To: jon@kana.stanford.edu Subject: Re: Sam Allan's data format Jon, This is embarrassing, but I can't find the directory that would have had the program to translate Sam Allen's tapes. I probably deleted it a few years ago. The little I remember is that the description Rob wrote on the tape was helpful, and I had to play around with blocking factors or some such thing. I have some version in Spyglass format on my mac, which seems to be: file1 - 122 stations by 4096 time points file2 - 122 stations by 4096 time points file4 - 122 stations by 4096 time points file5 - 122 stations by 4096 time points Perhaps the third file had a problem and couldn't be read? I'm not sure why I have 122 traces rather than 512. It's not sign bit, so I must have filtered it somehow to get rid of the highest frequencies. I can convert the spyglass to text and ftp it to you (about 4Mbytes per file), but you wouldn't have the original sign bit data, nor apparently the full set of stations. John Jon, I think it was sign bit, but I filtered it in time, and maybe averaged it in space to end up with non-sign bit data. John 122 is my processing of the tape. John From: John Vidale To: jon@kana.stanford.edu Subject: Re: Sam Allan's data format Jon, You jogged my memory, and I found some relevant pieces of paper. 1.5 kT of chemical explosive was set off at NTS at near midnight, sept. 22, 1993, at 12:01. The line was laid out in In Railroad Valley, Nevada, 126 miles away (I have a xerox of a map with some lines on it). There is a 10 or 20 s delay before recording. There were 56 channels in line, and 553 at N75E, with one channel dead for a total of 610 channels. The station separation is 45 feet. There are 5 X 132 sec chunks, with 0.25 to 0.5 seconds between them, which was the time to retrigger. The header is 64 bytes, there are 16384 samples per file (at 8 msec sampling rate). I recall 3050 blocks (5 by 610). I also recall 2112 bytes. The samples are integer*2. The header has sequential shot number, trace within shot, source flag, recording port-channel, source line name, receiver line name, receiver flag #, and an unused space, for a total of 4 shorts, 2 chars, 1 short, and 1 char = 8 numbers. I also have about 20 pages of xeroxes of sections. John Jon, Maybe this is clearer, or perhaps more confusing. What my notes say precisely is: each record: (then 3 boxes, in which Rob wrote:) 1. header 2. samples -> integer*2 3. eof then an indication that the pattern repeats for each record. also I see that the header is actually 64 bytes? short srcnum short trnum short srcflagnum short portchan char scrlinename[6] char reclinename[6] short reclinename char unused[42] John