oaCapture for OSX

[JamesF]

This was on the apple usb list:

That looks unpleasant.  I'm not really sure if it covers my use case, but it certainly looks like Apple have been mucking about with the low-level USB stuff and potentially broken something.

James

[JamesF]

hi james will my point grey blackly gige work with this software I'm using mountain lion

mark

Oooh, GigE Vision looks like a bundle of laughs.  The protocol is claimed to be "open", where "open" appears to have been redefined to require yearly fees to obtain a licence and an undertaking to never disclose the details to anyone who doesn't have a licence.  Which sounds a lot like "closed" to me.

There is a library that I might be able to use (though how portable it is remains to be seen) and I do have a somewhat beaten up PGR GigE camera so I will have a crack at getting them to work at some point, but I don't know quite where that point is for the moment.

James

[Davey-T]

274 still working OK today, still no Sun, if the software works as is ATM it is perfectly usable and I can put up with the slow startup and occasional forced quit, probably more annoying for you than me.

Dave

[NickK]

That looks unpleasant.  I'm not really sure if it covers my use case, but it certainly looks like Apple have been mucking about with the low-level USB stuff and potentially broken something.

James

My thinking too. I noted that there was some changes in things - mainly timings etc. Unfortunately with USB2 it's a host initiated polling protocol, so if the device doesn't follow that protocol.. you'll be using unreliable blind waiting.

[msh1]

Oooh, GigE Vision looks like a bundle of laughs.  The protocol is claimed to be "open", where "open" appears to have been redefined to require yearly fees to obtain a licence and an undertaking to never disclose the details to anyone who doesn't have a licence.  Which sounds a lot like "closed" to me.

There is a library that I might be able to use (though how portable it is remains to be seen) and I do have a somewhat beaten up PGR GigE camera so I will have a crack at getting them to work at some point, but I don't know quite where that point is for the moment.

James

ah well james as and when would be great tho

mark

[JamesF]

274 still working OK today, still no Sun, if the software works as is ATM it is perfectly usable and I can put up with the slow startup and occasional forced quit, probably more annoying for you than me.

Dave

It certainly is rather irritating.  And changing the cable for a much shorter one doesn't help.  USB2 hub next.

James

[JamesF]

And with a USB2 hub there is absolutely no delay on startup on both Mavericks and Yosemite on my MBP.  I still get the hang on exit, but I'm struggling not to see the slow start as a hardware incompatibility/OS level issue.

James

[JamesF]

Hah!  I just put in some debugging code to try to track down the cause of the hang on exit.  Just printing out which functions were being called when the quit button is pressed.  And now I can't get it to fail

I suspect that probably means this particular issue is down to my code.

James

[JamesF]

This was on the apple usb list:

...

Also, interestingly, assuming I queue a large number of these (all

calls return success) and immediately abort the pipe, only a small

handful of those are returned to the completion handler, the rest

'disappear'. The also feels new and unexpected.

...

I think it's possible I'm being bitten by something relating to this and it's what might well be causing the lock-up on exit.

Basically the UVC code has a queue of bulk transfers and at shutdown time cancels the transfers, waiting for the "transfer cancelled" notification to arrive before freeing the resources associated with each transfer.  But some of the notifications look like they never arrive so the application is waiting forever

I'll look into whether this is fixable if I cancel the transfers a little more slowly, but it's an ugly hack.

James

[NickK]

I think it's possible I'm being bitten by something relating to this and it's what might well be causing the lock-up on exit.

Basically the UVC code has a queue of bulk transfers and at shutdown time cancels the transfers, waiting for the "transfer cancelled" notification to arrive before freeing the resources associated with each transfer.  But some of the notifications look like they never arrive so the application is waiting forever

I'll look into whether this is fixable if I cancel the transfers a little more slowly, but it's an ugly hack.

James

You could use a ReadPipeTO or ReadPipeAsyncTO as this is the time out variant.

Note one other "fun" thing with the Mac USB implementation is that it returns packets.. but if there is one byte missing from the data block.. it won't return (unless Read..TO) nor will it return *any* of the data. This makes flushing buffers something of a nightmare on OS X. It's best to start with a largish buffer then if that fails, drop down in chunks until you're at a byte read.. once that's failed then you're flushed.

[JamesF]

Haha!  I have it licked (I hope

It appears that it is possible for several valid completed USB transfers to be delivered to my callback handler even after I've cancelled every single one of them.  That's not something I've seen happen before either on Linux or OSX.

Now I just have the start-up issue to deal with.  I'd be tempted to overlook it for a while were it not for the fact that it ends up with the colour Skyris cameras not being correctly identified.

James

[JamesF]

Don't know if you want / need suggestions but a pause button between the start / stop buttons ala Firecapture would be useful for Solar imagers for when the clouds fly by.

If the Sun puts in an appearance I will try a real life go, peeing down and blowing a gale ATM

Just for you, this is now done and will be in the next release

James

[Davey-T]

Just for you, this is now done and will be in the next release

James

Thankyou James, still haven't managed to try it in the field yet.

Dave

[JamesF]

I've not been out in ages myself.  We've had quite a few clear nights here recently, but the wind has been bitingly cold with wind chill figures getting close to double figures below zero at night and barely reaching zero even during the day).  I'm finding little to recommend being out in such weather.

James

[Davey-T]

Yes. I put the Skyris 274 on the 10" SCT to have a go at Jupiter but been too windy as the fork mount doesn't like the slightest breeze

Dave

[reznap]

Greetings James et al., not sure if this is the appropriate thread for this but thought I'd post here.  I'm new to this hobby and came across oaCapture, great substitute for substandard or nonexistent vendor apps, thanks for your efforts James!  I just picked up a NexImage Burst C (Aptina AR0132) running on OS X 10.11.3 and find that I'm only getting grayscale video.  It seems like all the camera controls work, it just displays and records in grayscale.  I did verify this is in fact the color and not monochrome version by testing on a PC, it displays color there just fine.  oaCapture see the camera as: (EUVC) Celestron NexImage Burst C

Is there anything I can do in order to get color working?  I'm happy to do any testing that might help.

Also, I know this software is geared toward video but would it be possible to add an option to capture single frames (or even capture a frame every n seconds) for those of us interested in getting snapshots?

Thanks again for your hard work!

[JamesF]

It could be that you're looking at the raw colour output (which probably will appear to be monochrome).  That's the way the data comes off the sensor and probably the best format to save it as because generating the colour frames using a decent algorithm is sufficiently slow that it might impact performance.  Some stacking applications will do this for you (eg. AutoStakkert!2), but if you need the data written in colour then you can use the demosaic item in the "Options" and "Settings" menus to enable it.

If you've done that and it still doesn't appear to be working, please let me know.

James

[reznap]

Silly me, selecting Demosaic under options did the trick.  Thanks for setting me straight.

Now that color is working I see there is a TIFF option for saving images like I was hoping for (I had only AVI and SER to pick from previously).  When I try to save as TIFF I get the error "Output is not writable" even though the destination directory will take AVI files.

Any thoughts?

[JamesF]

I believe I have a fix for that problem.

I am currently waiting for Apple to release 10.11.4.  They broke the USB subsystem when they re-wrote it for El Capitan and this next release allegedly fixes a number of issues that prevent cameras working properly, or prevent them from working at all.

If they don't get it released soon however I'll make a new release based on what's available now and I hope that should fix the problem for you.

James

[reznap]

Sounds good to me, thanks for the update.  On a related note, I noticed that libpng16.16.dylib is already inside the .app.  Assuming 10.11.4 fixes things, does this mean PNG will also be available for snapshots rather than only TIFF?  Expanding on that, will there be any possibility of adding an H.264 option for the video encoding?  I know it's lossy but I think many people would accept the tradeoff of reduced file size.  With hardware encoding (Quick Sync on Intel, VCE on AMD) becoming more common it seems like there could be some real benefit.  Just some food for thought.

Thanks again James.

[rofus]

Hi James,

 

I just bought an ASI120mm (so the usb2), but for some reason I cannot get it to work with oacapture (I use Yosemite always updated to latest version). I choose it as I read it was working on Mac, but for some reason I see on ASI forum that others have problems sometimes?

It seems the software can see it, but cannot get any preview or record any frame from it. I tried as a test also downloading any of the previous versions of oacapture, and they just crash when I connect the camera.

I also tried the firmware update from ZWO website, and loading the normal and 'compatible' framework did not do any difference. Do you have any idea? Do you have a debug version where I can see and help understanding where is the problem? I have some experience as developer and might help if needed

Thanks for the fantastic work you've done with Oacapture!

[JamesF]

I'm working on a new release at the moment that fixes a number of bugs and uses the latest ASI SDK (and adds support for at least some Point Grey GigE cameras).  Testing has reached the point where it's painfully slow and cumbersome though (not helped by the fact that my childrens' competitive swimming has claimed a huge chunk of spare time in the last eight or nine months).  I think after this release I shall have to decide I've reached version 1.0, start doing beta releases for testing and hope others will be willing to take up some of the burden.

I will do some testing with Yosemite.  There have been some changes somewhere recently that have meant a single binary that used to work fine on OSX 10.6 through 10.10 stopped functioning correctly on later releases.  10.11 has also been a bag of problems, but I'm hoping 10.11.4 has resolved those.

Debugging with the ASI cameras can be awkward because we have no visibility of what's going on inside the SDK which is where you really want it.  I'm not sure what I can do to help there.

Anyhow, I shall have a new release out before the end of this month if it kills me

James

[rofus]

Thanks James, I can only imagine how painful is the whole process!

As I said if I can be useful for some testing (ASI120MM on Yosemite and latest Macbook Retina) please let me know and I will happily test for you (I did some debug in the past for SGP, BYN and Digicamcontrol and their problems with Nikon cameras and virtual machines).

[Avdhoeven]

Dear James, first of all thanks for this great program! I'm planning to use it to image the Mercury transit next Monday. I only encounter a problem and I don't know how to solve it properly.

I have an ASI 174mm with a macbook air (2013 edition with 128gb ssd, El Capitan OS (but problems were the same with yosemite)) and as long as I use a usb 2.0 hub the program works perfectly with the camera, although only at 35fps at 1000x1000 resolution. When I directly connect the camera to the USB3.0 the problems start. As soon as I change a setting like recording location or some other setting with the camera connected the camera seems to hang. Also the program becomes very unstable and I can record sometimes 1, sometimes 2-3 movies before I have to shut the program down and start all over.

I have decided to go for the lower speed for now, because Monday I want stability, but I really wonder if it's a problem with my computer, or maybe a broader problem?

 

Thanks!

[JamesF]

Do you have the latest updates for El Capitan installed and are you using the most recent release of oaCapture?

There are all sorts of USB problems with El Capitan releases before 10.11.4 and the latest oaCapture release includes some fixes that are specific to OSX so if you don't have one or the other of those that might help.  Otherwise I can try testing the 174MM again.  I don't recall seeing any problems, but it's possible I missed something.

James