Reading bytes from a file

[Melissa Krueger]

Hi all,

I've got some old images that have a four byte header giving the image  
width and height. I'm trying to import the stacks via a macro that  
would read those bytes and pass them on to a call to open a Raw stack.  
This works fine for some of my stacks, but not others.

The charCodeAt function doesn't seem to reliably report the byte  
value. If I print out the first four bytes of a non-working image, I  
get: 225,1,63,1. If I do an octal dump from the command line (od -t u1  
-N 4) on the same file, I get: 135,1,63,1. The octal dump values give  
me the right width and height calculations.

Is there a way for me to get at those first 4 bytes without invoking  
charCodeAt?

I've included my code snippet in case anyone was interested.

     // Open a file and get x & y sizes
     myFile = File.openDialog("Choose an MCID file");
     str = File.openAsRawString(myFile);
     wide = charCodeAt(str, 0) + 256 * charCodeAt(str, 1) + 1;
     high = charCodeAt(str, 2) + 256 * charCodeAt(str, 3) + 1;

     // Open all files in the directory
     run("Raw...", "open="+myFile+" image=8-bit width="+wide+"  
height="+high+
         " offset=4 number=1 gap=0 open");

Thanks,
Melissa


--
Melissa A. Krueger, Ph.D.
Div Pulmonary & Critical Care Medicine
University of Washington
Seattle, WA 98195-6522

[Michael Schmid]

Hi Melissa,

what about using an InputStream instead? String operations always  
suffer from the problem of different character sets, and you may get  
conversion between them. See
   http://java.sun.com/j2se/1.4.2/docs/api/java/io/FileInputStream.html

   myInputStream = new FileInputStream(myFile);
   wide = (myInputStream.read() & 255) + 256 * (myInputStream.read()  
& 255);

or
   myInputStream = new FileInputStream(myFile);
   byte[] myByteArray = new byte[4];
   nBytesRead = myInputStream.read(myByteArray);
   //error checking for nBytesRead < 4 bytes should go here
   wide = (myByteArray[0] & 255) + 256 * (myByteArray[1] & 255);

See the readShort() method of AVI_Reader.java for slightly nicer code  
for 2-byte -> short conversion in Intel byte order.
   http://rsb.info.nih.gov/ij/source/ij/plugin/AVI_Reader.java

Hope this helps,

Michael
________________________________________________________________

On 29 Apr 2009, at 00:50, Melissa Krueger wrote:

> Hi all,
>
> I've got some old images that have a four byte header giving the  
> image width and height. I'm trying to import the stacks via a macro  
> that would read those bytes and pass them on to a call to open a  
> Raw stack. This works fine for some of my stacks, but not others.
>
> The charCodeAt function doesn't seem to reliably report the byte  
> value. If I print out the first four bytes of a non-working image,  
> I get: 225,1,63,1. If I do an octal dump from the command line (od -
> t u1 -N 4) on the same file, I get: 135,1,63,1. The octal dump  
> values give me the right width and height calculations.
>
> Is there a way for me to get at those first 4 bytes without  
> invoking charCodeAt?
>
> I've included my code snippet in case anyone was interested.
>
>     // Open a file and get x & y sizes
>     myFile = File.openDialog("Choose an MCID file");
>     str = File.openAsRawString(myFile);
>     wide = charCodeAt(str, 0) + 256 * charCodeAt(str, 1) + 1;
>     high = charCodeAt(str, 2) + 256 * charCodeAt(str, 3) + 1;
>
>     // Open all files in the directory
>     run("Raw...", "open="+myFile+" image=8-bit width="+wide+"  
> height="+high+
>         " offset=4 number=1 gap=0 open");
>
> Thanks,
> Melissa
>
>
> --
> Melissa A. Krueger, Ph.D.
> Div Pulmonary & Critical Care Medicine
> University of Washington
> Seattle, WA 98195-6522

[Wayne Rasband]

The charCodeAt() macro function should work as expected in the v1.42o
daily build, which fixes a bug in File.openAsRawString() that caused it
to not correctly read bytes with values greater than 127.

-wayne

On Apr 28, 2009, at 6:50 PM, Melissa Krueger wrote:

> Hi all,
>
> I've got some old images that have a four byte header giving the image
> width and height. I'm trying to import the stacks via a macro that
> would read those bytes and pass them on to a call to open a Raw stack.
> This works fine for some of my stacks, but not others.
>
> The charCodeAt function doesn't seem to reliably report the byte
> value. If I print out the first four bytes of a non-working image, I
> get: 225,1,63,1. If I do an octal dump from the command line (od -t u1
> -N 4) on the same file, I get: 135,1,63,1. The octal dump values give
> me the right width and height calculations.
>
> Is there a way for me to get at those first 4 bytes without invoking
> charCodeAt?
>
> I've included my code snippet in case anyone was interested.
>
>     // Open a file and get x & y sizes
>     myFile = File.openDialog("Choose an MCID file");
>     str = File.openAsRawString(myFile);
>     wide = charCodeAt(str, 0) + 256 * charCodeAt(str, 1) + 1;
>     high = charCodeAt(str, 2) + 256 * charCodeAt(str, 3) + 1;
>
>     // Open all files in the directory
>     run("Raw...", "open="+myFile+" image=8-bit width="+wide+"
> height="+high+
>         " offset=4 number=1 gap=0 open");
>
> Thanks,
> Melissa
>
>
> --
> Melissa A. Krueger, Ph.D.
> Div Pulmonary & Critical Care Medicine
> University of Washington
> Seattle, WA 98195-6522
>

[Melissa Krueger]

That works perfectly. Thanks!

Melissa

On Apr 29, 2009, at 7:24 AM, Wayne Rasband wrote:

> The charCodeAt() macro function should work as expected in the  
> v1.42o daily build, which fixes a bug in File.openAsRawString() that  
> caused it to not correctly read bytes with values greater than 127.
>
> -wayne
>
> On Apr 28, 2009, at 6:50 PM, Melissa Krueger wrote:
>
>> Hi all,
>>
>> I've got some old images that have a four byte header giving the  
>> image width and height. I'm trying to import the stacks via a macro  
>> that would read those bytes and pass them on to a call to open a  
>> Raw stack. This works fine for some of my stacks, but not others.
>>
>> The charCodeAt function doesn't seem to reliably report the byte  
>> value. If I print out the first four bytes of a non-working image,  
>> I get: 225,1,63,1. If I do an octal dump from the command line (od -
>> t u1 -N 4) on the same file, I get: 135,1,63,1. The octal dump  
>> values give me the right width and height calculations.
>>
>> Is there a way for me to get at those first 4 bytes without  
>> invoking charCodeAt?
>>
>> I've included my code snippet in case anyone was interested.
>>
>>    // Open a file and get x & y sizes
>>    myFile = File.openDialog("Choose an MCID file");
>>    str = File.openAsRawString(myFile);
>>    wide = charCodeAt(str, 0) + 256 * charCodeAt(str, 1) + 1;
>>    high = charCodeAt(str, 2) + 256 * charCodeAt(str, 3) + 1;
>>
>>    // Open all files in the directory
>>    run("Raw...", "open="+myFile+" image=8-bit width="+wide+"  
>> height="+high+
>>        " offset=4 number=1 gap=0 open");
>>
>> Thanks,
>> Melissa
>>
>>
>> --
>> Melissa A. Krueger, Ph.D.
>> Div Pulmonary & Critical Care Medicine
>> University of Washington
>> Seattle, WA 98195-6522
>>

[Harry Parker]

Hi Melissa,

I'd try just opening the image temporarily with the File.openAsRawString(path, 4) function to read the 1st 4 bytes.
Then close the file and reopen it as a raw image as before.

 --
Harry Parker
Senior Imaging Systems Engineer



________________________________
From: Melissa Krueger <[hidden email]>
To: [hidden email]
Sent: Tuesday, April 28, 2009 6:50:50 PM
Subject: Reading bytes from a file

Hi all,

I've got some old images that have a four byte header giving the image width and height. I'm trying to import the stacks via a macro that would read those bytes and pass them on to a call to open a Raw stack. This works fine for some of my stacks, but not others.

The charCodeAt function doesn't seem to reliably report the byte value. If I print out the first four bytes of a non-working image, I get: 225,1,63,1. If I do an octal dump from the command line (od -t u1 -N 4) on the same file, I get: 135,1,63,1. The octal dump values give me the right width and height calculations.

Is there a way for me to get at those first 4 bytes without invoking charCodeAt?

I've included my code snippet in case anyone was interested.

    // Open a file and get x & y sizes
    myFile = File.openDialog("Choose an MCID file");
    str = File.openAsRawString(myFile);
    wide = charCodeAt(str, 0) + 256 * charCodeAt(str, 1) + 1;
    high = charCodeAt(str, 2) + 256 * charCodeAt(str, 3) + 1;

    // Open all files in the directory
    run("Raw...", "open="+myFile+" image=8-bit width="+wide+" height="+high+
        " offset=4 number=1 gap=0 open");

Thanks,
Melissa


--Melissa A. Krueger, Ph.D.
Div Pulmonary & Critical Care Medicine
University of Washington
Seattle, WA 98195-6522