• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

More good results for the FA decoder!!!

John Dyson

Active Member
Joined
Feb 18, 2020
Messages
172
Likes
90
Just listen (sorry, must be snippets.)
The decoder is available for free -- and is probably the most complex piece of audio software in your quiver.

Demos:
https://www.dropbox.com/sh/tepjnd01xawzscv/AAB08KiAo8IRtYiUXSHRwLMla?dl=0

Software:
https://www.dropbox.com/sh/1srzzih0qoi1k4l/AAAMNIQ47AzBe1TubxJutJADa?dl=0

I suggest listening and comparing before commenting. Some people wont like the decoded version, but
most that I have run into, and my some 25 users do. (I have no marketing, so just word of mouth.)

Be kind -- but feel free to do full analysis, esp considering temporal behavior -- multi-band compressors/expanders are never flat on average, but
the consumer recordings are not flat below -20dB either.

I don't intend to insult anyone's library -- remember the 'digital sound' from the 1980s? People have gotten used to it -- but it is still there.

Enjoy --- the user base does...
 
OP
J

John Dyson

Active Member
Joined
Feb 18, 2020
Messages
172
Likes
90
Announcement about the difference between an original coming from the mix, and what the consumer gets...
This comes after 5yrs of work, probably not important to most people, but here IS the answer, unless someone else can give the reason. This is from a post elsewhere, so it wont look 'quite right':

MAJOR ANNOUNCEMENT -- meeting one of my personal goals... Probably not important to most people though....

After all of these years I have figured out the reason for the FA encoding!!!
Below, I'll explain the test used to determine the answer

ANSWER: FA mostly just blurs the signal in a subtle way... Probably for IP protection...

As simple as that: it blurs the signal.
How did I determine this likely fact?

1) Looking at the design, never figuring out how the compression should happen at a large scale, but compression IS happening at 10dB intervals, causing 'wobble'.
2) Just after getting the decoded result to match very well just *today*, I can NOT hear a difference when listening casually or when detailed stereo image isn't important to me.

However -- the original (normal consumer recording) signal IS blurred, and there IS a slight amount of compression.

What do I mean by 'blur'?

1) Adding DolbyA fog to the signal. This is a very, very subtle blur, but it very much covers up fine details.
2) Wobbles the stereo image.
3) Disconnects the HF from the LF, thereby causing the 'hash' that bothered me early on.

WHY DID IT TAKE SO LONG TO FIGURE OUT IT?
Decoding the mess requires massive extremely complex processing, and I never got it right till now. (less than 4 CPU cores need not apply.)
Reason for the long development: (hearing problems, not knowing the goal, frustrations from all kinds of sources including from those who are too skeptical.)

TODAY is the first time that the FA decoder, after immense amounts of processing, sounds like the original!!!

*** The difference between the original and the decoded version is that decoding subtantially removes BLUR.***

Release coming soon, after demos finish and after reviews give me the go-ahead....
(There might be one more iteration if reviews find a problem.)
 
OP
J

John Dyson

Active Member
Joined
Feb 18, 2020
Messages
172
Likes
90
Sound a bit too much like MQA.
Yes -- I do see a parallel, and our recordings have been infested with it since the middle 80's.
I truly am not sure if many/most people can hear the difference except:
1) slight tonal difference because instruments are better separated and not merged together.
2) maybe still an error in the bass, it is damned difficult to match the bass when there is likely bass rolloff AFTER FA encoding.
3) still not 100% happy with the sibilance, even though the descrambler phase does a good job of reconstituting HF detail.

The descrambler phase is very interesting, and the lack of descrambling was one of the reason for SOME of the problems in earlier versions. There were, of course, many other problems through the years.

About the descrambling action, it appears that some HF detail is bunched up into 7 segments (3kHz, 6k, 9k, 12k, 18k, 24k, 30k), and FA encoded material sounds similar to the original. Also, scrambling affords some HF compression. Expanding the detail is quite the trick, doesn't require conventional 'demod' but a rats nest of energy storing/releasing EQ.

The new phase adding the descrambler just happened about 1month ago when I gave up on trying to recover HF transients using the convetional layered DolbyA decoding scheme. Using some intuition and remembering that R Dolby also did a lot of video, tried using a video compression/expansion scheme, and it seemed to work. Of course, it took until now to find all of the 'settings'... Most all settings througout the decoder are based on the 3dB scheme (0.1875, 0.375, 0.75, 1.5, 3.0, 6.0)dB. It has been the correct plan, but like in the descramber, there needed to be 'tweaks' for best perf. There is a scaling parameter (1.0) that needed to be changed by <+-0.05 on the freq steps at/above 12kHz.

Sadly, the improved performance is SUBTLE, but true.
 
OP
J

John Dyson

Active Member
Joined
Feb 18, 2020
Messages
172
Likes
90
The decoder isn't quite as perfect as I hope, but is getting there and might actually be useful now.
What does it do? Lots of things to clean up recordings, but here is an especially egregious example from Olivia Newton John's, "Take me home, country roads". Suffice to say, these are only short snippets, and show an example of clarifying a complex chorus. This might not produce perfect results, but when listening to the entire recording, it melds very well, and does somewhat improve the experience. Other things like sibilance are also improved to some extent. The current bug might be slight (I mean, like 1-2dB) tilt up to 30kHz, which IS audible. (because the tilt starts around 12kHz.) I am trying to resolve the issue, and there ARE standard changes that follow a set of rules that I abide by, but somewhat confused by my hearing. As mentioned before, almost all EQ (like 99%) is based on standard values, and very very little 'tweaking'.

(Please try to avoid the Dropbox player -- it creates some blur by itself.)

Not decoded:

Decoded:

Again, not perfect, but is coming along and the output in general sounds very close to the FA original sans 'BLUR'.
Not everyone will like the decoded version, but at least this doesn't have compensation for cr*p headphones and my bad hearing anymore!!!
 
OP
J

John Dyson

Active Member
Joined
Feb 18, 2020
Messages
172
Likes
90
Finally, the 'FA' decoder really is working... Well, it recently has 'worked', but the final EQ was always botched -- but not so much now.
The magic answer was the 'descrambler', which really does some magic associated with EQ, HF signal expansion and things like that. The 'descrambler' allows sane presentation of the HF above about 4.5kHz.

Status: working reasonably well, sorely needs documentation update, perhaps 8-10 people know how to use it in current form. SW available, source will soon be available without asking. More useful docs coming in days, not weeks. Might even come in 'day'. If you try to use the decoder, remember to use the '--fa' switch, or, well, DolbyA mode doesn't sound very good for most recordings.

Location: URL: https://www.dropbox.com/sh/i6jccfopoi93s05/AAAZYvdR5co3-d1OM7v0BxWja?dl=0
All URLs given don't generally change, and haven't changed in probably over 1yr.


(ALL DEMOS ARE SNIPPETS -- IF YOU EVER SEE A FULL RECORDING IN MY PUBLIC REPOS, PLEASE TELL ME ASAP!!!)

Subdir for snippet demos: flacsnippet-REL9-V7.0B+5-0
Equvalent input snippets: same directory as above, but with INPUT in the name.
Decoder binary (soon matching uptodate source): https://www.dropbox.com/sh/5xtemxz5a4j6r38/AADlJJezI9EzZPNgvTNtcR8ra?dl=0
Speed: not fast, easily faster than realtime on the highest quality mode on my 10core. (Lots of precision Hilbert stuff going on for fog reduction.)
CAVEAT ABOUT THE DEMOS: don't expect anything major. the decoding result is subtle, direct A/B comparisons notwithstanding.

What are the final results of the working SW model?

Similar, almost indistinguishable response balance difference, freq balance short term sometimes a little different (esp LF).
More clean, precise sound. (HF, etc) Corrects 'fuzz' and 'blur' on consumer recordings.
Associated with the precise sound temporally with less 'wobble'/'fuzz', the stereo images esp on classical are less 'blurred', better locality.
Less hiss, esp on older recordings.
Different bass -- it is often different, but it appars to be actually correct. LF mixes differently with MF -- changes higher freq sound assoc with LF.
If there is a calibration offset (error), sometimes sibilance a little twisted/distorted, not usually horrible. In extreme cases, errors can cause a little 'garble'

How come it sounds so similar to the input, yet so much gain control & processing?

Whoever designed the scheme was a genius. Produces just enough damage to protect IP, but not so bad that most listeners are irritated from it.

Caveats:

The threshold of the layered gain curve is somewhat critical for the absolutely best, clean sound. Most recordings are created with threshold close to the correct settings, however, for example, some Supertramp uses a different/uncommon calibration level (e.g. 6 or 10dB difference, which is huge.) Most normal deviations of calibration are like: 0.005 or similar. My recent unmastered ABBA decodes were most happy with a calibration offset of approx '-0.008dB'. (On the other hand, the associated EQ on my attempt was ludicrous and embarassing. I should NOT have even tried to do any EQ. )

Is the decoder worth the trouble? Maybe...

Less so that I had hoped, but much of the time it can be worthwhile... Even Taylor Swift 'Shake it Off' is improved in a technical sense, but otherwise the recording is not worthwhile :) . Whoever designed the 'FA' scheme, I assume R Dolby, was truly a real genius to do such a precise kind of damage, without the result irritating most people. This 'FA' encoding is most likely a component of the 'Digital Sound' from the 1980s'. One can think of this 'FA' scheme, used since the 1980s, similar to MQA today, but more egregious and apparently totally secret. I have gotten nothing but pushback from those industry people who should know about it, but finally the decoder *IS* existence proof.

Will John (myself) be able to benefit fromusing the decoder?

Yes, because most of my hearing problem is associated with varying response balance. I can easily hear the damage that is corrected by the decoder. Much of the time, in the past during 'demos' , I would easily hear the corrections in the dynamics & hiss. Unfortunately for me, most other people were more distracted by the crazy resulting response balance. This is one reason in the past, I was very unpleasantly embarassed. I could hear the corrective actions, many other people were focused on the crazy response balance on the output. It would have been better to be able to communicate better with others, but the project has so much context and a lot of built-up frustration.

-----
Followup blather...

I don't think that too many people will be disgusted by the results now.

Traditional development tools & methods not always very applicable. This is NOT a simple fixed gain device nor has a fixed freq resp. Sine waves don't usually give useful results. This created quite the challenge. Near the end, started figuring out methods for 'estimating' the response curves, better avoiding the dependency on my lousy hearing. Still, mastering attempts are crazed folly, just that I sometimes am over-optimistic.

If I could hear well, this would have been working 2-3yrs ago AT LEAST.

Recently after finding that some kind of 'descrambler' step was needed, things got a lot easier. I could tell you long stories about how tricky this project was/is. At first, I didn't even know that DolbyA was involved, and THIS PROJECT is what spurred on the DolbyA project. Until recently, I did't even think about 'descrambling' and had to do a lot of research to understand a reasonable solution to the HF EQ problems. Simple EQ techniques can handle about 90% of the solution, but not much farther. Eventually figuring out the need for a 'descrambler' came from studying the spectral density and using my strange hearing.

BTW, the decoder for DolbyA encoded recordings is now MUCH MUCH more accurate. The FA project would still border on embarassing without the internal DA decoder fixes.

The source will be available in days, but it would take a person with double my IQ to figure it out. With the significant complexity and truly unique math/algorithms, the source needs to be cleaned-up and documented. To get past the 1st page or two, one would need to be a C++ expert also. My estimate is that 75% of the lines of code are SIMD operations in the guise of 'looks like normal math to me'.

Have at it... :)
 
Last edited:

Curvature

Major Contributor
Joined
May 20, 2022
Messages
1,069
Likes
1,341
What are you on about?

What is FA?

What is the problem you are trying to fix?

What is your solution?
 
OP
J

John Dyson

Active Member
Joined
Feb 18, 2020
Messages
172
Likes
90
What are you on about?

What is FA?

What is the problem you are trying to fix?

What is your solution?
There is a rather standard kind of 'processing' done to many/most/all consumer recordings.
The best way to describe the side effect is 'blur'. Effectively a minimally noticeable 'bunch of generation loss'.
The ubiquitious scheme uses compression as an element (but not a goal) and some special EQ techniques to be minimally audible, but still audible creates damage to the signal.
Purpose: very likely to provide plausibly good quality to consumers, while keeping the best-quality source IP private.
Time frame: appears to have started in the middle 1980s, and very likely the original reason for complaints about 'digital sound'.

Side effects: Subtle blur/fuzz of middle highs, stereo image less focused, a bit of modifying LF/MF relative levels, 'generally less clean' sound. (Probably more that I have forgotten.)

The method also has the tendancy to increase the loudness of tape hiss and other noise sources, but is NOT normal use of dynamic range compression nor limiting. For example, if you listen to many older recordings, the hiss that you hear is louder the natural level of tape hiss. Even though there is DRC in the process (therefore decoded by carefully crafted expansion elements), the device is not, as a whole, a 'compressor'. The tape hiss is increased as in compression, but actually more as a side effect of certain elements in the 'blurring' process.

Also appears to have a bit of nonlinear dynamic processing of the HF > about 4.5kHz (the knee for that step is actually 6kHz.) This HF scrambling reminds of certain video processing methods. The descrambling method is also 'existence proof'.

The processing is intricate, and mostly denied by the industry. But there is existence proof that the processing happens.

ADD-ON: the project mostly fixes the damage, and as a bonus, helps to remove whatever DolbyA fog created back in the studio. It is NOT a simple expander, EQ or anything trivial like that, perhaps the most complex piece of software used in audio (including a big DAW with all kinds of toys with it.)
 
Last edited:

dc655321

Major Contributor
Joined
Mar 4, 2018
Messages
1,597
Likes
2,235
There is a rather standard kind of 'processing' done to many/most/all consumer recordings.
The best way to describe the side effect is 'blur'. Effectively a minimally noticeable 'bunch of generation loss'.
The ubiquitious scheme uses compression as an element (but not a goal) and some special EQ techniques to be minimally audible, but still audible creates damage to the signal.
Purpose: very likely to provide plausibly good quality to consumers, while keeping the best-quality source IP private.
Time frame: appears to have started in the middle 1980s, and very likely the original reason for complaints about 'digital sound'.

Side effects: Subtle blur/fuzz of middle highs, stereo image less focused, a bit of modifying LF/MF relative levels, 'generally less clean' sound. (Probably more that I have forgotten.)

The method also has the tendancy to increase the loudness of tape hiss and other noise sources, but is NOT normal use of dynamic range compression nor limiting. For example, if you listen to many older recordings, the hiss that you hear is louder the natural level of tape hiss. Even though there is DRC in the process (therefore decoded by carefully crafted expansion elements), the device is not, as a whole, a 'compressor'. The tape hiss is increased as in compression, but actually more as a side effect of certain elements in the 'blurring' process.

Also appears to have a bit of nonlinear dynamic processing of the HF > about 4.5kHz (the knee for that step is actually 6kHz.) This HF scrambling reminds of certain video processing methods. The descrambling method is also 'existence proof'.

The processing is intricate, and mostly denied by the industry. But there is existence proof that the processing happens.

ADD-ON: the project mostly fixes the damage, and as a bonus, helps to remove whatever DolbyA fog created back in the studio. It is NOT a simple expander, EQ or anything trivial like that, perhaps the most complex piece of software used in audio (including a big DAW with all kinds of toys with it.)

Evidence of the “blur” claim?
 

RandomEar

Senior Member
Joined
Feb 14, 2022
Messages
331
Likes
765
...perhaps the most complex piece of software used in audio ...
[Press (X) to doubt]

I have read the posts in this thread. I still don't know what it is about. Some "secret" audio processing in all studios, to protect IP? What IP? If I have the recording and can listen to it, then what is there still to be protected? Why would this be secret in any way? There must be literally dozens of sound engineers on this forum, who would have to know this.

I don't get what the original problem is, that this "decoder" is meant to solve. I don't understand what it does. I'm sorry, but I can't see any congruent information in this thread - Do I simply lack the technical knowledge, or is this just mumbo-jumbo?
 

PierreV

Major Contributor
Forum Donor
Joined
Nov 6, 2018
Messages
1,437
Likes
4,686
The source will be available in days, but it would take a person with double my IQ to figure it out.
Oh, that wouldn't be a problem. Once I realized I had cognitive issues and my thoughts were a bit blurry, I developed a cortical descrambler that restored my brain's full potential.
And, as a bonus, I could stop using my tinfoil hat.
 

dc655321

Major Contributor
Joined
Mar 4, 2018
Messages
1,597
Likes
2,235
Oh, that wouldn't be a problem. Once I realized I had cognitive issues and my thoughts were a bit blurry, I developed a cortical descrambler that restored my brain's full potential.
And, as a bonus, I could stop using my tinfoil hat.

I took that comment to mean he’s terrible at writing software. But your response is way better :D
 
OP
J

John Dyson

Active Member
Joined
Feb 18, 2020
Messages
172
Likes
90
Evidence of the “blur” claim?
I'll select some of the better demos -- in general, the demos that I provide are both 'best' and 'worst case' and 'inbetwen'.
The 'blur' (and I don't have a better way of describing it) is a combination of a lot of subtle changes. It produces a lack of precision in the recording.

I don't know how many of the readers were around when there were complaints about 'digital sound' back in the 1980s -- I was, and was an engineer at Bell Labs who understands ALL of that stuff. Digital was NOT so bad for a person accomodated to analog to be able detect things like Gibbs effect or other digital thingies (or even the assocated clipping.) BTW -- Gibbs is NOT ringing. However, there was *something else* going on. Please refer to the Dire Straits CDs from that period. Those are often considered definitive nowadays, and the damage in the recordings was NOT digital per-se.
 
OP
J

John Dyson

Active Member
Joined
Feb 18, 2020
Messages
172
Likes
90
[Press (X) to doubt]

I have read the posts in this thread. I still don't know what it is about. Some "secret" audio processing in all studios, to protect IP? What IP? If I have the recording and can listen to it, then what is there still to be protected? Why would this be secret in any way? There must be literally dozens of sound engineers on this forum, who would have to know this.

I don't get what the original problem is, that this "decoder" is meant to solve. I don't understand what it does. I'm sorry, but I can't see any congruent information in this thread - Do I simply lack the technical knowledge, or is this just mumbo-jumbo?
The processing is likely not a 'secret', but is probably not well known. I am not into conspiracy theories or anything like that, but rather a hard-core engineer, less of an audiophile nowadays.

With earlier history doing recordings, I knew what studio sound was like. Also, I respected/enjoyed the direct to disk recordings from Sheffield Labs/etc. In fact, back then, vinyl generally sounded reasonably good, with the expected defects. Then, later on in the 1980s, as a Bell Labs engineer not easily fooled, I detected something wrong with many (most/all) digital consumer recordings that I found. It wasn't 'digital sound' because that claim was somewhat specious not even describing the recording defect at all. However, something was wrong with the dynamics of almost any digital recording that I found. To me, the defect was distracting enough to start decreasing my interest in HiFi, because true HiFi was hard to find. Then, my CS and Video hobby started. The idle recreation time/resources encourged development of a previously well known OS, and playing with my D9/DVCAM/SVHS/LD video equipment. HiFi hobby diminished not nearly zero.

In about, 2011/2012, when I had some time, I happened across a CD, noticing the audio problem again. THAT started the quest about WTF is causing the damage. With experience in dynamics processing anyway, I did a quick analysis and noticed some curious processing going on... Took a few years for a first step about what was going on, but understood the problem very early. Was also dissuaded and misdirected by skepticism assocated with cluelessness and possibly in some cases, diversion.

Answering other comments: Not into tinfoil, instead I am a hardcore DSP/CS/EE, not into conspiracy theories, and the only people out to get me are certifiably insane, and those who are prejudiced or ignorant are to be pitied. Otherwise, all is nice and comfortable!!!
 
Last edited:
OP
J

John Dyson

Active Member
Joined
Feb 18, 2020
Messages
172
Likes
90
About the demos... When they are provided in a day or two, they will be more 'best case' examples so that the improvement is easier to detect/hear. My normal demos are intended to be 'honest' from a statistical standpoint and not intended really to 'sell' anything. On the other hand, when there are strongly skeptical people, they'll sometimes pick the 'worst case' to argue against an improvement. There will always be cases where such processing will give minimal or even no benefit. Sometimes even the distributed/encoded consumer materials will sound more 'normal', but not actually be 'natural'. In a general sense, decoded/cleaned-up recordings are not always the best to listen to.

So, by approx Weds, I'll have some 'best case' examples, maybe even some mixed cases, where there is technical improvement, but so what?

Skepticism is always okay, if it is associated with open mindedness and a dash of wisdom. I have been my own worst enemy also, from excess enthusiasm but also misdirected by unreliable hearing and not knowing the actual project goal. Perhaps another flaw: programming, EE and DSP are very easy for me -- the application is NOT easy for me though.
Originally getting started: there was a definite observation about 'dynamics processing' and the quality being not as good as it should be. So, I defined the nebulous goal to recover the original recording. Frustratingly, I did not know what a corrected recording sounds like. It has been many years since I was involved in audio (all the way from recording, to audio, to Bell Labs consumer products engineering research.) Back in the past, and now, I could always hear that the recording was damaged, didn't really know the recording 'behind' the damage. At first, didn't even know the exact character of the damage other than there was some damned fast attack/release dynamics processing somewhere in the mix.

So, demos in a few days... More/different examples might also be helpful later in the process.
Suffice to say: the decoding REALLY DOES work, the decoding isn't always perfect, but is very close to being so.
 
Top Bottom