Taperssection.com
Gear / Technical Help => Recording Gear => Topic started by: kleiner Rainer on February 18, 2018, 04:56:38 AM
-
found this some days ago:
http://www.vlsi.fi/en/support/evaluationboards/vs1005devboardextension1.html
Records up to 96 kHz 24-bit stereo PCM to SD Card
Maximum tested SD card capacity 256 GB
Uses VS23S040 as a 4-Mbit audio buffer
Graphical user interface
Uses stdaudioin so can be configured to use either VS1005's own ADCs or external ADC
Dynamic range 98 dB(A) with VS1005's own ADC
For longer recordings, can split audio to multiple 2 GB files with zero sample loss, overcoming FAT32 file size limitations
Recording time on 64 GB card from over 30 hours (96k/24b) to over 92 hours (48k/16b)
Automatic file system recovery if power is lost while recording
Full source code available
Manufacturers forum:
http://www.vsdsp-forum.com/phpbb/viewtopic.php?f=13&t=2210
Greetings,
Rainer
-
Looks good. 275 Euro for development board. (+ analog break out board needed)?
It says up to 5 channels. That is interesting...
In the meantime, how about a Raspberry Pi with a standard USB device attached.
Write to external USB.
The only problem is making sure USB card, if disconnected, keeps/recovers the recording.
-
The Pi is doable. You can record 24 bit 96k stereo WAVs from the command line to a max size 32GB micro SD, it could be a simple bit bucket behind a pre. arecord will automatically split the file at 2GB also. It could be as basic as one button to start arecord, one to stop the process, and a light while arecord is active.
Doing it that simple you would need to understand what your levels would need to be going into the Pi. I've never tried this but the tools and functionality are there in the simplest sense from the command line.
-
Pi is certainly doable, does need some care to prevent clicks/pops/dropouts.
Using it with an Echo2 USB interface, controlling the pi3 by iPhone [with ssh] over Wifi; stunning sound.
-
Wow, that is awesome.
Sorry to thread hijack. What are the click/pop/dropout issues with the Pi, connections I assume?
-
Running the pi3 'default' I lost audioframes, writing to SDcard seems a bit slow [or buffering not up to speed].
Some kernel tweaks added, no issues anymore.
-
its not cheaper or better "cost/benefit" to use a usb audio interface (like the behringer umc404, 4 in, 24-Bit/92 kHz) and use Pi or a Android phone to do the recording?
i don't know how much a umc404 cost compared to this device (im in brazil) but here they are very cheap compared to any recording device
-
It says up to 5 channels. That is interesting...
Interest piqued!
5 is just enough for my application, although I wish it would do 6, or 7, or 8 channels for other applications I also have in mind.
[lurk mode on]
-
Stupid question: Is there some proprietary firmware for this device or does it also support current Linux distributions such as Debian Stretch (or one of its derivatives)? Does this board require any specific proprietary audio or video drivers? If you've git to compile your own kernel, proprietary drivers can be a major PITA.
-
The VS1005 Eval unit doesn't have a CPU in the standard sense but uses the VS1005 DSP to serve this purpose. There are Linux platforms available for some DSP's but you would probably want to avoid this as it would add unnecessary overhead.
The firmware is specific to the VSDSP.
Here is the development software:
http://www.vlsi.fi/en/support/software/vside.html
-
Interesting and you're right. The VS1005 is a pretty nice dedicated audio processing chip http://chipdesign.ulitzer.com/node/2154458 (http://chipdesign.ulitzer.com/node/2154458) However, the ADE is Windows only with a Wine option. Not *nix friendly.
-
It says up to 5 channels. That is interesting...
Where did you find this? In a cursory look at the documentation and forum I couldn't find anything on its potential channel capacity.
Implementation of this beyond the dev-kit is over my head. Question for the group- what would be the potential minimum size of a minimal-function 5ch (6ch if possible) recorder based on this. Figure incorporating a 5 or 6 ch battery box / buffer amp ahead of it in the same enclosure for use with unbalanced low-voltage supply microphones. Candy-tin?
-
The VS1005 Eval unit doesn't have a CPU in the standard sense but uses the VS1005 DSP to serve this purpose. There are Linux platforms available for some DSP's but you would probably want to avoid this as it would add unnecessary overhead.
The firmware is specific to the VSDSP.
Here is the development software:
http://www.vlsi.fi/en/support/software/vside.html (http://www.vlsi.fi/en/support/software/vside.html)
It seems to have some 8 bit (Atmel?) processor.
Yes, very specific DSP codes too.
Now it would seem appropriate to use an ARM chip.
Either Raspberry Pi or BeagleBone platform...
-
The VS1005 Eval unit doesn't have a CPU in the standard sense but uses the VS1005 DSP to serve this purpose. There are Linux platforms available for some DSP's but you would probably want to avoid this as it would add unnecessary overhead.
The firmware is specific to the VSDSP.
Here is the development software:
http://www.vlsi.fi/en/support/software/vside.html (http://www.vlsi.fi/en/support/software/vside.html)
It seems to have some 8 bit (Atmel?) processor.
Yes, very specific DSP codes too.
Now it would seem appropriate to use an ARM chip.
Either Raspberry Pi or BeagleBone platform...
I guess for "our purposes" you'd need a device with a low-latency kernel and a decent amount of RAM for jackd (instead of ALSA).
While this is definitely possible with a Raspberry PI, you'd need to tweak a lot of kernel parameters
https://wiki.linuxaudio.org/wiki/raspberrypi
The more powerful, Intel-based UP^2 board looks interesting though. And it features the same 40 pin GPIO as the RPi.
-
I just scanned the web store. Reminds me of the old Heath Kits my Dad used to play around with for his ham radio hobby when I was a kid. Almost as expensive as a manufactured component, but you have to build it yourself. My Dad loved doing it, so if you're into DIY projects for their own sake, this could be pretty fun. Case construction is the part I'd struggle with. Hope somebody here makes one and posts pics and samples.
-
A modern digital Heath Kit seems an accurate assessment. I'm not so interested in having yet another DIY project for its own sake, but rather as a possible way of obtaining something not commercially available which targets specific recording needs.
-
I couldn't find much info for multi channel Pi recording, but for 2 channel there is an interesting Wolfson Audio card for the Pi that allows Analog line in/out, S/PDIF in/out and it seems like you do not have to compile a low latency kernel to use it. For £19+8 for the card's carbon fiber enclosure that isn't bad for a small two channel setup, or more if using a mixer.
Using a phone with SSH to control the Pi as mentioned above it would make a nice low profile rig. This may be something I want to play with in the near future.
http://uk.farnell.com/wolfson-audio-card-raspberrypi (http://uk.farnell.com/wolfson-audio-card-raspberrypi)
-
The Audio Injector Octo soundcard looks interesting.
It features 8 RCA outs and 6 RCA ins.
https://www.kickstarter.com/projects/1250664710/audio-injector-octo-surround-sound-for-the-raspber
There's also the PiSound interface that currently supports MIDI only (but if memory serves, there was a discussion about XLR ins, too).
https://blokas.io/