Recommendation for audio hardware

Hello, all.

Thanks for all the replies, I probably should have been a bit more
specific about what I was after.

I pretty much expected that there wouldn't be a motherboard out
there with that kind of audio capability. To be honest, I'm not
EXACTLY sure what I'm looking for, only that I'll know it when I find
it! I'm actually doing this as a three man project, I'm the
software developer of the group and I'm sorting out the 'computery'
side of the hardware as the electronics wizard doesn't have much
specific experience with computer hardware.

We're not going to bother with anything as complex as an LCD
panel. The front panel will basically be one gain knob per
channel (that's one XLR and one jack input, as on any standard
mixer) and then one volume and pan per channel for monitoring
only. Each channel would be recorded as a single mono audio
file by whatever OS is running on the board (probably a stripped
down NetBSD). There'll also be a strip of LEDs to show the overall
input level. Nothing complex.

There'll be one USB (or Firewire) port so that the disk can be
mounted on any computer over the cable, although I've not
decided on a filesystem type as FAT32 has a 4gb filesize limit.

I can't be more specific about the power consumption yet, only
that it needs to be a low as possible as we intend to use the
unit for outdoor location recording and hope to get about 10-12
hours constant use out of it per session. It'll be powered with
portable power cells outside, inside it can just use an ordinary
wall socket (obviously).

The reason for this project is frustration at the quality of the
rest of the recorders on the market. Each recorder seems to
have some fatal deficiency, like too-high power consumption
due to the manufacturer shoving every bell and whistle
possible into the design, or not having enough inputs, or
not having a USB connection. None of them seem to really
fit the criteria. Even the closest unit we found (a Fostex
something-or-other) only had four inputs and used a FAT32
filesystem, limiting each individual recording to an hour.

I'll have a read through all of the replies and have a look
at the individual suggestions in a bit.

cheers!
MC

artifact.one@googlemail.com wrote:
> 
... snip ...
> 
> We're not going to bother with anything as complex as an LCD
> panel. The front panel will basically be one gain knob per
> channel (that's one XLR and one jack input, as on any standard
> mixer) and then one volume and pan per channel for monitoring
> only. Each channel would be recorded as a single mono audio
> file by whatever OS is running on the board (probably a stripped
> down NetBSD). There'll also be a strip of LEDs to show the overall
> input level. Nothing complex.
> 
> There'll be one USB (or Firewire) port so that the disk can be
> mounted on any computer over the cable, although I've not
> decided on a filesystem type as FAT32 has a 4gb filesize limit.
> 
> I can't be more specific about the power consumption yet, only
> that it needs to be a low as possible as we intend to use the
> unit for outdoor location recording and hope to get about 10-12
> hours constant use out of it per session. It'll be powered with
> portable power cells outside, inside it can just use an ordinary
> wall socket (obviously).
> 
> The reason for this project is frustration at the quality of the
> rest of the recorders on the market. Each recorder seems to
> have some fatal deficiency, like too-high power consumption
> due to the manufacturer shoving every bell and whistle
> possible into the design, or not having enough inputs, or
> not having a USB connection. None of them seem to really
> fit the criteria. Even the closest unit we found (a Fostex
> something-or-other) only had four inputs and used a FAT32
> filesystem, limiting each individual recording to an hour.

Why don't you dedicate one file per channel, and then you can use
FAT32 with separate partitions per channel?  Now you can use large
HDs with suitable partitioning.  You also have the option of
multiple HDs.

-- 
Chuck F (cbfalconer at maineline dot net)
   Available for consulting/temporary embedded and systems.
   <http://cbfalconer.home.att.net>

On 26 Oct 2006 05:21:17 -0700, in comp.arch.embedded
artifact.one@googlemail.com wrote:

>Hello, all.
>
>Thanks for all the replies, I probably should have been a bit more
>specific about what I was after.
>
>I pretty much expected that there wouldn't be a motherboard out
>there with that kind of audio capability. To be honest, I'm not
>EXACTLY sure what I'm looking for, only that I'll know it when I find
>it! I'm actually doing this as a three man project, I'm the
>software developer of the group and I'm sorting out the 'computery'
>side of the hardware as the electronics wizard doesn't have much
>specific experience with computer hardware.
>
>We're not going to bother with anything as complex as an LCD
>panel. The front panel will basically be one gain knob per
>channel (that's one XLR and one jack input, as on any standard
>mixer) and then one volume and pan per channel for monitoring
>only. Each channel would be recorded as a single mono audio
>file by whatever OS is running on the board (probably a stripped
>down NetBSD). There'll also be a strip of LEDs to show the overall
>input level. Nothing complex.
>
>There'll be one USB (or Firewire) port so that the disk can be
>mounted on any computer over the cable, although I've not
>decided on a filesystem type as FAT32 has a 4gb filesize limit.
>
>I can't be more specific about the power consumption yet, only
>that it needs to be a low as possible as we intend to use the
>unit for outdoor location recording and hope to get about 10-12
>hours constant use out of it per session. It'll be powered with
>portable power cells outside, inside it can just use an ordinary
>wall socket (obviously).
>
>The reason for this project is frustration at the quality of the
>rest of the recorders on the market. Each recorder seems to
>have some fatal deficiency, like too-high power consumption
>due to the manufacturer shoving every bell and whistle
>possible into the design, or not having enough inputs, or
>not having a USB connection. None of them seem to really
>fit the criteria. Even the closest unit we found (a Fostex
>something-or-other) only had four inputs and used a FAT32
>filesystem, limiting each individual recording to an hour.
>
>I'll have a read through all of the replies and have a look
>at the individual suggestions in a bit.
>
>cheers!
>MC
You could try a mac mini thing with Take Vos's boom recorder software,
it is very highly regarded in the film industry
http://www.vosgames.nl/products/BoomRecorder/
and is often use on shoots, on a DC powered trolley


martin

CBFalconer wrote:
>
> Why don't you dedicate one file per channel, and then you can use
> FAT32 with separate partitions per channel?  Now you can use large
> HDs with suitable partitioning.  You also have the option of
> multiple HDs.
>

I thought FAT32 had a hard limit of 4gb per file?

The way I was planning to do it was one directory per channel, and
when you press record, a new file is created in the directory:

  channel1/
    0001.wav
    0002.wav
  channel2/
    0001.wav
    0002.wav
    0003.wav
  channel3/
  channel4/
  channel5/
  channel6/

The file is closed when the record button is pressed again. Etc.

4gb of 16 bit 44khz audio works out at about 13 hours, unless I'm
mistaken*.

MC

* The Fostex unit's hour limit wasn't just down to the FAT32 filesystem
apparently, the software on it is notoriously buggy.

martin griffith wrote:
> On 26 Oct 2006 05:21:17 -0700, in comp.arch.embedded
>
> You could try a mac mini thing with Take Vos's boom recorder software,
> it is very highly regarded in the film industry
> http://www.vosgames.nl/products/BoomRecorder/
> and is often use on shoots, on a DC powered trolley
>

Without starting a holy war, I'd like to try and steer cleer of
proprietary
solutions unless I'm forced to.

How do you get six inputs on a mac mini? External firewire device?

MC

On 26 Oct 2006 09:23:36 -0700, in comp.arch.embedded
artifact.one@googlemail.com wrote:

>martin griffith wrote:
>> On 26 Oct 2006 05:21:17 -0700, in comp.arch.embedded
>>
>> You could try a mac mini thing with Take Vos's boom recorder software,
>> it is very highly regarded in the film industry
>> http://www.vosgames.nl/products/BoomRecorder/
>> and is often use on shoots, on a DC powered trolley
>>
>
>Without starting a holy war, I'd like to try and steer cleer of
>proprietary
>solutions unless I'm forced to.
>
>How do you get six inputs on a mac mini? External firewire device?
>
>MC
Yep, M Audio are good, and Boom recorder can handle up 32 tracks I
think, and with full metadata

You wont be able to make a descent audio interface as cheap as M Audio


martin

martin griffith wrote:
> On 26 Oct 2006 09:23:36 -0700, in comp.arch.embedded
>
> You wont be able to make a descent audio interface as cheap as M Audio
>

I do actually use a Delta 66 in my workstation. They're very nice
cards.

I would still like to try to put something together though, even if
it's only as a learning experience. M-Audio don't seem to have anything
that would quite fit our needs.

Initially I did think about just a laptop and a firewire interface, but
the
ultimate goal for this device is to do a 24 hour unattended recording
and I don't see that happening for various reasons.

MC

artifact.one@googlemail.com wrote:
> CBFalconer wrote:
>>
>> Why don't you dedicate one file per channel, and then you can use
>> FAT32 with separate partitions per channel?  Now you can use large
>> HDs with suitable partitioning.  You also have the option of
>> multiple HDs.
> 
> I thought FAT32 had a hard limit of 4gb per file?

I believe it does, but your recording will spread over multiple
files, each up to 4GB.  Thus the total recording space will be
(num-of-channels * 4GB).  The separate partitions (assuming MSDOS)
will each be dedicated to a particular partition.  To regain the
overall multi-channel stream you have to read multiple files, and
keep them in sync.

-- 
Chuck F (cbfalconer at maineline dot net)
   Available for consulting/temporary embedded and systems.
   <http://cbfalconer.home.att.net>

On Thu, 26 Oct 2006 08:54:55 -0400, CBFalconer <cbfalconer@yahoo.com>
wrote:

>artifact.one@googlemail.com wrote:

>> There'll be one USB (or Firewire) port so that the disk can be
>> mounted on any computer over the cable, although I've not
>> decided on a filesystem type as FAT32 has a 4gb filesize limit.

>Why don't you dedicate one file per channel, and then you can use
>FAT32 with separate partitions per channel?  Now you can use large
>HDs with suitable partitioning.  You also have the option of
>multiple HDs.

While the FAT32 _file_ size limit is 4 GiB, the _partition_ size limit
is much larger. 

Only with FAT16, the 2/4 GiB partition size limit would apply.

If a large disk is divided into 6-7 separate partitions each recording
a single channel, the write head would constantly seek between the
partitions, slowing down the writing process. Writing to adjacent
files in the same partition or even into the same file will reduce the
disk head seek times.

A few partitions may be useful when a single day's work is recorded
into each partition, thus, when the job is done and the data is
transferred to some other safe media, the partition can be cleared
entirely and there would not be a need to defragment that partition
before reuse. 

Even if a few (2-3) partitions are used and defragmentation would be
needed, the defragmentation can be done one partition at the time. 

Also with more than one partition/disk, there would be less files in
the partition and thus the risk for directory and disk allocation
table corruption is less.  

Paul