startCapture(withOutputWidth:outputHeight:subframe:) Doesn't yield any frames on iOS SDKs

Help & Questions
1

Hello,

I’m curious to know in what cases startCapture(withOutputWidth:outputHeight:subframe:) fails to produce a frame. I’ve been playing with this API for quite some time now; the behavior is intermittent. Should there be any specific sequence it should be called?

Any details on this would be helpful.

Thanks

2

It wont produce a frame if deepar isn’t initialised or if there is no input frame. Are you getting different results?

3

I overcame this with some refactoring around. I think I started capturing earlier than starting the camera.

But I see a different issue now. If you confirm the following message is relevant to the DeepAR iOS SDK, I can create a new thread to discuss this further. I see The following statement on the console just after I start capturing the frames.

CVOpenGLESTextureCacheCreate failed (error: 0xffffe5fb)

and

2023-05-19 12:50:38.941394-0400 Together[34348:2919407] The input tensor width must be aligned to 64-byte.
2023-05-19 12:50:38.942519-0400 Together[34348:2919407] The input tensor width must be aligned to 64-byte.
2023-05-19 12:50:38.943789-0400 Together[34348:2919407] The input tensor width must be aligned to 64-byte.
2023-05-19 12:50:38.976624-0400 Together[34348:2919407] The input tensor width must be aligned to 64-byte.
2023-05-19 12:50:38.977837-0400 Together[34348:2919407] The input tensor width must be aligned to 64-byte.
2023-05-19 12:50:38.978983-0400 Together[34348:2919407] The input tensor width must be aligned to 64-byte.
4

Yes i think it is related to the SDK, but those warnings should be safe to ignore.

1 Like