be extended >>> further by this patch. >>> >> I'll get to your other questions and observations on the portaudio >> changes in a bit, but this slow starting thing is driving me bonkers. > [...] >> Also, does that machine have many VST and/or Ladspa effects installed? > > I did a bit of experimenting on that (the slowest machine I have now > is the netbook with Win 7 Starter). The figures I've been quoting for > start times are with minimal VST or LADSPA effects. > > The quick answer is no discernible difference trying about 300 > LADSPA DLL's. > > 300 VST's adds about 2 seconds to the Desktop launch and 5 seconds > to the netbook launch (assuming of course I am not rescanning). > > I'm not quite clear from: > http://bugzilla.audacityteam.org/show_bug.cgi?id=368 > > whether we are loading VST's on launch or not when we don't scan. > > I'm figuring we are loading, because people with unlicensed plug-ins > always get to see the nag screens every Audacity launch, scan or not. > Yep, looks like Audio Units, VSTs, and Ladspa plugins are loaded at each launch. Only VSTs that have been previously scanned will be loaded. Any misbehaving AUs or Ladspas can cause Audacity issues during startup. Audacity "shouldn't" crash at startup due to misbehaving VSTs, but I could be wrong about that. Leland |
From: Leland L. <ll...@ho...> - 2013-01-03 15:16:39
|
On 1/3/2013 1:10 AM, Leland Lucius wrote: > Hi Gale, > > On 1/1/2013 11:21 PM, Gale Andrews wrote: >>> Now, in regards to your normal 10 second startup. We have slow startups >>> here too. Have y'all an idea of why this happens? I've looked into it >>> a little and it seems to be related to i18n, but I didn't dig very hard yet. >> >> On Ubuntu 12.10 on my laptop, Audacity HEAD starts in two seconds, >> and it's almost as quick on OS X (Mac Mini), so other things being >> equal it must be Windows-specific. >> >> It is also related to machine speed. On slow Windows machines you >> can wait up to 45 seconds now, so I guess that will be extended >> further by this patch. >> > I'll get to your other questions and observations on the portaudio > changes in a bit, but this slow starting thing is driving me bonkers. > > Can you check something on one of your machines where the startup is > slow? Take a look at your audacity.cfg and tell me what the > Locale/Language is set to? If it is, change it to "en" and try starting > Audacity to see if there's a difference. If it starts faster, wait for > a few minutes and try again. Nevermind Gale, wild goose chase, but I haven't given up yet!!! :-) Leland |
From: Steve t. F. <ste...@gm...> - 2013-01-07 05:06:43
|
Probably a question for Leland: When Audacity supports WASAPI, will Audacity (with appropriate hardware) support 24 bit recording on Windows as opposed to 16 bit padded to 24 bit? Steve On 3 January 2013 15:16, Leland Lucius <ll...@ho...> wrote: > On 1/3/2013 1:10 AM, Leland Lucius wrote: >> Hi Gale, >> >> On 1/1/2013 11:21 PM, Gale Andrews wrote: >>>> Now, in regards to your normal 10 second startup. We have slow startups >>>> here too. Have y'all an idea of why this happens? I've looked into it >>>> a little and it seems to be related to i18n, but I didn't dig very hard yet. >>> >>> On Ubuntu 12.10 on my laptop, Audacity HEAD starts in two seconds, >>> and it's almost as quick on OS X (Mac Mini), so other things being >>> equal it must be Windows-specific. >>> >>> It is also related to machine speed. On slow Windows machines you >>> can wait up to 45 seconds now, so I guess that will be extended >>> further by this patch. >>> >> I'll get to your other questions and observations on the portaudio >> changes in a bit, but this slow starting thing is driving me bonkers. >> >> Can you check something on one of your machines where the startup is >> slow? Take a look at your audacity.cfg and tell me what the >> Locale/Language is set to? If it is, change it to "en" and try starting >> Audacity to see if there's a difference. If it starts faster, wait for >> a few minutes and try again. > > Nevermind Gale, wild goose chase, but I haven't given up yet!!! :-) > > Leland > > > > ------------------------------------------------------------------------------ > Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS, > MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current > with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft > MVPs and experts. ON SALE this month only -- learn more at: > http://p.sf.net/sfu/learnmore_122712 > _______________________________________________ > audacity-devel mailing list > aud...@li... > https://lists.sourceforge.net/lists/listinfo/audacity-devel |
From: Leland L. <ll...@ho...> - 2013-01-07 05:21:27
|
On 1/6/2013 11:06 PM, Steve the Fiddle wrote: > Probably a question for Leland: > When Audacity supports WASAPI, will Audacity (with appropriate > hardware) support 24 bit recording on Windows as opposed to 16 bit > padded to 24 bit? > It basically will since the WASAPI or the WDMKS portaudio hostapis do not force the sample format to 16-bit like the MME and DirectSound fellas do. But, from what I've read, when using WASAPI in shared mode (the default for portaudio), Windows will convert the samples from the sound card to float internally, do whatever processing it deems necessary, and then convert it back to the requested format. WDMKS or WASAPI exclusive mode do not do those conversions. Either way, you're definitely going to get better than 16-bit if the sound cards supports it. Leland |
From: Leland L. <ll...@ho...> - 2013-01-07 05:25:15
|
On 1/6/2013 11:21 PM, Leland Lucius wrote: > On 1/6/2013 11:06 PM, Steve the Fiddle wrote: >> Probably a question for Leland: >> When Audacity supports WASAPI, will Audacity (with appropriate >> hardware) support 24 bit recording on Windows as opposed to 16 bit >> padded to 24 bit? >> > It basically will since the WASAPI or the WDMKS portaudio hostapis do > not force the sample format to 16-bit like the MME and DirectSound > fellas do. > > But, from what I've read, when using WASAPI in shared mode (the default > for portaudio), Windows will convert the samples from the sound card to > float internally, do whatever processing it deems necessary, and then > convert it back to the requested format. WDMKS or WASAPI exclusive mode > do not do those conversions. > > Either way, you're definitely going to get better than 16-bit if the > sound cards supports it. It also depends on the default format you select for shared mode in the mixer app: |