I’ve just installed Audacity 2.1.2 from the Xubuntu repository (v16.04 LTS, Amd64) on a Dell XPS-17 laptop with Intel Core i7-2670QM processor 2.20 GHz with 8GB RAM and 2x 500MB HDDs. It has a 3GB NVIDIA GeForce GT 555M Graphics Card. There is a 12GB swap partition and a 27GB Linux system partition with more than 18GB free space. I am trying to record from an EZcap USB Cassette Capture audio cassette drive connected through a USB port on the PC which has a Li battery installed and is plugged into the mains power. I can get the setup to record OK but whatever I do I get short gaps in the recording of about 25ms each +/-50%, about 0.6secs apart. The tapes are old and of poor quality but there are no gap-clicks either listening to the cassette player or on Playthrough; only on the Audacity recorded data when it is replayed.
I’ve changed the audio buffer to 10ms and to 1000ms and switched off Playthrough but it makes no difference.
I’ve swapped the Open Source drivers for NVIDIA’s own, to no avail.
I’ve uninstalled the CPU monitor which refreshes every 250ms and that makes no difference.
Looking at the CPU trace, CPU usage for any core selom reaches 40% and most of the time only 1 or 2 cores are above 20%. Less than 750MB RAM is being used out of a total of 7.7GB, with no Swap activity. There are no spikes.
Attached is a screen shot of the audio trace from my last effort.
Any clues or ideas as to what I could try next would be most welcome.
Tks in anticipation
Tks for quick reply.
No I haven’t. Not sure how I’d go about that. I’ve looked for a traditional-style sound card but I keep coming back to the NVIDIA graphics card.
If I were to record a simple sonnd stream through the internal mike, would that work? What might be a suitable App to use?
As you will guess, this is a foreign land to me, full of dragons and bear-traps.
Chasing that down I note that according to Synaptic, I’m using the ALSA drivers v1.0.25 (installed from the Xubuntu distro) and according to http://www.alsa-project.org/main/index.php/Changes_v1.0.22_1.0.23
support for ALC655 was added in v1.0.23.
Your further guidance would be greatly appreciated. Meanwhile I’ll continue to poke about to see if I stumble on a solution.
Further to my first reply which is still with the mods I have done a test which may be helpful:
I started by downloading Gnome-sound-recorder and used the internal mic to record a piece from the radio. When I looked at the (.ogg) trace with Audacity there were no breaks although the amplitude of the trace was quite low. Encouraged by this I added an external mic (mono) and turned up the gain. When I recorded through Audacity using the HDA Intel PCH: ALC665 Analog (hw:0,0) setting in Edit/Preferences/Devices, there were no breaks (see attached traces).
This says to me that the problem lies with the handling of the digital input via the USB as the Playthrough sound is fine. I’ve tested all the USB ports on the PC and the result is the same.
I ran another test with exactly the same software build, but X86 rather than Amd64, with a much less capable CPU and no graphics card (Dell Inspiron-530 desktop) and got the same result. I also have a (rarely used) Win10 partition on the laptop. This morning I installed Audacity 2.1.2 which worked OK from the off with 44100Hz recording rate. That was going to have to be a very reluctant workaround until you struck gold.
I shall break out the Coke in celebration
Many of the common USB audio devices use a fixed 48000 Hz clock (built into the hardware) for timing when to measure each consecutive sample value. In order to operate at 44100 Hz, the data needs to be resampled from 48000 samples per second to 44100 samples per second. How exactly this conversion happens is for the hardware and its drivers to sort out. In your setup, that’s not working correctly. It seems that the device is providing too few audio samples to fill the audio buffers, so you are getting short periods of silence at the end of each audio buffer.
By telling Audacity to use 48000 Hz sample rate, we bypass the problem by calling for audio at the “native” sample rate for the device of 48000 samples per second.