|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
67.235.228.24
In Reply to: RE: Traffic "John Barleycorn Must Die" posted by Dave_K on September 25, 2015 at 03:39:59
You are welcome!
It could be an honest transfer from a tape.
In this particular case I am going to trust the software. The fact that all eight songs have a cut-off point of 22kHz is not a coincidence in my opinion.
-------------------------------------------------------------------
Big speakers and little amps blew my mind!
Follow Ups:
> It could be an honest transfer from a tape.
That is indeed what it looks like.
> The fact that all eight songs have a cut-off point of 22kHz is not a
> coincidence in my opinion.
With respect, you are misjudging the graphs you have presented. There is
no clear cutoff at 22kHz that I can see. Instead, you have the decreasing
signal content up to that frequency, with then analog tape noise extending
out to the actual sampling cutoff an octave higher in frequency.
This album was made on an early multitrack analog recorder, using most
likely 1" tape. The actual track width was about the same as that on a
cassette and I know from my own experience in recording studios back then
that the basic sound quality was not great. Without Dolby-A, the analog
noisefloor was pitiful.
John Atkinson
Editor, Stereophile
Hi John,
Thanks for posting. I understand your points about musical content falling off at 22kHz naturally. I also see them were most of the content drops off at 24 to 26kHz, which also seem very natural.
I will post all eight tracks and their accompanying test results. Please give me a few minutes. Your analysis would be appreciated.
Jamie
-----------------------------------------------------------------------
Big speakers and little amps blew my mind!
This is the first track, "Glad."
I think it helps JA's point, but it is also the highest testing of all tracks.
> File: 01-Glad.flac
Audio format: PCM
Bit depth: 24 bit
Sample rate: 96.0 kHz
Cut-Off Frequency: 24.4 kHz
TPL Left: 0.1 dB
TPL Right: -0.6 dB
TPL Mid: -0.3 dB
TPL Side: -6.1 dB
RMS Left: -19.6 dB
RMS Right: -19.7 dB
RMS Mid: -20.1 dB
RMS Side: -30.2 dB
CREST Avg.: 10.7 dB
IS L/M: 1
IS R/S: 0
Max. M-Loudness: -11.4 dB
Max. S-Loudness: -12.9 dB
Integrated Loudness: -16.8 dB
Loudness Range: 9.6 dB
PLR Avg.: 8.4 dB
------------------------------------------------------------
Big speakers and little amps blew my mind!
Track 2
This also could support a tape transfer, or the minor amount of noise after 20.9kHz could have been generated by running the output of a dac through a tube preamp and then into a his-res ADC.
I found the ad I was looking for that describes just that process. I was about to give it a separated thread when I saw JA's post. Linked below.
>
File: 02-Freedom Rider.flac
Audio format: PCM
Bit depth: 24 bit
Sample rate: 96.0 kHz
Cut-Off Frequency: 20.9 kHz
TPL Left: 0.1 dB
TPL Right: 0.1 dB
TPL Mid: -0.1 dB
TPL Side: -6.9 dB
RMS Left: -18.9 dB
RMS Right: -18.5 dB
RMS Mid: -19.1 dB
RMS Side: -29.1 dB
CREST Avg.: 10.3 dB
IS L/M: 2
IS R/S: 1
Max. M-Loudness: -10.1 dB
Max. S-Loudness: -11.1 dB
Integrated Loudness: -15.8 dB
Loudness Range: 9.9 dB
PLR Avg.: 8.2 dB
--------------------------------------------------------------
Big speakers and little amps blew my mind!
I would like to add regarding tracks one and two, there is no evidence of a natural roll-off. Not all songs have them, but one reason I test all songs on an album is to look for patterns or trends. Here you have a pattern of tracks that dip some, and then settle into a rather level noise floor. Something that is not natural.
Track 3.
Here we see that dip again, and then a leveling off past the cut-off line. I'm waiting to see at least one track on every album the exhibits a natural roll-off.
> File: 03-Empty Pages.flac
Audio format: PCM
Bit depth: 24 bit
Sample rate: 96.0 kHz
Cut-Off Frequency: 21.0 kHz
TPL Left: 0.0 dB
TPL Right: 0.1 dB
TPL Mid: -0.0 dB
TPL Side: -7.1 dB
RMS Left: -19.3 dB
RMS Right: -18.7 dB
RMS Mid: -19.3 dB
RMS Side: -31.8 dB
CREST Avg.: 10.5 dB
IS L/M: 0
IS R/S: 2
Max. M-Loudness: -10.0 dB
Max. S-Loudness: -11.6 dB
Integrated Loudness: -16.5 dB
Loudness Range: 6.6 dB
PLR Avg.: 8.3 dB
-----------------------------------------------------------------
Big speakers and little amps blew my mind!
Track 4
This one supports my opinion. Cut-off of 20kHz. After that it is totally bereft of high frequency content. Nothing but a quick roll-off of music then to a very low level noise floor. To me there is no way this could be a transfer from a tape. Some might say this is because the original tape is not viable, but I have a possible alternate answer for you. They go to certain lengths to make their files look good, and then they forget to do one or two tracks. Don't underestimate people, not even professionals. Otherwise they wouldn't have lost the original to begin with.
File: 04-I Just Want You To Know.flac
Audio format: PCM
Bit depth: 24 bit
Sample rate: 96.0 kHz
Cut-Off Frequency: 20.0 kHz
TPL Left: -0.4 dB
TPL Right: -2.2 dB
TPL Mid: -2.1 dB
TPL Side: -8.8 dB
RMS Left: -20.0 dB
RMS Right: -20.9 dB
RMS Mid: -21.1 dB
RMS Side: -28.7 dB
CREST Avg.: 10.5 dB
IS L/M: 0
IS R/S: 0
Max. M-Loudness: -13.7 dB
Max. S-Loudness: -14.1 dB
Integrated Loudness: -17.2 dB
Loudness Range: 6.0 dB
PLR Avg.: 8.1 dB
-------------------------------------------------------
Big speakers and little amps blew my mind!
Track 5.Cut-off of 20.7hKz. Again we see a low cut-off with a steep curve. Some type of energy above Fc but negligible. It looks level and unnatural.
> File: 05-Stranger To Himself.flac
Audio format: PCM
Bit depth: 24 bit
Sample rate: 96.0 kHzCut-Off Frequency: 20.7 kHz
TPL Left: -0.1 dB
TPL Right: -0.8 dB
TPL Mid: -0.7 dB
TPL Side: -8.0 dB
RMS Left: -18.2 dB
RMS Right: -18.2 dB
RMS Mid: -19.0 dB
RMS Side: -26.1 dB
CREST Avg.: 10.3 dB
IS L/M: 0
IS R/S: 0
Max. M-Loudness: -11.8 dB
Max. S-Loudness: -12.7 dB
Integrated Loudness: -15.1 dB
Loudness Range: 4.3 dB
PLR Avg.: 7.9 dB
--------------------------------------------------------------------
Big speakers and little amps blew my mind!
Edits: 09/26/15
Track 6The software was not able to establish an Fc for this track. I think it is because there is a large amount of musical energy captured in the spectrograph.
That's seems a little odd considering that JBMD is the quietest, most sublime, most understated song on the whole album. You wouldn't think it would show huge amounts of musical content.
The other three screenshots for this file I have already posted. They are the ones with the blue cursor. I guess people can read the file names alright?
I think those three graphs, once you squelch some of the energy, also show an Fc of something like 22kHz.
I think this is a mild version of a technique I have termed, "radar jamming." They simply flood the file with high frequency noise to make it look like a very good capture. Sort of like real radar-jamming they try to hide the target (Fc.) There is no way JBMD has that much native musical energy.
> File: 06-John Barleycorn (Must Die).flac
Audio format: PCM
Bit depth: 24 bit
Sample rate: 96.0 kHzCut-Off Frequency: --- kHz
TPL Left: -0.1 dB
TPL Right: -0.1 dB
TPL Mid: -0.2 dB
TPL Side: -7.8 dB
RMS Left: -20.1 dB
RMS Right: -19.8 dB
RMS Mid: -20.3 dB
RMS Side: -30.6 dB
CREST Avg.: 9.5 dB
IS L/M: 0
IS R/S: 0
Max. M-Loudness: -8.4 dB
Max. S-Loudness: -9.5 dB
Integrated Loudness: -17.1 dB
Loudness Range: 10.8 dB
PLR Avg.: 7.0 dB
----------------------------------------------------------------
Big speakers and little amps blew my mind!
Edits: 09/26/15 09/26/15
I have to hurry. Having to bargain for time now.
Track 7.
This graph could support the tape transfer position, but not very well. There is still that steep decline with an Fc of 19.3 kHz. Funky things going on the right of Fc.
> File: 07-Every Mother's Son.flac
Audio format: PCM
Bit depth: 24 bit
Sample rate: 96.0 kHz
Cut-Off Frequency: 19.3 kHz
TPL Left: -0.1 dB
TPL Right: -0.6 dB
TPL Mid: -1.0 dB
TPL Side: -6.0 dB
RMS Left: -16.9 dB
RMS Right: -18.5 dB
RMS Mid: -18.8 dB
RMS Side: -23.7 dB
CREST Avg.: 10.0 dB
IS L/M: 0
IS R/S: 0
Max. M-Loudness: -9.3 dB
Max. S-Loudness: -10.5 dB
Integrated Loudness: -14.5 dB
Loudness Range: 6.6 dB
PLR Avg.: 8.0 dB
----------------------------------------------------------------
Big speakers and little amps blew my mind!
Track 8
The last one goes to me. There is next to zero musical content above the Fc of 22.2kHz. Maybe they got tired of faking hi-res files, and gave up on the last one. Deadlines. smh
Only on JBMD do I see anything I could consider a hi-res file, and JBMD don't look to good.
I welcome thoughts on these matters.
> File: 08-Sittin' Here Thinkin' Of My Love.flac
Audio format: PCM
Bit depth: 24 bit
Sample rate: 96.0 kHz
Cut-Off Frequency: 22.2 kHz
TPL Left: -1.2 dB
TPL Right: -2.1 dB
TPL Mid: -2.4 dB
TPL Side: -7.1 dB
RMS Left: -19.7 dB
RMS Right: -17.6 dB
RMS Mid: -20.5 dB
RMS Side: -22.9 dB
CREST Avg.: 9.6 dB
IS L/M: 0
IS R/S: 0
Max. M-Loudness: -11.4 dB
Max. S-Loudness: -13.2 dB
Integrated Loudness: -15.9 dB
Loudness Range: 4.8 dB
PLR Avg.: 8.0 dB
-----------------------------------------------------------
Big speakers and little amps blew my mind!
Darn it, same track I had trouble with the first time. Reran it.
Big speakers and little amps blew my mind!
Now I'm jacking it all up.
--------------------------------------------------
Big speakers and little amps blew my mind!
Last one, then gotta go. I hope everyone posts away.
Here are some shots from Yes "Fragile." Some don't exhibit a lot of roll-off, because that happens further up and to the right of the chart.
Fragile I am confident is a transfer from tape to hi-res. Look at the organic randomness of the spectrographs. Fc is 43.4kHz.
After this review I still believe the source of JBMD was a 16/44 digital master. I certainly could be wrong.
> File: 09-Heart Of The Sunrise.flac
Audio format: PCM
Bit depth: 24 bit
Sample rate: 96.0 kHz
Cut-Off Frequency: 43.4 kHz
TPL Left: 0.1 dB
TPL Right: 0.4 dB
TPL Mid: -0.1 dB
TPL Side: -2.0 dB
RMS Left: -15.0 dB
RMS Right: -14.9 dB
RMS Mid: -16.0 dB
RMS Side: -21.7 dB
CREST Avg.: 9.9 dB
IS L/M: 2
IS R/S: 55
Max. M-Loudness: -6.9 dB
Max. S-Loudness: -7.5 dB
Integrated Loudness: -11.8 dB
Loudness Range: 10.5 dB
PLR Avg.: 7.8 dB
------------------------------------------------------------
Big speakers and little amps blew my mind!
> After this review I still believe the source of JBMD was a 16/44 digital
> master. I certainly could be wrong.
Again, with respect, you are misreading your own graphs. There is no sharp
cut-off at 22kHz with the "John Barleycorn Must Die" spectra. The sharp
cutoff due to the antialiasing converter occurs an octave higher. And while
the noisefloor of the Traffic transfers is higher than a 16-bit floor, it is
due to the presence of analog tape noise.
Your graphs suggest that the Traffic transfers appear to be high-
sample-rate transfers from an analog master, just as with the Yes Fragile
graphs you posted. Except that in that case, the original master had a
lot more HF content.
John Atkinson
Editor, Stereophile
Here are four files that look suspicious to me. If you see any that you think are upsampled from a lower res copy, then please say so. I want to learn how to read these test results better.
Thanks,
Jamie
---------------------------------------------------------------
Big speakers and little amps blew my mind!
> Here are four files that look suspicious to me.
Thank you.> If you see any that you think are upsampled from a lower res copy, then
> please say so.
None appeared to be upsampled.> I want to learn how to read these test results better.
What you are looking for is not the sort of gentle reduction in signal
level with increasing frequency that you are describing in your graphs
but the sharp cutoff at half the sample rate due to the antialiasing
filter. See the graph at the link below, which has two such cutoffs, one
at 16kHz and one at 22kHz, despite being a 96kHz file.
John Atkinson
Editor, Stereophile
Edits: 09/29/15
There is one more I would like you to take a look at, then I will leave you alone. I didn't include it, because I thought it an obvious upsample.
It has an almost vertical cutoff, but looking at it again it has a Fc of 26.1kHz, so it's past the halfway point for the antialiasing filter.
Your thoughts?
Thanks in advance.
Jamie
------------------------------------------------------------------
Big speakers and little amps blew my mind!
Well, I guess that's what you call a brickwall filter. It is very much a filter. Nothing like I've tested so far. It shouldn't be too hard to spot in the future.
I will be more circumspect with my analysis from now on.
Wort was the lifeblood of JB wasn't it? :)
Thanks John.
------------------------------------------------------------------
Big speakers and little amps blew my mind!
> I guess that's what you call a brickwall filter. It is very much a filter.
> Nothing like I've tested so far. It shouldn't be too hard to spot in the
> future.
Yes that sharp cutoff is is exactly what you should be looking for.
> I will be more circumspect with my analysis from now on.
Even so, your posting the graphs is very useful.
> Wort was the lifeblood of JB wasn't it? :)
I am not the world's best typist - the W and S keys are too close to
each other. for "wort" read "sort" :-)
John Atkinson
Editor, Stereophile
None of them look suspicious to me.
Here are some examples of what upsampled 44.1k looks like in SACD format:
http://archimago.blogspot.com/2013/07/list-suspected-44-or-48khz-pcm.html
Upsampled 44.1k in hi-res PCM format would look the same except without the ultrasonic hump in the noise floor above cutoff.
Dave, thanks for the input and the link. I will give it a another good look. Most of the brick-walls are very pronounced and some not. Chime in on a graph I post later if you want.
----------------------------------------------------------------------
Big speakers and little amps blew my mind!
Thank you for your input John. I will defer to you on the recording technology of 1970. If you say 1" tape, then I will go with that. You know much better than me what the Island Records studios might have been like at that time. Cut me a little slack that I didn't know some tape formats dropped off at 22kHz like 16/44.
I do wish you had chosen a couple of the tracks, and broken down what you saw.
I think I will have to buy and test the Bruce Springsteen album "Nebraska." I understand he recorded that album alone in a hotel room with a RadioShack four track cassette deck. To me it's a good sounding album.
I will post to you a couple of graphs that I believe are upsampled lower res files. I hope you will take the time to comment on them.
Thanks,
Jamie
---------------------------------------------------------------
Big speakers and little amps blew my mind!
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: