DolbyA decoding feedback -- 'feral' examples (yes/no)

[sandyk]

4 hours ago, John Dyson said:

Listen to 'please mr please' or 'Let me be there' -- very closely.

John

At  a quick listen K1 is noticeably better with the both tracks that you suggested.

More open sounding right from the start , albeit a little softer sounding , with her voice cleaner sounding too. (first converted from .flac to .wav)

 

Regards

Alex

[pkane2001]

6 hours ago, John Dyson said:

 

 

Okay -- the demos with the 'K1' and 'K1A' suffixes do not have the bumps that you are talking about.  Also, I spent more time in setting the calibration level -- anyone listening should get the sense that every last bit of detail on the recording is being reproduced (in the flac version.)   All of the errors were in my sox script and the calibration setting, not in the decoder.

 

In the K1* versions, I applied no sibilance correction (there is often a bump in the 9kHz range which is artificial sibilance -- in the leaked DolbyA material.)   I don't know if the 9kHz bump is intentional, or it results from enhancement causing a 'friction' distortion between the HF0/HF1 bands in the DolbyA HW encoding.  (The HF0/HF1 bands actively interact in the 9kHz range -- the dance must be precise.)

 

The calibration change brings out a lot more detail also.

 

The detail in the results are closer to what a master tape will provide, but still the leaked DolbyA consumer stuff is not 100% trustworthy, so sometimes there is a little excessive 'love' added.   Significant phase errors can be troublesome even though the DHNRDS uses phase insenitive detectors, still too much EQ done by the 'mastering' can cause phase shifts with untoward effects.

 

Listen to 'please mr please' or 'Let me be there' -- very closely.  Methinks that every last bit of detail in the recording (modulo the 16 bits and 48k) is reproduced there.  (There is a certain 'sweet' sound that happens when the DA decoder is set correctly.)  When the HF0/HF1 dance lines up perfectly -- the sound is astonishingly good.  (It is sometimes tricky to get that precision on feral material.)

 

https://www.dropbox.com/sh/d2dyuonx8fk3pwv/AAArjQgYoo_Skpy8YSiOuxpha?dl=0

 

John

 

 

Hi John,

 

K1A Country Road is a different portion of the track, makes it harder to compare to the undecoded and earlier versions.

K1 version does look much cleaner now (blue is undecoded):

 

Can't tell yet if I can hear the difference, as I only listened to these on my Mac built-in audio/speakers.

 

[John Dyson]

35 minutes ago, pkane2001 said:

 

Hi John,

 

K1A Country Road is a different portion of the track, makes it harder to compare to the undecoded and earlier versions.

K1 version does look much cleaner now (blue is undecoded):

 

Can't tell yet if I can hear the difference, as I only listened to these on my Mac built-in audio/speakers.

 

The distortion should always have been 'clean', but the frequency response was damaged by my stupid sox scripts.  (A DolbyA hardware during decoding does create odd distortions/dynamics effects  -- the DHNRDS cannot produce the same kind of distortion.*)   The small decrease in HF in your diagram is to be expected, even though most of the gain reduction should be at the instantaneously lower levels.  At 0dB, the DA doesn't decrease the signal, but at -40dB, the resulting signal will be -50dB or -55dB (on a per-band basis.)  So, a DolbyA decoding device will be flat at 0dB, but (on the MF band) will start decreasing the gain at about -20dB input level.

 

* The DHNRDS design 'sort of' has the DolbyA detector as almost a plug-in.   With a minor source modification, the DolbyA attack/release can be removed, and it is possible to run the software TOTALLY without the slow-down of the attack/release.  Basically, the level detection can directly drive the gain control without the filtering of attack/release -- only the anti-MD aspects being enabled.  The DHNRDS can do 'infinitely fast' attack/release without casually noticeable distortion (the grinding effect from effectively 1msec attack/release doesn't exist!!!)   It is REALLY weird that way.  Almost any other expander with 1msec attack/release WILL sound distorted (because of modulation effects.)

 

The only issue:  natural HF0/HF1 dance does cause a deadening effect, but the increase in distortion to run with a 1msec attack/release is apparently very minor.  The HF0/HF1 dance is built-in to the design, disabling it for the experiment, would have required more changes.  I added the crazy-heroic anti-MD capability because of anecdotal reports about loss of detail on DolbyA encode/decode operations.  That loss of detail is caused by an effect with similar sound as tape saturation.

 

The DA decoder infrastructure is being adapted for the Telcom C4 -- but with the post-anti-IMD and trajectory shaping disabled.  Only the input-anti-IMD code will initially be enabled.  However, if we determine that the C4 decoding results will be improved by the extreme anti-MD code, then I'll re-enable it.  The C4 project will likely be used much less often - so I don't want to invest the mega-heroic efforts and extreme amount of time that the DA effort required.

 

The DHNRDS DA software wasn't originally intended to run without the dynamic DolbyA attack/release, but by disabling the DolbyA attack/release math,  it can show that the dynamic distortions simply cannot exist when using the DHNRDS DA design.  Of course, insanely bad use of the gain control signal (bad calibration) can produce something that sounds like distortion, but using the DHNRDS as a simple expander without the DolbyA parameters was amazing (I did the bypass mentioned above on a lark -- just to see what happens -- oddly, it really worked well.)  Most of any slightly apparent distortion comes from the HF0/HF1 bands interfering with each other -- necessary for proper DolbyA decoding, but bad for GP expansion.  When doing a DolbyA decode -- it is important that the HF0/HF1 bands dance together perfectly, because of the original source material and DolbyA decoding device need to sychronize for best clarity/less grain in the sound.   A true DolbyA cannot synchronize perfectly, but instead has a looser match to the attack/release, thereby softening the detail.

 

* I actually added an experimental 'extra NR' capability where perhaps 1 to 1.1 trhough 1 to 1.5 dB additional expansion can be added with the 'infinitely fast ' (1msec) attack/release.  Frankly, it deadens the sound too much -- but might be helpful in extreme cases of hiss.  Maybe the feature should be better optimized, or reserve an actual implementation of the 'infinitely fast' attack/release expander for a separate program?

 

John

 

[fas42]

Just listened to the undecoded , K0, and K1 flac versions of Have You Ever Been - the RAW is the closest to hearing what was in the recording space, K0 and K1 progressively move towards hearing it via an AM radio - IOW, the sparkle and life of the accompanying instruments, etc, is being deadened just a bit more each time.

[sandyk]

6 minutes ago, fas42 said:

Just listened to the undecoded , K0, and K1 flac versions of Have You Ever Been - the RAW is the closest to hearing what was in the recording space, K0 and K1 progressively move towards hearing it via an AM radio - IOW, the sparkle and life of the accompanying instruments, etc, is being deadened just a bit more each time.

 

 The RAW file is helping to compensate for glaring deficiencies in your laptop setup !

[John Dyson]

12 minutes ago, fas42 said:

Just listened to the undecoded , K0, and K1 flac versions of Have You Ever Been - the RAW is the closest to hearing what was in the recording space, K0 and K1 progressively move towards hearing it via an AM radio - IOW, the sparkle and life of the accompanying instruments, etc, is being deadened just a bit more each time.

We can disagree -- that is okay.   The 'sparkle' is called HF compression - too many people are used to it, and it is regrettable.

Additionally, the difference of only a dB or so at the high levels kind of argues against 'closer to AM radio'.  AM radio is MORE compressed, while the DA output has 10dB of active expansion between 3k and 9k and 15dB of VERY ACTIVE expansion between 9k and 20+kHz.   The both bands have gains hitting 0dB much of the time, which shows that the peaks care more clean & distinct, rather than 'bunched up' as in compressed material.   (There is also expansion at the low end -- while the MF band is only active at very low levels as specified by DolbyA behavior.)

 

Remember:  the sadly leaked DolbyA material in the 1980s were the beginnings of the very evil loudness wars in distributed music starting in the  1990s.

 

John

[fas42]

28 minutes ago, sandyk said:

 

 The RAW file is helping to compensate for glaring deficiencies in your laptop setup !

 

Alex, the key glaring deficiency of high end rigs I come across is that the treble is badly handled - so something I prize highly is anything, no matter how "basic", that gets this area right - listen to a recording with, say, a trumpet in full flight, being driven hard by a player going for the big notes - "ambitious rigs" typically make this sound like the aim is to torture the listeners into quivering submission, just to escape from the distortion, .

 

Doesn't have to be this way - get a system into good shape, and then it makes sounds just like hearing the real instrument being revved to the max. Which, strangely enough, means that everything you listen to just "sounds right". Every time. And I reckon that's not too bad a thing, really ... .

[kumakuma]

1 hour ago, fas42 said:

Just listened to the undecoded , K0, and K1 flac versions of Have You Ever Been - the RAW is the closest to hearing what was in the recording space, K0 and K1 progressively move towards hearing it via an AM radio - IOW, the sparkle and life of the accompanying instruments, etc, is being deadened just a bit more each time.

 

Wow. I couldn't disagree more.

 

The "sparkle" in the RAW version sounds quite unnatural and started to make my head hurt after just a few minutes of listening.

[fas42]

9 minutes ago, kumakuma said:

 

Wow. I couldn't disagree more.

 

The "sparkle" in the RAW version sounds quite unnatural and started to make my head hurt after just a few minutes of listening.

 

My aim is to recover the impact of hearing the sounds as if you were in the actual space where the recording took place - the higher the treble content of that sound, the worse it will sound if it is not as accurately reproduced as possible. Meaning, that "sparkle" will hurt and sound unnatural if not well handled - the end result of making this happen as well as is possible, is that replay has all the bite and transient impact of live music making - I listen for "sparkle" being unnatural as the distortion giveaway that the system is not working right; every time you resolve the underlying issues, listening becomes effortless, and the "sparkle" is now doing its job as it should.

[kumakuma]

13 minutes ago, fas42 said:

 

My aim is to recover the impact of hearing the sounds as if you were in the actual space where the recording took place - the higher the treble content of that sound, the worse it will sound if it is not as accurately reproduced as possible. Meaning, that "sparkle" will hurt and sound unnatural if not well handled - the end result of making this happen as well as is possible, is that replay has all the bite and transient impact of live music making - I listen for "sparkle" being unnatural as the distortion giveaway that the system is not working right; every time you resolve the underlying issues, listening becomes effortless, and the "sparkle" is now doing its job as it should.

 

I just transferred two of the ONJ songs to my high-end digital audio player and listened to them again using both high-quality IEMs and headphones.

 

I stand by my previous post. The RAW versions are both unnatural and unpleasant to listen to.

[fas42]

And that's fine ... I also just tried again, and the processed ones, by comparison, turn the instrumental backing into a bland backdrop; they were far less interesting as a musical element, meaning that the vocals became the focus - this is why I use the AM radio analogy.

 

I find the marvellous thing with all the classic pop recordings that we all know inside out, is that with clear reproduction the vocal lines which were everything when we learnt them by heart, are now joined by the instrumental lines; how each instrument has a special quality in the whole, and has an interest factor which is just as great as the obvious pop lyrics, and melody. This is the "surprise factor" which lifts the most banal songs of the era up many notches, and makes them a constant delight to listen to.

[John Dyson]

1 minute ago, fas42 said:

And that's fine ... I also just tried again, and the processed ones, by comparison, turn the instrumental backing into a bland backdrop; they were far less interesting as a musical element, meaning that the vocals became the focus - this is why I use the AM radio analogy.

 

I find the marvellous thing with all the classic pop recordings that we all know inside out, is that with clear reproduction the vocal lines which were everything when we learnt them by heart, are now joined by the instrumental lines; how each instrument has a special quality in the whole, and has an interest factor which is just as great as the obvious pop lyrics, and melody. This is the "surprise factor" which lifts the most banal songs of the era up many notches, and makes them a constant delight to listen to.

If you like the sound, the buying a compressor would help (or side chain with EQ -- to emulate the HF compression.)  They even sell fake-DolbyAs for vocal enhancement -- but please, NOT ALL OF THE TIME!!!

 

I really do suggest adding a sidechain with EQ -- I was having a discussion with someone in a bar one day (many, many, many years ago) -- he was from a culture who would still be dancing to disco music today...  He LOVED the sound of dynamic range compression -- it gave him a bit of buzz, and that is okay because sometimes we find certain things that we like.  I think he even had a compressor for his personal system (he was wealthy & indulged from the family who started Cummins engine -- a person with an interesting and VERY different background than myself.)

 

My 'buzz' isn't even so much for listening, but my current thing is to support/preserve the legacy of a brilliant time of the music scene, so that people in the future will be able to refer-to and hear music as it was originally produced.   That REALLY is my current interest.  If there is a compressor in the way between me and the original production -- then I don' t like it for the purpose of my project.  It was a big turn-on when we started getting interest from the music libraries (of the corporate/government levels.)  THAT is the goal for me.  The ability to present the material in a technically ACCURATE, historical quality context has been a goal since the DHNRDS project started.

 

Similarly, the DolbyA sound that is often being listened to today was not likely even heard by the original recording engineer (unless he was troubleshooting a problem.)  That incessant  'compressed' sound appeared because of serendipty -- unintentional, and just the way that things happened.  But that sound (even if that is something that I liked to listen to)  is NOT my own goal.   However, the 'sharp', almost 'raspy' sound of Karen Carpenters later recordings (Albums 1971 and after), sometimes resulted from ab-using DolbyA units on her vocals.  Would I choose to do that today?  Certainly not, because there are more clean, better ways to get some enhancement today*

 

* Probably one reason why they did use DolbyA in the past, and still some people wishing to do things the old fashioned way still use DolbyA encoder emulation. Why do that do that?:  it doesn''t sound as bad as a simple compressor with similar parameters, and R Dolby did an ingenious design.  (If someone is really interested in what Dolby did, I can do a paragraph or two write up.  The design does show that Dolby had an uncommon understanding of semiconductor electronics and the way that hearing works in the early 1960s.  Many *even competent* engineers didn't start seriously developing that level of understanding until the 1970s -- he was really ahead of his time.)   I'll even assert that it is very likely that few -esp younger ones- EEs today (who even design low noise preamps and the ilk) understand the circuitry to the level of what R Dolby actually implemented some 50+yrs ago.

 

 

John

 

[fas42]

I'll just point to this thread on Gearslutz, where those who are skeptical, make the sort of points I would make if I were in the industry - https://www.gearslutz.com/board/mastering-forum/1265710-about-dolbya-why-seems-lurk-consumer-realm-2.html .

 

I have no interest in using the excuse that because playback is harsh, or unnatural on some systems, that there is an inherent fault of Dolby A not being properly decoded - the recording is what it is, and before ascribing subjective unpleasantness to poor mastering, I would need to hear the qualities of whatever system was being used to subjectively assess the 'feralness' of some recording ... sorry about that, John, .

[sandyk]

31 minutes ago, fas42 said:

I would need to hear the qualities of whatever system was being used to subjectively assess the 'feralness' of some recording ... sorry about that, John, .

 Frank

 If you had been reading what John has been saying, you would realise that he is actually able to readily MEASURE the differences due to the lack of Dolby A decoding. The tricky part is where they attempted to remaster a Dolby A recording without using the original master tapes with their calibration level tones, more often than not botching it up.

Alex

[fas42]

51 minutes ago, sandyk said:

 Frank

 If you had been reading what John has been saying, you would realise that he is actually able to readily MEASURE the differences due to the lack of Dolby A decoding. The tricky part is where they attempted to remaster a Dolby A recording without using the original master tapes with their calibration level tones, more often than not botching it up.

Alex

 

Alex, one thing I haven't seen is any "ready measurement" - if the recording deliberately used processing, including Dolby A itself as an intended effect, where does one separate the "valid" processing from the "invalid" processing?

[sandyk]

1 minute ago, fas42 said:

 

Alex, one thing I haven't seen is any "ready measurement" - if the recording deliberately used processing, including Dolby A itself as an intended effect, where does one separate the "valid" processing from the "invalid" processing?

 

 A good question. However, there would be very few Recording Engineers that would do that, as most material would sound terrible ,except perhaps some original Abba recordings from Polar.

[John Dyson]

3 hours ago, sandyk said:

 

 A good question. However, there would be very few Recording Engineers that would do that, as most material would sound terrible ,except perhaps some original Abba recordings from Polar.

I do know that some recording engineers initially responded (including my project partner) with a bit of indignation that DolbyA leaks occurred.  It would be considered to be really 'bad form' in the past to blindly use DolbyA en-masse on the recording.   Some of those rather somewhat 'chortling' engineers have come around to being friends with me.   It would not have even been thought of to use DolbyA for anything but an enhancement.

 

I CAN say that the irritating pop examples 'Shake it Off' by the self-absorbed Taylor Swift and the song 'Call me, Maybe' from Carly Rae Jepson do APPEAR (I write APPEAR) to have been so very highly compressed that the DHNRDS can make really clean sense of the material (you can hear details in 'Shake it off' that were obscured by the modulation effects of the extreme compression after 'processing' with the DHNRDS.)   They might be using a DolbyA or DolbyA emulation for these hyper compressed pop recordings - but they are really unnatural sounding recordings anyway.  Below are beginning snippets of DHNRDS 'decoded' copies of the songs -- but the quality standards being technically excellent in some ways, show poor taste wrt the extreme compression.

 

*More and more knowledgeable people nowadays are accepting of the fact that the DolbyA leaks occur (ALOT), especially once the trick of EQ being used as a step  in the ersatz mastering process.  In fact, a guy doing 'greatest hits' releases actually wrote me on a public forum that the leaks have happened 'A LOT'.

06-Shake It Off-A.mp3 03 Call Me Maybe.mp3

[sandyk]

Hi John

 Taylor Swift may be easy on the eyes, but I don't find her too easy on the ears though.

 Perhaps an acquired taste ?

 Regards

Alex

 

[John Dyson]

6 hours ago, fas42 said:

I'll just point to this thread on Gearslutz, where those who are skeptical, make the sort of points I would make if I were in the industry - https://www.gearslutz.com/board/mastering-forum/1265710-about-dolbya-why-seems-lurk-consumer-realm-2.html .

 

I have no interest in using the excuse that because playback is harsh, or unnatural on some systems, that there is an inherent fault of Dolby A not being properly decoded - the recording is what it is, and before ascribing subjective unpleasantness to poor mastering, I would need to hear the qualities of whatever system was being used to subjectively assess the 'feralness' of some recording ... sorry about that, John, .

 

Some people like crummy sound -- and that is okay Frank.

 

Thank you about the pointer to that forum -- one of the 'skeptics' on that forum has gotten a DHNRDS decoder -- no complaints so far :-).

 

 

John

[John Dyson]

15 minutes ago, sandyk said:

Hi John

 Taylor Swift may be easy on the eyes, but I don't find her too easy on the ears though.

 Perhaps an acquired taste ?

 Regards

Alex

 

Hah!!  I agree about the looks, but my irritation with her is also at the intellectual & social maturity level.   Spoilt little child and arrogant teenage girl comes to mind.

 

The 'interesting' thing about 'Shake it Off' and apparently other material of that ilk (or YUCK!!!) is that they use lots of HF compression that some dead-ears (probably need the boost after being desensitized) like to hear.

 

Frankly, it is apparant that some people become desensitized to the subtle nature of certain recordings -- after too much HF compression and boost.  I really try to avoid stuff like 'Shake it Off' and 'Call Me, Maybe' (dont' have to try hard to avoid actually) because it does desensitize hearing.  PLEASE DON"T LISTEN TOO MUCH, EVEN TO THE CLEANED-UP VERSIONS.   That stuff is aural poison, and actually worse than traditionally leaked/feral DolbyA.

 

John

 

[John Dyson]

Given this is a similar subject -- I have an example that shows the 'DolbyA' decoding fog, vs the DHNRDS 'much less' fog.  It isn't the most extreme, but it is an example that I just ran across.  It is from one of the more 'disparaged' groups, so I made it very short (so hearing isn't polluted :-)).  But also, this is not cherry picked -- this example isn't the most extreme case, just convenient...

 

John

fog.flac nofog.flac

[John Dyson]

Another DolbyA fog example... (JUST FYI -- 'fog' is one major reason why I wrote the DHNRDS.)

There are small differences in compression, but the slight additional compression used on the commercial version is NOT the cause of the fog -- unless it was VERY sloppy compression.   That level of 'fog' is the evil 'modulation distortion' that the DHNRDS tries to solve.

 

John

 

 

 

fogV1.flac nofogV1.flac

[fas42]

'Fraid I like "crummy", "foggy" sound, John  - the qualities that make ABBA special are significantly reduced in the nofog versions, and especially the lilt in the voices has been lost to an unacceptable degree - I won't be a customer, sorry.

[sandyk]

47 minutes ago, fas42 said:

'Fraid I like "crummy", "foggy" sound, John  - the qualities that make ABBA special are significantly reduced in the nofog versions, and especially the lilt in the voices has been lost to an unacceptable degree - I won't be a customer, sorry.

 

 Perhaps they are when using a Laptop with tiny speakers, but this is an Audiophile forum where most members use something way better for their main listening sessions, and most prefer to hear them as the original Recording Engineer intended them to sound, NOT due to a stuff up later on  by a mastering Engineer.

[fas42]

So you've got a good handle on what Michael Tretow was after, then?