Fascinating. Thanks for explaining this. I wasn't 100% sure what non-linear meant in this context and didn't want to derail, but hey, might as well. So the output after having been remapped using decode_lp_sample or similar is 16-bit. How did e.g. Mr Taylor in that function arrive at those values for this look-up table though?
May 2 2017 3:54 PM, "Dave Rice" <[log in to unmask]> wrote:
> Hi Eli,
> As noted, the use of 12 bits is not linear but it's a selection of some of the values of the 16 bit
> range that are mapped to 12. AFAIK all of the DAT rippers that support 12 bit do so with a lookup
> table in order to map the 12 bit value to the corresponding 16 bit value. For instance see the
> decode_lp_sample function, https://github.com/andrew-taylor/read_dat/blob/master/read_dat.c#L1067,
> which is used in both DatXtract and read_dat.
> Dave Rice
>> On May 2, 2017, at 3:15 PM, Eli Bildirici <[log in to unmask]> wrote:
>> Hey Lou and Ted,
>> Appreciate it but I'm trying to avoid the quality loss, however minor, of a D-A-D conversion. Once
>> the 12-bit files are created, resampling to a more standard format e.g. Redbook should be trivial
>> by using a good resampler like SoX (or Izotope if you could afford it) right? In any case there is
>> also Redbook and possibly 48kHz content on these tapes and it just seems silly to do D-A-D on
>> those, lacking even the motive of moving to a more standardized format. I'm only even asking for
>> this because I figured it would be beneficial to know an alternative to the preferred way of making
>> these copies (DSS drives and ReadDAT or DATXtract). Of course if neither of these ways work out,
>> doing a D-A-D conversion will have to do, and probably won't make much of a difference given that
>> there is very little content here that isn't straight-up voice (we're a radio show after all). But
>> it's still not ideal :/ In any case, thanks for chiming in.
> May 2 2017 2:48 PM, "Ted Kendall" <[log in to unmask]> wrote:
> If memory serves, the 12-bit convertors were non-linear anyway. Best to
> convert to something civilised via analogue.
> May 2 2017 2:24 PM, "Lou Judson" <[log in to unmask]> wrote:
> Sigh, nonstandard formats are a problem!
> But as a small time archivst myself, I would take the analog outut of the deck and re-encode it at
> 16 bit for archiving purposes. There is no way you waould actually want “bit-perfect” 12 bit files,
> as that would just pass the problem down to your successors.
> Play the DATs, record the output in a normal format, and it will be far better than any kluge you
> can do with 12 bit digital.
> I used Sony 12 bit 32kHz for some extended recordings back in the day, and this was the best way I
> found for the transfer.
> I’ll be interested to see if anyone has a better suggestion!
> Lou Judson
> Intuitive Audio
> On May 2, 2017, at 10:56 AM, Eli Bildirici <[log in to unmask]> wrote:
>> Hey ARSClisters,
>> I'm an intern in the archives department at Democracy Now! and have come across a couple hundred
>> mixed-frequency DAT tapes, many of which have 12-bit/32kHz content on them. (And yes, this is the
>> same issue an archivist from DN! posted about nearly ten years ago, here
>> (http://www.cool.conservation-us.org/byform/mailing-lists/arsclist/2007/07/msg00380.html).) We'd
>> like to make bit-perfect digital copies, but don't have a SCSI DDS drive at the moment, and in any
>> case would like to know if there's a reliable alternative path to digitally copying DATs,
>> especially those recorded at a bit-depth of 12, via S/PDIF and capture cards. I attempted to copy a
>> portion of a tape recorded at Redbook rates using our main deck - a Sony PCM-R500 - with an M-Audio
>> FireWire Solo and Audacity to record, but this was full of pops and clicks that, I assume, are the
>> result of the M-Audio failing to lock onto the tape deck's clock. (The pops and clicks only show up
>> during recording and are not on the tape itself. Actually, I suspect Audacity may in part be to
>> blame, since the pops and clicks are absent when monitoring via the M-Audio control panel app on OS
>> X...but I digress.) I've also found that external sound cards based on the C-Media CM6206 chipset -
>> cheap and plentiful on eBay - reportedly support 16-bit/32kHz input, as does the S/PDIF input of
>> our aging PowerMac G5. But that doesn't help us with 12-bit content, and nor do we know whether
>> clock syncing will pose an issue and how to deal with it if it does. My suspicion is that, while
>> there might be an economical and reliable solution to recording DAT tapes with 16-bit content (at
>> least at 44.1 or 48 kHz sample rates) without a DDS drive, that we're out of luck for 12-bit
>> content - but, any advice/confirmation/refutation would be most welcome.
>> Eli Bildirici
>> (347) 837-8337
>> Eli Bildirici
>> (347) 837-8337