Null test 88.2/24 and 44.1/16

[Jud]

57 minutes ago, bachish said:

I know that the high res file, technically, has all the original info (ultrasonics) 

 

Insofar as the studio mics were able to capture it, and it wasn't removed by the ADC(s) during sigma-delta modulation or any downsampling.

[Jud]

1 hour ago, bachish said:

Yes, I'm not saying the wav file has a perfect representation of the original source recorded.

 

I am comparing listening back to either the high res, original version vs the down sampled version.  

 

Right, just wanting us to remember that A/D is the flip side of what goes on in a DAC - first converted from analog to digital with sigma-delta modulation, then (with most ADCs) converted to PCM at 352.8 or internally/externally downsampled to the more common resolutions, 192 or 176.4kHz or below.  So even the "hi res original" from the studio is almost always downsampled.

[Jud]

7 hours ago, bachish said:

My understanding is that -80db below the signal is considered inaudible and covered by the masking effect.

 

If you were trying to hear noise, that's correct. But we're trying to hear music.  How much very low level musical detail is the noise masking?  (Maybe none, considering the noise level of a typical listening room, but it's the way of conceptualizing noise and masking that I'm thinking of.)

[Jud]

I fully agree with what you both are saying.  Let me see if I can't also encourage you to think about this in a slightly different way, not necessarily for this particular example, but in general.

 

You're with a friend at a restaurant.  She wears a hearing aid.  She complains about the noise of conversation at the other tables.  You say "It's OK, it's not like we can overhear what they're saying."  She replies, "But I'm not worried about overhearing them.  They're making it more difficult to hear you."

 

Later, she mentions the A/C coming on and the noise that makes.  You say, "But the noise of conversation is louder."  She replies, "Yes, but both of them are making it more difficult for me to hear."

 

In other words: I'm encouraging thinking about noise masking music, rather than vice versa, since the music is what we're trying to listen to; and thinking about noise sources as additive, not as one excluding or drowning out another.

 

Now it may well be that in a typical situation, any additional noise from filtering, especially with good software like iZotope, amounts to a fart in a hurricane compared to the noise level in a typical listening room, so though additive, its contribution might be utterly negligible.  No problem with that at all.  I just want to encourage a mindset that looks at ability to hear musical detail, rather than masking and whether we can notice audible distortion.

[Jud]

21 hours ago, Jud said:

 

Right, just wanting us to remember that A/D is the flip side of what goes on in a DAC - first converted from analog to digital with sigma-delta modulation, then (with most ADCs) converted to PCM at 352.8 or internally/externally downsampled to the more common resolutions, 192 or 176.4kHz or below.  So even the "hi res original" from the studio is almost always downsampled.

 

Following up on this -

 

Here are some recordings available from NativeDSD, and you can filter your search to come up with only those albums/tracks available in DSD256: https://www.nativedsd.com/new_browse

 

@bachish, @pkane2001, I don't know whether you're able to do the same analysis with one of these tracks, but it would be interesting to see whether a recording that presumably had experienced less processing would show any greater difference to a 44.1kHz resolution downsampled file than the 88.2kHz original you used before.

[Jud]

10 minutes ago, crenca said:

@Jud, on the contrary very on topic.  As much as I appreciate what the OP, Paul, have done and your cogent suggested expansion, the fact is most of this analysis has been done before and is (largely) non-controversial and accepted.  In Audiophiledom however much is claimed about not only about the integrity of sampling/down-and-up conversations, etc. so much as how all this does not matter because it is over those (realtively) short USB connection (or ethernet, etc.) that all the Audiophile grade endpoints, USB cables, and internal EM/RF "noise" do their thang to the sound (positively or negatively).  Suddenly, in this last 10FT or so of digital (analogue of course at the physical layer) the "sound" of various files/connection method suddenly bloom up, including for yourself with whatever cheap USB mod your using currently.  There is a rather large disconnect between your faith in your own experience and the faith you are putting in this topics computational analysis because the fact is these files suffered all sorts of EM, RF, poor cables, and other assaults that you claim you "hear" in the last 10ft....  

  

 

I don't care.   

 

I'm just interested in the very nice analysis these two kind folks have done, and looking forward to anything further they're able to do along the same lines.  But to save these people the trouble, if you have got a link to the types of measurements they've done, comparing a DSD256 recording and a 44.1kHz downsampled version, that would be great.  If including DSD is a problem, there are the 2L comparison tracks that would enable comparative measurements of 352.8 and 44.1kHz files: http://www.2l.no/hires/index.html?

[Jud]

Just now, crenca said:

The assault from the EM/RF inside these computers on the files in question can hardly be imagined.

 

On the contrary, its analog is experienced in all the threads polluted by fruitless debates.  It would be so very nice to keep this thread free of that sort of thing.

[Jud]

39 minutes ago, pkane2001 said:

Why do you think that DSD captures would be subject to less processing than PCM that @bachish captured and processed himself?

 

Hi Paul -

 

I was thinking to capture the recording at an earlier stage of the A/D process.  The way I understand the files from NativeDSD originate and are edited  (@bmoura and @tailspn will know more), during editing either only the portions being edited or the entire file would be converted from sigma-delta modulated format to PCM for some operations, then the relevant portions or the entire file would be converted with a modulator to DSD256 format, and no additional conversions/decimation would have been done.  Or with the 2L files, the "DXD" (352.8kHz) files would have come from the ADC already converted in their entirety to PCM, edited in that format, and no further conversions/decimation done from there.  This would be in contrast to an 88.2kHz file, which I believe would have to be decimated from some higher PCM resolution.

 

If all this is correct (and I'm not certain it is), then I'd be interested to know whether there's much (if any) more difference between one of these files and a 44.1kHz version, versus the 88.2kHz/44.1kHz comparison.

[Jud]

10 hours ago, bachish said:

 The integrity of audio files as they pass through digital cables and get processed in the computer seems to be assumed in recording circles. 

 

There are a few folks in recording circles who fuss about such things, but they are definitely outliers. 

[Jud]

9 hours ago, bachish said:

 

My hypothesis is...it probably would compare quite well to the original in the audible range much like this test did.

 

I personally have little doubt of it, but I'm curious to see whether there's any difference, and if so, where. 

[Jud]

Just now, bachish said:

 

Ok, interesting. I suppose it isn't too hard to test. I could take an original recording that is on the SD card in my field recorder that hasn't been transfered yet, upload it to the computer, load it into programs and export over and over, transfer between external hard drives, many times, and upload and download from OneDrive several times. 

 

I could then take it, flip the phase and load it into my DAW and load, just one time, the original from the SD card into the DAW and see if they null. Would be interesting.

 

 

And/or you could ask Cookie.   

 

Whether I agree with what she thinks on various matters (bearing in mind she's an experienced producer, I'm just some guy), I really like the quality of the recordings she produces. 

[Jud]

14 minutes ago, bachish said:

Yes, I'm sure she does excellent work. I'm definitely going to take a listen.

 

You mean ask her to do the test or ask her opinion?

 

The various things she thinks make a difference to eventual sound quality.

[Jud]

First things first, I'm hoping in the midst of our more usual discussions that it will be possible to go a little further along with @bachish's and @pkane2001's exploration of the resolution question to see what if any difference might be obtained in a comparison between Cookie's files  and a 44.1 version, or a 2L DXD file and the 44.1 version.

 

Two further steps might be interesting: comparisons with a 44.1 file produced by @bachish's own decimation filter; and with a 44.1 file produced by someone's version of a simple "bad" decimation filter of the type a major music company might use, that I've seen criticized occasionally here.  For the latter, @mansr @Miska, @PeterSt, Cookie, any ideas? 

[Jud]

52 minutes ago, PeterSt said:

 

Hi Jud - Is this about the Q word ?

 

I don't think so.  I've seen various critical comments over the years regarding the filters used in getting releases by the major music companies from the studio to CD format.  For example, I recall Miska saying many (most?) of the filters used ring.  That type of thing.

[Jud]

2 minutes ago, mansr said:

I would assume they use whatever their DAW, e.g. Pyramix, provides. Why do you believe it would be "bad"?

 

Because I have in the past read negative remarks about some of this filtering by people like @Fokus and Miska. 

[Jud]

9 minutes ago, pkane2001 said:

 

All do, this is the nature of band-limited digital filtering. Sharper filters ring more. I know Miska measures DACs well into MHz, so yes, you can see all the ugly artifacts in the inaudible range. But do they matter?

 

In testing DeltaWave, I found some recordings that had noise pushed out to well above 88KHz range by noise shaping, at nearly the same level as the original (audible) signal. Perhaps that's an example of a poorly designed filter? And yet, it's still ultrasonic energy that most likely doesn't affect the audible range... May destroy a tweeter or two, if using a high-bandwidth amp, though

 

I'm curious as to what recordings.   

[Jud]

42 minutes ago, mansr said:

Specifics, please.

 

As mentioned in my response to Peter, Miska has said many of these filters ring.  See Fokus' description in point #2 in his comment here: 

 

Edit: One more from Fokus - see his comment here re aliasing in an SDM ADC:

 

 

[Jud]

13 minutes ago, mansr said:

All filters "ring," Miska's too. I want you to explain why you qualified downsampling filters used in typical DAWs as "bad." How do they differ from a filter you would deem "good"?

 

To be as precise as I can be (as a layperson) about what I'm asking:

 

Miska and Fokus have made some criticisms (both of ringing; Fokus of aliasing, see edit to my last comment) of typical ADC filters.  I don't know what decimation filters Cookie used for her DSD/44.1k comparison, or 2L used for its DXD/44.1k comparison; and I am supposing the decimation filter @bachish used might be different from either.

 

I would be curious to see how if at all these various decimation filters - typical, Cookie's, 2L's, the OP's - might change the measurements in the hi-res/44.1k comparison.

[Jud]

18 minutes ago, mansr said:

Nobody is talking about ADCs here.

 

OK, then I've managed to become confused about what goes on between when the signal hits the mics and when it becomes a CD or a 44.1k download (not an unusual thing for me).  Is the chain - sigma-delta conversion, then decimation - the same when @bachish records and in the music @pkane2001 is analyzing as occurs to produce the music and files Fokus and Miska were talking about, or not?

 

18 minutes ago, mansr said:

 

You made the assertion that the filters used by DAWs such as Pyramix when exporting to CD format are somehow "bad." Maybe they are, but the discussions you've referenced are about other topics and neither support nor refute this idea.

 

I have no idea what the difference is between what Fokus and Miska are talking about and what occurs with filters in DAWs (I never mentioned filters in DAWs, or Pyramix specifically, since I didn't know enough to do so).  I've read Fokus and Miska talking about ringing and aliasing, and in my mind the approximate sophistication level of the analysis that went on was something like "Ringing and aliasing bad!"

 

Mea maxima culpa.  Absolvo a me.

 

Good, now that we're hopefully done with this little side discussion about the state of my knowledge: What I'm curious about, for anyone who feels like being helpful, is whether and to what extent music in 44.1k from the "majors" has had different filtering used in its creation than the 44.1k examples 2L and Cookie have put up, and if so, what if any differences might be seen in the types of comparison measurements @bachish and @pkane2001 are doing.

[Jud]

And now we are well and thoroughly down the usual drain.  Paul and @bachish, if/when you'd like to post any more measurement comparisons or have further discussions about them, I'll await that with interest.

[Jud]

47 minutes ago, crenca said:

But such a thing requires access to both the "master" and knowledge of which specific software/DAW and the settings (filters, etc.) used.

 

If you were looking at a specific track, sure.  I was thinking more along the lines of using a generic filter, of the type that Fokus describes in the comments I linked, on one of the 2L DXD files to get a 44.1k file, for example, and comparing that to the 44.1k example posted by 2L, to see whether there's any measurable difference in the files.

[Jud]

1 minute ago, bachish said:

 

Yes, no doubt she is a great engineer. What she has described reminds me of what recording engineer Hudson Fair does, recordist for the Chicago Symphony Orchestra.  He also records to analog tape and mixes in an analog mixer. He says it sounds better than mixing on computer and, no doubt, it would sound different but subject to one's aesthetics. But, IMO, what gives Cookie her distinctive sound isn't so much tied to DSD vs PCM but what she is doing in the analog domain prior to converting to digital. That is where the magic happens. 

 

If interested, to see Hudson Fair's website see http://www.hudsonic.net/. Some interestimg pics there.

 

Thank you for the link. 

 

Yes, that's kind of what I was getting at in my initial comment about Cookie - that regardless of whether one agrees or disagrees with anything she says, it's indicative of a degree of care to produce a quality product that's borne out in the result.

[Jud]

Returning the favor of the Hudson Fair link - I've always liked the quality of the recordings Mark Knopfler's been involved with.  Here are a couple of links with some information about his (quite non-purist!) recording techniques:

 

https://www.premierguitar.com/articles/Producer_Chuck_Ainlay_on_Recording_Mark_Knopfler

 

http://guyfletcher.co.uk/the-diaries-2/recording-tracker-2014/

 

Edit: I saw your previous post after I posted my comment.  Looks like my "non-purist" remark was right in line with what you're saying.

[Jud]

Yeah, been going to the Infinitewav site for years.   

 

There are some other SRCs that don't look great on some of the tests.  But what I like it for is to visualize  things like the relationships between steepness, phase, and ringing, for example.

[Jud]

1 hour ago, bachish said:

Sound on Sound does have more in depth reviews but not all reviews are equally so.  I sent you a PM.

 

Apologies for the OT: did you happen to see the interview with Giles Martin about the making of the remastered 50th anniversary Sgt. Pepper?  https://www.soundonsound.com/techniques/inside-track-sgt-peppers-lonely-hearts-club-band

 

Fun quote:

 

"According to Okell, he and Martin began the remix process by listening to the 1967 mono and stereo mixes of each song, which would be at the top of each Pro Tools session. Okell would then recreate these as a starting point. 'There were no mix notes or anything like that, so it really was a question of listening carefully. We used Mark Lewisohn’s book [The Complete Beatles Recording Sessions] but he’s not an engineer and some details have been missed out, or are incorrect. Obviously, the content of each four-track tape was scribbled on the box, and they’d write ‘best’, and ‘4 to 4’ for a bounce, with the tape number it was bounced to, and the varispeed info, but that was it. And even these notes were sometimes incorrect, because things were changing all the time."