In dankelley/oce: Analysis of Oceanographic Data

This is mainly a message to R.E., but I will CC it to others in this discussion of decoding Nortek AD2CP files in the R package called "oce". I am looking again at the ad2cp file that I sent before (attached again, for completeness).

I am confused as to why you (R.E.) say

I could not find any sequences of a5 0c 23 in the attached .ad2cp file.

because I see that sequence when examining the file in two different ways. I'll show the two methods in the next sections.

Examining the file in R

The R code

    buf <- readBin("ad2cp_01.ad2cp", "raw", n=1e6)

reads the file (it's 50,000 bytes, but I ask for a million to be sure to get to the end). Then I look at two spots in the file: (a) the start and (b) at byte 6997.

Start of file

The R code

    buf[1:10]

gives output

    a5 0a a0 10 f2 11 f6 8f b9 72

which we may read directly (i.e. the R function readBin() reads the bytes as they are, without alternating for the endianness of the machine)

• 0xa5: sync char
• 0x0a = decimal 10: number of bytes in the header
• 0xa0 = decimal 160: ID code, meaning that this is a text sequence

and this makes sense, as this first data chunk holds the instrument-setup commands.

Partway through file

The R code

    buf[6996 + 1:12]

gives output

    a5 0c 23 10 a0 3e 00 00 92 24 86 35

which I read as

• 0xa5: sync char
• 0x0c=decimal 12: number of bytes in the header
• 0x23: ID code that is undocumented in the 2022 manual, but which I am told is a code that will be used in future systems. (This data file may be from a test instrument.)

Using "od" command to see file contents

I am on an intel-based mac, and so "od" emits bytes in flipped order. Again, we look at two spots

Start of file

Using the unix command

    od -j 0 -N 10 -A d -vx ad2cp_01.ad2cp

yields

    0000000      0aa5    10a0    11f2    8ff6    72b9

Here, the first value is a decimal counter, and after that are the bytes (in reversed order within each 4-character unit). I read this as

• 0xa5 = sync char
• 0x0a = decimal 10 = number of bytes in header
• 0xa0 = ID for a string record (which holds the instrument-setup header)

In other words, I get the same interpretation with "od" as I get R.

Partway through file

Using

    od -j 6996  -N 10 -A d -vx ad2cp_01.ad2cp

reveals

    0006996      0ca5    1023    3ea0    0000    2492

which, as above, yields

• 0xa5: sync char
• 0x0c=decimal 12: number of bytes in header
• 0x23: unknown ID, as above

Commentary

Using two different methods, I get the byte sequence a5-0c-23 in this file. I think this file may be from a test instrument, and that would make me think I can ignore the format until it is documented, but I am confused as to why you don't see the sequence. I do not have a Windows machine so I don't know how to suggest similar work there.

dankelley/oce documentation built on May 8, 2024, 10:46 p.m.