Tweet
I have a set of music that I thought might be relevant to the recent discussions about recording media. I recorded the Grateful Dead in concert from my seat in the audience on 12/18/1973. I was about 75 feet from the right speakers. I used an early Sony lectret condenser mic and a nice mono cassette recorder. Many years later (mid-90s) the Grateful Dead released the soundboard recording of that show to the tape trading community. I obtained a cassette copy of that show which was apparently 4th
generation from the soundboard tape (which was on open reel).
I recently acquired a CD which was a copy of one based on wav files created from the first generation cassette.
I took the CD and made a copy of one song on MD using optical in. Then I played the CD through my stereo and taped the same song using a powered mic and the line in jack, and an unpowered mic and the mic in jack. Then I made mp3 copies of the CD's wave file at 64, 128 and 256 kbps encoding rates.
I took all of these files (which all were about 1 minute and 11 seconds of the exact same portion of the concert)and did a spectrum analysis. Here's what I got:
The analysis was kind of interesting because the soundboard CD shows first of all,
very little music above 10000 Hz even though there is piano, lead guitar, etc. It sounds very good and very natural.
In comparing the files, I used the CD from the soundboard as the standard. The MD copy came closest to the standard, and only departed a bit by apparently emphasising the 5k-10k range (e.g. at 5k, cd =-65dB, md=-62dB). The MD mic and line in were nearly identical, and were next closest in
matching the standard.
The MP3 encoding at 256 was next and was a good match at the low frequencies, but had
low response at 1k, but then matched the standard at 5k, and was very
close up through 10k.
The 4th gen cassette, original audience tape, MP3 encoding at 128 and were all
fairly close in frequency response. All had slightly different characteristics, but none was clearly better than the other.
The MP3 64kbps encoding was clearly a mess (audibly as well as in the specs),
and had a loss of information at most levels, with almost none above 5k.
Anyway, I thought you might find this interesting. It was helpful to me.
Dan
generation from the soundboard tape (which was on open reel).
I recently acquired a CD which was a copy of one based on wav files created from the first generation cassette.
I took the CD and made a copy of one song on MD using optical in. Then I played the CD through my stereo and taped the same song using a powered mic and the line in jack, and an unpowered mic and the mic in jack. Then I made mp3 copies of the CD's wave file at 64, 128 and 256 kbps encoding rates.
I took all of these files (which all were about 1 minute and 11 seconds of the exact same portion of the concert)and did a spectrum analysis. Here's what I got:
The analysis was kind of interesting because the soundboard CD shows first of all,
very little music above 10000 Hz even though there is piano, lead guitar, etc. It sounds very good and very natural.
In comparing the files, I used the CD from the soundboard as the standard. The MD copy came closest to the standard, and only departed a bit by apparently emphasising the 5k-10k range (e.g. at 5k, cd =-65dB, md=-62dB). The MD mic and line in were nearly identical, and were next closest in
matching the standard.
The MP3 encoding at 256 was next and was a good match at the low frequencies, but had
low response at 1k, but then matched the standard at 5k, and was very
close up through 10k.
The 4th gen cassette, original audience tape, MP3 encoding at 128 and were all
fairly close in frequency response. All had slightly different characteristics, but none was clearly better than the other.
The MP3 64kbps encoding was clearly a mess (audibly as well as in the specs),
and had a loss of information at most levels, with almost none above 5k.
Anyway, I thought you might find this interesting. It was helpful to me.
Dan
Comment