Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: [TROLLBAIT] Dare I start another vinyl topic? (Read 41397 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Re: Dare I start another vinyl topic?

Reply #75 –
Hmm, I've quickly checked the sample files in Baudline, and the noisefloor is actually pretty high across the board. I've tested against a youtube video I've downloaded with youtube-dl, and that noisefloor is actually lower on my end, so I don't know what happened there.

Some files have a ~17kHz peak for the entire length of the sample.

Eberhard Weber has that, Frankie's file has that, too, but it's slightly frequency shifting, which I think is down to it being played from a Vinyl, perhaps? The amplitude in this case is much lower than in the first file of that peak. In fact all the frequencies wobble back and forth regularly.

Saint-Saens has by far the lowest noisefloor. Kevin Ayers' has a much lower noisefloor than any of the other files, too, but nothing compared to the Saint-Saens file. The Saint-Saens file also has the 17kHz peak, although much lower than the previously mentioned files. Some classical instruments have pretty high overtones, and the lowpass filtering does them no good: The vibrato "hits" the filter edge and is simply clipped which creates a weird "on-off-keying" kinda effect in the upper range, not really audible, but it's just one of a more obscure artifact, and one reason you don't want a steep filter cut-off.

I'll add the spectrograms for you to check out.


Re: Dare I start another vinyl topic?

Reply #77 –
Another for good measure
Loudspeaker manufacturer

Re: Dare I start another vinyl topic?

Reply #79 –
Quote from: cliveb on
Quote from: old tech on
Not very confident of my results.  I listened in front of my PC rather than streaming the files to the hi fi but here it goes...
1. Vinyl                  5. Digital
2. Digital              6. Digital
3. Digital              7. Vinyl
4.Vinyl                  8. Digital
Thanks for taking the trouble, but since I didn't number the files, I'm not sure which files your numbers refer to.
Assuming it is the order they are found in the ZIP file, then we have:
1. Eberhard Weber
2. Frankie
3. Gordon Giltrap
4. Kevin Ayers
5. King Crimson
6. Queen
7. Saint-Saens
8. Tchaikovsky
Does my numbering match yours?
Yes. I numbered them in the order of the unzipped files.

Re: Dare I start another vinyl topic?

Reply #80 –
Quote from: polemon on
Some files have a ~17kHz peak for the entire length of the sample.
That's strange, since I brickwalled all of the files at 16kHz.

Quote from: polemon on
I'll add the spectrograms for you to check out.
Thanks for adding these, because it explains my confusion.
It looks like your Baudline spectrograms think that these files have a sample rate of 48kHz. They are actually 44.1kHz.
16kHz at a 44.1kHz sample rate will appear to be 17.4kHz if assumed to be 48kHz sample rate. I checked a few files and there are some with a peak at 15.6kHz (=17kHz if taken as sampled at 48kHz).
I guess that the peak, which is slightly below the filter cutoff, could be an artefact of the CoolEdit FFT filter I used?

Re: Dare I start another vinyl topic?

Reply #81 –
Quote from: dc2bluelight on
So...you have mostly recordings made pre-1978 then?  The industry rapidly swung to digits after that.  Even vinyl released in the late 1970s through the end of the first vinyl era was recorded digitally first.
I would agree as far as classical is concerned.
Jazz: I couldn't say.

But as far as rock and pop is concerned, I think you are way out. Apart from a few outliers (eg. Ry Cooder's 1979 Bop til you Drop), rock and pop were routinely tracked to analogue tape well into the 90s. The earliest rock album in my collection that I know for sure was digitally recorded is Joe Jackson's Big World, released in 1986. (Maybe Dire Straits' 1985 Brothers in Arms was also digital?). But the overwhelming majority of rock albums from those years were still analogue recordings.

Re: Dare I start another vinyl topic?

Reply #82 –
Quote from: polemon on
Nope. Oversampling is not done through filtering or interpolation as such, oversampling is a re-sampling of a reconstructed signal.
....
[lots of technical stuff incuding some impressive looking mathematical equations]
Doesn't alter the fact that your resampled recording includes frequencies that are not in the original files.
Something must have gone wrong.

Re: Dare I start another vinyl topic?

Reply #83 –
Quote from: cliveb on
Quote from: polemon on
Some files have a ~17kHz peak for the entire length of the sample.
That's strange, since I brickwalled all of the files at 16kHz.

Quote from: polemon on
I'll add the spectrograms for you to check out.
Thanks for adding these, because it explains my confusion.
It looks like your Baudline spectrograms think that these files have a sample rate of 48kHz. They are actually 44.1kHz.
16kHz at a 44.1kHz sample rate will appear to be 17.4kHz if assumed to be 48kHz sample rate. I checked a few files and there are some with a peak at 15.6kHz (=17kHz if taken as sampled at 48kHz).
I guess that the peak, which is slightly below the filter cutoff, could be an artefact of the CoolEdit FFT filter I used?
Well, technically, there's no such thing as a "brickwall filter" it's just a sharp cut-off low-pass filter. A Gibbs-overshoot would appear right at the cutoff, this peak that we're seeing there is unlikely it comes from the filtering process. Can you give us the un-bandlimited samples?

i'm using a simple loopback from Pulseaudio and it's piped into baudline. It just takes whatever the DAC settings are forwarded to. Hence, the spectrogram is also single channel, etc. I use it like that so it doesn't care where a signal is coming from, etc. Otherwise I'd have to record a file every time, etc.

Baudline has been around for ages: https://en.wikipedia.org/wiki/Baudline It's often used by HAM radio operators, etc.
You can load your FLAC files directly into it, if you want. It unfortunately expects /dev/audio, and PulseAudio's wrapper for that is crap. So piping it is.

I'll see if inspectrum compiles, though, apart from GNU radio this seems to be the best effort right now.

Re: Dare I start another vinyl topic?

Reply #84 –
Quote from: cliveb on
Doesn't alter the fact that your resampled recording includes frequencies that are not in the original files.
Something must have gone wrong.
Only saw your post now. Yeah, the Pulseaudio wrapper doesn't adjust the sampling frequency the DAC is set to. Your best bet is use Matlab or similar, though or just load the file.

Anyway, can you give us the un-altered sample of Eberhard Weber - Chorus Part IV.flac? I.e. not band limited, etc.?

Re: Dare I start another vinyl topic?

Reply #85 –
Quote from: polemon on
Anyway, can you give us the un-altered sample of Eberhard Weber - Chorus Part IV.flac? I.e. not band limited, etc.?
I could, but then a simple frequency analysis would give away its source (since I have already stated that the vinyl rips had energy at slightly higher frequencies than the CD rips).

However, I can tell you that the 15.6kHz peak is indeed present in the non-filtered sample - so NOT an artefact of the CoolEdit filter. Why its presence or otherwise is significant for the purposes of this test escapes me.

Re: Dare I start another vinyl topic?

Reply #86 –
Quote from: ajinfla on
Remind me again, why are you quoting me above in your list of haters?
You mean this one?:
Quote from: ajinfla on
I mean music, like certain orchestral works, with the widest dynamic range, deep bass (below 20hz), etc., played at realistic levels, to expose the limitations of vinyl...which electronic/rock/pop music that vinyl clutchers mainly listen to, won't expose.
Remember, not talking "preference" here, but the physical reality limits of the soundfield, that only the audiophile deaf could miss.
Perhaps I'm trying your tactic and did it for fun, just to give you something else to dispute.

Or perhaps I have misinterpreted your phrase "physical reality limits of the soundfield, that only the audiophile deaf could miss" as some sort of claim that vinyl is so badly flawed that it should be easy to spot. That's how it seemed to me. But we have already established that what you write often baffles me, so I could easily have got the wrong end of the stick. Perhaps you meant that it's only well recorded classical with wide dynamic range and deep bass that the audiophile deaf will miss?

Quote from: ajinfla on
Btw Clive, when you listen to your cleansed digitized vinyl, do you feel an irresistible urge to surf the net and play games, etc, shortly after hitting play?
Just curious..
There he is again, that mischievous little imp. Is this the thread equivalent of photo-bombing?

Re: Dare I start another vinyl topic?

Reply #87 –
Transparency only is important if you know how the source sounds. So it's only important for acoustic instruments recordings.
If you listen to a pop-rock album, you can not know how the source is, how the master is. It will sound however the people in the mixing table want it to sound.
So if you listen to that album through a vinyl record, the sound you're hearing can not be compared with anything.
If you listen to an orchestral, or string quartet recording, you can compare it with the real sound of an orchestra or string quartet, that you know because you've heard it live multiple times.
So the flaws of vinyl as a format, are irrelevant for pop-rock. It doesn't matter how bad the mastering is, or how different it ends up in the vinyl, because the sound of the vinyl will not be compared with anything, it will be an autonomous entity, a "particular sound". The digital edition of that album will not be better, just different, because there's not an original sound to be faithful to.
So the discussion about sound fidelity, about transparency, is just a waste of time with people that doesn't care about transparency in the first place, i.e. pop-rock aficionados.

Re: Dare I start another vinyl topic?

Reply #88 –
Quote from: cliveb on
However, I can tell you that the 15.6kHz peak is indeed present in the non-filtered sample - so NOT an artifact of the CoolEdit filter. Why its presence or otherwise is significant for the purposes of this test escapes me.
It's not, but I'm curious what it is. I remember someone saying something that it's noise picked up by the recorders while a CRT screen was running, I'm kinda wondering if that's it.

https://www.reddit.com/r/audioengineering/comments/2mszh4/whats_up_with_this_spike_at_around_157khz_on_the/

It seems the peak in your files is rather low, compared to this: https://i.imgur.com/6MVtUct.png

Re: Dare I start another vinyl topic?

Reply #89 –
Quote from: cliveb on
Quote from: dc2bluelight on
So...you have mostly recordings made pre-1978 then?  The industry rapidly swung to digits after that.  Even vinyl released in the late 1970s through the end of the first vinyl era was recorded digitally first.
I would agree as far as classical is concerned.
Jazz: I couldn't say.

But as far as rock and pop is concerned, I think you are way out. Apart from a few outliers (eg. Ry Cooder's 1979 Bop til you Drop), rock and pop were routinely tracked to analogue tape well into the 90s. The earliest rock album in my collection that I know for sure was digitally recorded is Joe Jackson's Big World, released in 1986. (Maybe Dire Straits' 1985 Brothers in Arms was also digital?). But the overwhelming majority of rock albums from those years were still analogue recordings.
But digital delay lines were finding their way onto lathes by around 1977, initially retrofitted and later incorporated in newer builds.  So how can you be sure that your album was not converted to digital and back (and at 14bits at that) unless you are sure that the record is a pre 1977 pressing?

Btw the way, there were many digital rock and pop recordings prior to 1986 (and yes Brothers in Arms was a 16/44 recording), a lot of the 1980s acts recorded digitally by the time CDs were released.  Of course there was still a lot of analog recordings around that time too.

Re: Dare I start another vinyl topic?

Reply #90 –
Quote from: cliveb on

I would agree as far as classical is concerned.
Jazz: I couldn't say.

But as far as rock and pop is concerned, I think you are way out. Apart from a few outliers (eg. Ry Cooder's 1979 Bop til you Drop), rock and pop were routinely tracked to analogue tape well into the 90s. The earliest rock album in my collection that I know for sure was digitally recorded is Joe Jackson's Big World, released in 1986. (Maybe Dire Straits' 1985 Brothers in Arms was also digital?). But the overwhelming majority of rock albums from those years were still analogue recordings.
Do you have proof of these statements?

Re: Dare I start another vinyl topic?

Reply #91 –
Quote from: cliveb on
Perhaps you meant that it's only well recorded classical with wide dynamic range and deep bass that the audiophile deaf will miss?
Yes, as I was concurring with what this guy said:
Quote from: cliveb on
I'm not suggesting that an LP and CD of the same recording would be indistinguishable - of course they would sound different (and especially so for classical).
He said only audiophile deaf type would miss the difference between digital/vinyl with classical. I agree with him. Do you?

Quote from: cliveb on
Quote from: ajinfla on
Btw Clive, when you listen to your cleansed digitized vinyl, do you feel an irresistible urge to surf the net and play games, etc, shortly after hitting play?
Just curious..
There he is again, that mischievous little imp. Is this the thread equivalent of photo-bombing?
No, serious question. Do you? Remember, you agreed with Funkstar about the whole ADD thing listening to digital vs total focus with live vinyl...which you don't do. Interesting "hypothesis".
Paradox?

Btw, did you have a chance to listen to my posted files?
Loudspeaker manufacturer

Re: Dare I start another vinyl topic?

Reply #92 –
Quote from: old tech on
But digital delay lines were finding their way onto lathes by around 1977, initially retrofitted and later incorporated in newer builds.  So how can you be sure that your album was not converted to digital and back (and at 14bits at that) unless you are sure that the record is a pre 1977 pressing?
Digital delay was only required when cutting lacquer from a digital master.  Otherwise the analog chain already had the preview head in place.  I hired a major mastering studio for two projects in the mid 1980s, both on digital masters.  The studio owned the Sony PCM-1630 (Video > PCM > DAC unit), but rented the mating digital preview delay on a per-project basis.  It was fully 16 bit, also a Sony unit, matching the 1630.    Analog projects did not use it then, as they already had the delay head.

Aside from that, the digital delay was transparent relative to analog tape's noise and distortion, it didn't impose any inherent effects.  Citing preview delay in this discussion is tangential in any case.
Quote from: old tech on
Btw the way, there were many digital rock and pop recordings prior to 1986 (and yes Brothers in Arms was a 16/44 recording), a lot of the 1980s acts recorded digitally by the time CDs were released.  Of course there was still a lot of analog recordings around that time too.
Yeah, kinda my point.  DDD was a big deal, a marketing plus. 

Re: Dare I start another vinyl topic?

Reply #93 –
Quote from: ajinfla on
No, serious question. Do you? Remember, you agreed with Funkstar about the whole ADD thing listening to digital vs total focus with live vinyl...which you don't do. Interesting "hypothesis".
Paradox?
Yes, sometimes I might have music on in the background while I do something else (such as browse the net, do some ironing, etc).
But the crucial point here is that the music is the secondary thing. It's only on because it's so easy to play it when you have network streamers around the house. On those occasions when I put on some music for the purpose of listening, I pay attention.

Back in the days when I had to put on a physical disc (vinyl or CD), it only ever happened when I wanted to listen. I never had it on as a background while doing something else. (Actually, there is one caveat - as a teenager, if I had a visitor of the fairer sex, then music got put on, but my attention frankly wasn't focussed on it).

Quote from: ajinfla on
Btw, did you have a chance to listen to my posted files?
Not yet. I've spent too long fielding responses on this thread.

Re: Dare I start another vinyl topic?

Reply #95 –
Quote from: cliveb on
However, I can tell you that the 15.6kHz peak is indeed present in the non-filtered sample - so NOT an artefact of the CoolEdit filter. Why its presence or otherwise is significant for the purposes of this test escapes me.
Me as well.

Re: Dare I start another vinyl topic?

Reply #96 –
Quote from: ajinfla on
Quote from: cliveb on
Not yet. I've spent too long fielding responses on this thread.
They are examples of why I "hate" vinyl I suppose.
OK, I've now taken a cursory listen to your files. Not sure what to say, really. They sound good to me, typical examples of well recorded classical music. But my (limited) experience of classical is that it is nearly always well recorded within the constraints of the technology available at the time.

The only piece I also have is the Firebird Suite (Robert Shaw, Atlanta Symph Orch, Telarc, on CD). The extract you posted sounds marginally clearer than the Telarc, but to be frank the difference in performance is vastly more significant.

Although I'm quite partial to bombastic Russian stuff, the 1812 has never been my cup of tea, so I don't have it.

Re: Dare I start another vinyl topic?

Reply #97 –
Quote from: cliveb on
They sound good to me, typical examples of well recorded classical music. But my (limited) experience of classical is that it is nearly always well recorded within the constraints of the technology available at the time.
Disagree, there are plenty horrible classical recordings of every era. However that isn't the point. The point is that if you are going to seriously compare differences of vinyl vs digital, don't handcuff the digital. Show what it can do as my samples did, yours didn't.
The "especially classical" per you.

Quote from: cliveb on
Although I'm quite partial to bombastic Russian stuff, the 1812 has never been my cup of tea, so I don't have it.
Once again, totally misses the point. Try those canon shots on a TT, tell us what happens. There's your answer.

p.s. I hope Ralph didn't blow up his audiophile "hifi" system playing that track. I have personally witnessed a >$300k "audiophile" system explode/shut down playing that very track, which my non-believer "lofi" systems handle with ease.
That's what makes the whole audiophile/vinylphile/studiophile, etc etc. thing so darn funny. :))

Ok, so when do we find out which of your tracks were which?
Loudspeaker manufacturer

Re: Dare I start another vinyl topic?

Reply #98 –
Quote from: dc2bluelight on
Digital delay was only required when cutting lacquer from a digital master.  Otherwise the analog chain already had the preview head in place.  I hired a major mastering studio for two projects in the mid 1980s, both on digital masters.  The studio owned the Sony PCM-1630 (Video > PCM > DAC unit), but rented the mating digital preview delay on a per-project basis.  It was fully 16 bit, also a Sony unit, matching the 1630.  Analog projects did not use it then, as they already had the delay head.
I don't believe DDL was only relevant to cutting from digital recordings, though I am not a subject matter expert.  One of the main benefits DDL is that the preview delay was more precise and noise free, regardless whether the source material was a digital or analog recording.  Also, it doesn't surprise me that DDL was 16 bits by the time we were in the early 80s, the earlier implementations were not.
Quote from: dc2bluelight on
Aside from that, the digital delay was transparent relative to analog tape's noise and distortion, it didn't impose any inherent effects.  Citing preview delay in this discussion is tangential in any case.
Tangential to an extent but it illustrates that digital processing does not need to change the sound of an analog recording, even at 14 bits.  The DDL story is enough to confirm that the vinylphile's obsession with all analog is irrational and pure placebo.  What is tangential though is comparing a vinyl record to analog tape, for example, a copy of an analog tape to another also does not need to change the sound (apart from a slight generational loss) but the sound will always change when cut to a vinyl record.

Re: Dare I start another vinyl topic?

Reply #99 –
Quote from: old tech on
I don't believe DDL was only relevant to cutting from digital recordings, though I am not a subject matter expert. 
It wasn't.  But the onset of universal use of DDL for all lacquer cutting is just a bit overblown.  It didn't suddenly appear then become universally applied to all mastering.  It was expensive, memory was expensive, and thus for about a half decade or so it was adopted slowly, often rented per project, and about the time it was cheap enough and good enough to be universally adopted for all mastering, vinyl was already fading out. 
Quote from: old tech on
One of the main benefits DDL is that the preview delay was more precise and noise free, regardless whether the source material was a digital or analog recording.
Precision of delay time was already adequate using the analog preview head.  DDL in mastering was used for automatic lathe pitch control, pre-adjusting the land between grooves.  Precision was relatively easy, and not an issue even for analog preview machines.  Just a matter of linear tape speed and head spacing.  The issue of DDL audio performance was very significant because it was the delayed signal that actually got cut.
Quote from: old tech on
Also, it doesn't surprise me that DDL was 16 bits by the time we were in the early 80s, the earlier implementations were not.
You'll have to cite a reference for that one.  14 bit delays in mastering would have always required truncating two LSBs, as every pro digital recording system from 1976 (Soundstream) on was 16 bits.  Only the original EIAJ "prosumer" stuff was 14 bits, and Sony took care of that too.  Any EIAJ material that made it to release would have been edited on a PCM-1600-based 16 bit system. The Sony delay product was introduced along with their original PCM adapter unit, the PCM-1600, which was part of nearly every CD production chain, and part of the lacquer mastering process from its inception. It was 16 bits.  All DASH recorders were 16 bits, and most production was at 44.1 because resampling was a problem in the early days. 

14 bit DDLs also have a failure in design logic.  DDLs were based on memory chips of the time, and they were all structured in 4 bit or 8 bit bytes, so 16 bit delay would have been a natural, 14 bit would have been unnecessary and of no advantage. 

In fact, the only reason 14 bit digital audio ever existed at all was cheaper ADCs and DACS, which in turn drove the early EIAJ 14 bit video-based formats.  Sony found room for the extra two bits by dropping a bit of error correction code, IIR.  But no matter, it all became 16 bit format for editing and mastering.  Thus the DDL was highly likely to be 16 bits as well.  Let me know if you find a confirmed variance.  One possibility would be a DDL designed specifically to replace the preview head arrangement which was an expensive mod to an analog deck.  Might have been a cheap alternative. But I doubt it proliferated.
Quote from: old tech on
Tangential to an extent but it illustrates that digital processing does not need to change the sound of an analog recording, even at 14 bits.  The DDL story is enough to confirm that the vinylphile's obsession with all analog is irrational and pure placebo.  What is tangential though is comparing a vinyl record to analog tape, for example, a copy of an analog tape to another also does not need to change the sound (apart from a slight generational loss) but the sound will always change when cut to a vinyl record.

Actually, you have it a bit backwards in some ways.  Analog tape copies have significant generational loss in distortion, noise and FR in particular.  FR response issues can be partially compensated for if recognized and analyzed in detail, but usually were not.  Setup tones were 1kHz and 10kHz, no spot frequencies or sweeps to go by.  The lacquer/vinyl system had the potential for being inherently flatter FR, and the noise spectrum was in many ways less audible than that of certain tape speeds, but it did introduce other losses.  It's maximum HF modulation was quite limited geometrically, and separation wasn't nearly as good as tape.  Vertical groove distortion was a problem exacerbated by wear, but off the press, not too bad.  All in all a direct-cut from live vinyl could (and did) sound better than a second or third generation analog tape.  Listen to any of the early 1970s Sheffield direct to disc stuff if you doubt that, and compare to their CDs made from tape backups of the same sessions. 

Edit: after some extensive googling...the original Studer mastering delay unit was indeed 14 bit, updated to 16 by the second generation.  Sorry, don't have dates.  The Ampex ADD-1 was 16 bits (ca 1979).  The reason DDL wasn't adopted quickly for analog only projects was the cost of memory.  One on-line referencehttp://jcmit.net/memoryprice.htm has the cost/Mb at almost $37,000 in 1977, dropping to $10,000 by January of 1979.  That should scale the cost of a 1.8 second DDL. These things were more expensive than the tape-based system with no advantage until you started to cut from digital masters.