Announcing a new DolbyA decoder -- more useful than you might think

John Dyson

Member
Jul 2, 2018
41
1
13
I am making a new DolbyA decoder (finally working nearly perfectly) that is good for removing that so-called digital sheen or harsh compressed sound so common in older recordings (usually from before the 1990s.) Before thinking -- naw, that cannot have happend, well the best thing that I can say is to try the decoder on some of the most harsh sounding recordings. You might not 'long' anymore for vinyl aafter listening to the results of a digital copy of an old recording with the decoder. Some caveats is that the decoder uses a LOT of CPU, for example it takes a whole core of my Haswell CPU to run realtime at 48k or 96k. It also runs directly at the sample rate that you are using -- and it doesn't do very well at 44.1k or much above 96k, even though it DOES work at 176.4 and 192k, but at slightly lower quality. Best sample rate is 64k or 72k, but that requires using something like sox to do the conversion. (There are numerous DSP reasons why 64k to 72k are best, but that is the sweet spot.)

The spectograms and spectrums match a real DolbyA 360 very closely, and actually produces less harmonic (of course) and intermodulation distortions than a real DolbyA. Some of the superior distortion performance results from carefully crafting the attack/decay and performs the gain control operations in a more complex and less distortion way than just multiplication or division.

There is no 'digital' or 'computer' sound endemic to this processor, and the amount of material available is amazing. I even have some stuff from a big 'high quality' digital recording supplier that needs decoding.

Cost: nothing other than CPU time on your fairly recent Intel Box running Windows 10, Linux available on request, even though Linux is the actual development and code testing platform. Availability: Now -- for download.

I can provide lots of deep technical details, but this is NOT a direct implementation of the HW DolbyA -- that is almost impossible, but is a re-engineered design that actually works and sounds VERY similar to a real DolbyA.

This is ONLY a decoder, because I doubt there is really much interest in much more DolbyA encoding, and I didn't want to spend the time to do it.

Distribution - from my repository, updated whenever needed (sometimes a couple of times per day when I make mistakes). Once it has sit on the repo longer than a day or so, it is stable.
The program is free, no nagware, no timeouts -- just a program. The file is a zip file with needed dlls. Uncompress the zipfile (usually just with the windows interface), and place all of the
constituents on the directory where you want to keep and use it.

Help - just post a question, answer will come as quickly as possible.

Examples -- on the repository, look for simple filenames for the processed version, Filenames starting with ORIG- for the pre-decoded versions. The examples will only be there for a few days, but the results are impactful.

In the pre-decoded versions, listen for hiss (usually worse at the beginning), the harsh 'digital' sound, and too much compressed treble.

For the decoded versions, listen for often lush and clean sound. Be careful to listen to the decoded versions first -- if you listen to the originals
first, your ears will sometimes adjust to the harsh sound.

Repository location: https://spaces.hightail.com/space/tjUm4ywtDR

Have fun!!!
John
 

John Dyson

Member
Jul 2, 2018
41
1
13
This is a really, really good update. I am getting into the tweaks and tuning towards perfection. The code is stable now, expecting to make it a real 'release' in a few weeks. It really needs lots of testing and usage before blessing it. Also, if this idea of errant 'DolbyA' encoding becomes better accepted, I plan to do what is necessary to help transfer some technology to those who can better implement a GUI or plug-in application. This tool is not trivial, especially difficult to make it so very accurate in SW.

Here are the two significant changes:
1) re-instituting the attack prefilter. The attack times are the same, but now the distortion sidebands are minimized before being created, rather than just being postfiltered. This would be most interesting to the 44.1k and 192k users, but the improvement to 48k-96k should be noticeable in this 'easy' sample rate range. I find my VERY difficult material to be improved. As usual, if you run at 72k (e.g. using sox), the program does a slightly better job. 96k is second best, but takes 96/72 more time.
2) better inter-channel timing for attack/decay. Before this correction, the threshold had to be set very accurately to avoid a very minimal phasing on strong 's' sounds. It was noticeable on Scarborough Fair and the Pink Panther theme. Now, changing the threshold over a fairly wide range -- even though incorrect -- should not have as strong a phasing type sound. (It really wasn't 'phasing' but an obvious change in freq resp.)

I also have a warning about using SOX... When you have a very hot .flac file on input, and you use the technique of using a gain of -6dB before the decoder, and then use '--ingain=6' in the decoder command line -- it is actually better to use the '-v 0.5' switch on SOX instead of 'gain -6.' I was finding that 'gain -6' doesn't avoid potential mild clipping on input for very hot .flac files.

As recently usual, the location of the decoder is below. The examples have been updated.




https://spaces.hightail.com/space/tjUm4ywtDR
 

John Dyson

Member
Jul 2, 2018
41
1
13
Update on the DolbyA decoder. Well good news, and some intelligence from the 'mastering' world.
Firstly, there is interest now on the 'archives' front -- that is, those people/organiizations responsible for keeping the heritage recorded, and the DolbyA compatible decoder just might be used to help with that goal.
Next, I have several independent items of intelligence that INDEED DolbyA encoded material is leaking out to consumers, and NOT JUST leaking out. In some cases, it is due to a choice by the distributors to avoid the cost of the decoding operation, often from a digital tape. Also, since it is very clear that there is DolbyA material leaking out EVEN FROM THE LIKES OF HDtracks -- repository with proof below -- I seem to detect that there might even be some mastering done incompletely to avoid the effort.

I am not claiming the 'laziness or cheapness' aspect about the HDtracks offerings, however -- the quality no matter if it is DolbyA encoded, appears to be pretty good. No laziness there -- the quality is good, but it is DolbyA encoded. Here is my repo with two short examples from their 'Carpenters' offering... The 'UNDECODED" version is directly from the downloaded recording, and 'DECODED' version is after being processed by the DolbyA compatible decoder.

This is almost 100% proof that the material is DolbyA encoded!!! (even from a truly premium outfit like HDtracks) I am happy, because I got better quality than I thought that it was initially, because it can be decoded/cleaned-up!!!

Location: https://spaces.hightail.com/space/ru9jbJ5oCI
 

Tapetech

Well-Known Member
Apr 10, 2014
142
23
328
Fairfax, VA
Question: All of these digital files you have with (aledgedly) Dolby encoded music. Do all the files also have 200 nWb/m test tones?

If you have a digital file of a recording that is Dolby A encoded and that same digital file does not have a 200 nWb/m test tone, then you cannot properly decode that file (using either a hardware or software decoder). Dolby NR decoding (whether it be A, B, C, S or SR) is 100% linked to the flux level of the music. If you digitize an encoded Dolby recording without a flux level test tone, you loose the very important fluxivity reference and therefore can't properly decode the file.

Yes, you can guess at the proper playback level for Dolby decoding, but that is not accurate. Without the flux reference tone on your file, you are just guessing.
 
Last edited:

Tapetech

Well-Known Member
Apr 10, 2014
142
23
328
Fairfax, VA
I have several independent items of intelligence that INDEED DolbyA encoded material is leaking out to consumers, and NOT JUST leaking out. In some cases, it is due to a choice by the distributors to avoid the cost of the decoding operation, often from a digital tape.

So you are saying that " to avoid the cost" of Dolby decoding, music distributors like HDtracks are releasing albums without the proper Dolby A decoding?

That is very hard to believe. There is very little "cost" involved for decoding a tape (that is encoded with Dolby NR). All real mastering engineers have fully refurbished, fully functional Dolby decoders.
 

John Dyson

Member
Jul 2, 2018
41
1
13
So you are saying that " to avoid the cost" of Dolby decoding, music distributors like HDtracks are releasing albums without the proper Dolby A decoding?

That is very hard to believe. There is very little "cost" involved for decoding a tape (that is encoded with Dolby NR). All real mastering engineers have fully refurbished, fully functional Dolby decoders.

Please refer to my demo, and I have information from people who are 'in the know' and I trust very highly (they have no benefit, in fact the opposite) for misinforming me.
It really weirded me out about 1yr ago when I started seeing/hearing what I have been seeing/hearing. (I can provide NUMEROUS demos, including failed decodes to show that the DolbyA decoder often screws up if the material is not DolbyA encoded.)

So, it actually took me a few months to think about it, and try to find other atlernatives to my conclusion before finally accepting the FACT that DolbyA material IS leaking out to the consumer. My DolbyA decoder project is NOT targted to consumers, and already has strong interest in both the 'professional' community and in an offshoot of the normal commercial situation. (I don't want to hint strongly at the area where interest has been shown, just simply that it is a fact.)

I don't really have a strong interest (other than informing) to encourage/discourage people believing me, and I fully respect the notion that individuals might not believe me at first, and as a partial 'input' to those who might not immediately believe me -- I have made a good faith 'effort' to provide some evidence, which is at one of my repository sites.

* Listen first to the 'Hurting Each Other' undecoded version, then listen to the decoded version. That is a pretty clear example, showing the decrease of hiss, and a little better dynamics.
* For an example about how STEALTH the DolbyA encoding can be, listen to the 'Top Of The World' example, and note how it is plausible that it is not DolbyA encoded (but it really is -- the rest of the album is DolbyA encoded.)

The only thing that I can see that votes against DolbyA encoding is that appears that each music file on that album appears to be normalized, and on most DolbyA encoded albums, each song is usually recorded at a different peak level, and the DolbyA settings are the same from song to song.

Other than the fact that it does appear that there was SOME final mastering because of the normalization, usually DolbyA encoded albums are NOT normalized.

I am not sure if each song on the Carpenter's album example has a different DolbyA level (threshold/tone level) or not -- it is a real pain to figure out the DolbyA level for every song on an album. It can take me a few minutes for each song in the more difficult/subtle cases.

For the DolbyA tone, there is a 'threshold' setting that is related to the tone leval as optionally displayed by the program.

I NEVER plan to sell the decoder to consumers -- might happen, but definitely not in the plans. All of the interest is commercial, and when I have broached the subject with actual recording engineers, it has either been 'acceptance, because they knew about it', 'disbelief, thinking that I am a crazy person', 'disgust, because so many people might be cheated', and probably a lot of other responses.
I just don't want to be called a crazy person -- so, believe what you want. However, I am truly trying to inform.

(BTW, those people who worry about buying fancy op amps -- except in cases where they are relaly needed -- real DolbySR units have LF411 type opamps for almost everything, pretty poor quality when compared with the good TI units. What I am trying to say -- sometimes people are trying to solve a real problem, but also it is easy to spend too many resources in the wrong place -- never really fixing the problem. This DolbyA thing can be a fix for a lot of between 1966-1990 material, and is really easy to do now with the available software. (The really premium version of the software is NOT freely available, but there is some stuff out there which can really help.)

So -- I don't gain anything in any way if people believe or disbelieve, and I accept that it might take a while for it to 'sink in', becuase it took me months to accept the idea. However, I didn't have anyone supporting my opinion, so I had to do the research myself (and luckly, I am a very, very experienced techie with a very good ear.)

Demo location: https://spaces.hightail.com/space/ru9jbJ5oCI

John
 

Tapetech

Well-Known Member
Apr 10, 2014
142
23
328
Fairfax, VA
I am not sure if each song on the Carpenter's album example has a different DolbyA level (threshold/tone level) or not -- it is a real pain to figure out the DolbyA level for every song on an album. It can take me a few minutes for each song in the more difficult/subtle cases.

So you are setting the Dolby threshold level by ear (since there is no Dolby tone on the file). That is quite inaccurate and can result in mistracking. I hope you don't expect other people to (successfully) do that as well if they use your Dolby A decoding software. Dolby encoded files have to have enclosed Dolby tones in order to be correctly decoded.
 

John Dyson

Member
Jul 2, 2018
41
1
13
So you are setting the Dolby threshold level by ear (since there is no Dolby tone on the file). That is quite inaccurate and can result in mistracking. I hope you don't expect other people to (successfully) do that as well if they use your Dolby A decoding software. Dolby encoded files have to have enclosed Dolby tones in order to be correctly decoded.

I am very well aware of the 'setting threshold by ear' issue, and it is time consuming for me. The use of the decoder SHOULD NOT be day-to-day, but rather one-off to have a good copy of what you have purchased (if in the consumer world without tones, it can take a few minutes per album.)
With tones, the program allows you to measure them, and then enter the correct level -- so for pro use (the actual market) it is no trouble.

I can set it within about 0.1-0.2dB by ear on a lot of material. Some material really confounds me, and I cannot authoratively say what the setting should be
I have also found that even off by 1dB, the sound is better than the DolbyA compression itself (well, most of the time.) MISTRACKING IS OFTEN BETTER THAN TERRIBLE COMPRESSION.
There is one recording that I have -- I must say, if it is off by more than 0.1dB, it doesn't sound right. I would suspect that a real HW DolbyA would have a similar adjustment sensitivity for that material.


Pleae refer to the HDtracks example -- the threshold level was even totally weird (a bit challenging to set correctly.) Most commercial recordings seem to require a setting bimodal (relative to 0dB input level of -13.8dB +-1dB and another peak at about -15.4dB).
DolbyA recordings with 'tones' also seem to come in at about -13.5dB range when copied to digital with proper gain reference, etc.

Oddly, the distribution peaks for commercial recordings are about 1.5dB apart.

If you listen to my demo, that was set manually and by ear (listen for the s sounds -- the highs tend to get clogged up if not set correctly), also no jerking or gating effects. The threshold for that was about (forget the exact number) -8dB or so.

With the decoder, being so very easy to use & runing realtime, an iterative run is also easy. Not like an old DolbyA unit where you might have to rewind/play/rewind/play -- just issue the command again with a different level setting. I use command line editing, so most of the time spent is in the listening to some difficult parts of the music.

BTW -- the decoder in question also has MUCH less distortion (esp intermod) so material like vocal chorus come out amazingly clearly (not the big blob sound.) Also subtle and intense instruments (Herb Alpert or Sergio Mendes stuff) come out with really good detail. NO HIGH BOOST, but real detail.

John
 

John Dyson

Member
Jul 2, 2018
41
1
13
Sorry for the 2nd reply. I forgot about a repo with some recent examples, so I just uploaded them to a temporary site. These are demos, but only mp3, so lose significant amounts of clarity (many of the examples are prone to spatial shifting or mashing together vocals or quickly repeated echoed voices being mashed into a single vocal.) Mp3 is especially bad at the quickly (ms) repeated vocals. Real DolbyA hardware will also tend to smash things together into softened blobs. WIthout any HF boost the SW decoder is mostly better in those regards. I cannot claim ZERO intermod, but I can claim very little intermod. Here is the temporary repo: https://spaces.hightail.com/space/z3H68lAgmJ

BTW -- some of the material was EQed by the distributor, so I had to guess and reverse some of the EQ -- but the material originally DID have the dolbyA HF compression -- which is a major way of detecting it other than just excess HF in the sound.
 

John Dyson

Member
Jul 2, 2018
41
1
13
Update on the decoder -- there have been major/fantastic quality updates -- one of the targets is archivists, and the quality standards must be near perfection. As it now, the decoder can take such problematical material as many of the ABBA choruses and clarify the vocals tremendously. On other material, where there might be a loud/insdistingushable chorus when using an original DolbyA decoding, the new decoder does a much, much better job. Since there is a bifurcated market -- I am manitaining the 'better than original DolbyA' on my original distribution sites. The 'incredible' version is being upgraded to work with RIF64/BEXt type headers, longer files, etc. The sound quality has hit the point where it is very comparible with the original master tape in some ways. (I have some material which is altered, but comes from master tape instead of original DolbyA decoding, and the resulting 'compatible decoder' decoding sounds very much like the master tape rather than a Dolby decode.
The big technical differences include -- much more accurate tracking of the attack/decay curves. The original line of decoders used an approximation, and the new versions are really dead on (I had to sit down and do some serious math, instead of doing an approximation.) Now, the new one has the same decay as a real DolbyA full stop. The attack is different, but with a purpose. There is a flag to disable the special attack, but you really never want to use that. The intermod is as fully eradicated as the decoder can do --w hen considering that the original encoding does its own fair share of intermod, and the phase accuracy above about 100 Hz isn't good enough to accurately cancel the original intermod. Below the 100Hz range, the intermod cancellation is perfect -- the difference can be heard between the original decoder version and the current pro version.
The currnently available decoder is very useful if you don't like the DolbyA compression sound and have lots of 1960s-early 1990s material (some with the DolbyA encoding -- about 25% or more.) It is better than leaving the DolbyA compression, but is not perfect.
It probably isn't worthwhile to upload any decoded examples for the new version of the decoder, but anyone is interested -- I can do so. In some cases, flac is almost definitely required instead of mp3, so I could only upload a few good examples (the decoder is really that good -- in some cases it can produce results better than 200kmp3 can.
Also, if you have any material that you might be interested in decoding (thinking that it might be DolbyA encoded -- harsh high end, EQ doesn't work well, sometimes distorted below 80Hz), then contact me and I'll do a decode for you. However, I cannot distribute the decoder any more until we get the business case set up. This 'pull-in' has occurred because of surprising interest in the recording community (esp the newer versions), and I am even working on source obscuring for people who want to write a plugin. (There iare some super-proprietary methods for distortion removal/elimination, while maintaining enough sidebands to function correctly.)

So -- the DolbyA compatible decoder project is becoming too big for me --having to unplanned update the I/O interface to be more compatible for the pro community (believe me, I wouldn't do the updated interface code unless it was necessary!!!)
Thanks -- again -- I can offer decoding help if there is a frustrating problem for you -- and I keep everything private/proprietary when needed.

John
 

John Dyson

Member
Jul 2, 2018
41
1
13
Last minute note -- I have uploaded some examples (needed for something else.) ABBA hasn't sound so good since they were in the studio before the processing/DolbyA. They actually sound clean/clear -- not perfect, but also not that ugly sound. If you like ABBA, you might even want to keep some of the examples (the 256k mp3s do NOT do the quality justice, but still blow away everying ncluding the Polar and/or ABBA studio stuff or whatever.) The -demo files are the ones that have been fully processed so far. Just listen for the choruses -- the short MamaMia example clarifies what I am talking about.

https://spaces.hightail.com/space/xghqJodgrj
 

astrotoy

VIP/Donor
May 24, 2010
1,547
1,017
1,715
SF Bay Area
John, thanks for all your work. I have a pretty large collection of prerecorded 15ips 2 track tapes, many safety masters. Over the years I have run across quite a few Decca safety masters that were dolby A encoded. I stayed away from them since I didn't have a dolby A decoder. I do have an excellent dolby B decoder for the consumer tapes that were released in dolby B. Could I use your dolby A decoder to play back dolby A encoded tapes in real time. If so, I would start acquiring the dolby A encoded tapes. I could also use it to create non-dolby A tapes.

Thanks, Larry
 

John Dyson

Member
Jul 2, 2018
41
1
13
John, thanks for all your work. I have a pretty large collection of prerecorded 15ips 2 track tapes, many safety masters. Over the years I have run across quite a few Decca safety masters that were dolby A encoded. I stayed away from them since I didn't have a dolby A decoder. I do have an excellent dolby B decoder for the consumer tapes that were released in dolby B. Could I use your dolby A decoder to play back dolby A encoded tapes in real time. If so, I would start acquiring the dolby A encoded tapes. I could also use it to create non-dolby A tapes.

Thanks, Larry

NOTE: the repository was updated a few hours after the original post -- I grabbed the wrong files -- these WILL blow any ABBA listener away!!!

The DolbyA compatible decoder does run in realtime on X86_64 type machines of Haswell (4700 series) are newer. It 'works' on older machines -- but is slower. There is an SSE2 version mostly for ATOM machines, and AVX version for the newer type machines.
There are two versions of the decoder available -- one which is 'consumer' quality with a few problems -- but sounds better than no decoding. There is another one that does miracles -- decoding better than a real DolbyA. I have some ABBA examples that are almost impossibly good but come from close to the same source as everything else.
The 'free' one is avaialble at: https://spaces.hightail.com/space/tjUm4ywtDR with a filename of da-fixA-win-06jul2018.zip.
If it generally does what you want, but kind of misses the mark for absolutely the best, I have a 'perfect' or 'better than perfect' version that I can make available with a special build and your name encoded into it. The 'high quality' version works, but is limited to normal .wav files (doesn't handle the BEXT or RF64 stuff yet, but that is coming soon.) It works nicely with the normal, historical .wav files using 16bit, 24bit integers, or 32bit floating point, and can be forced to produce the floating point output if you have to use the 16 bit input. It also works with 44.1k - 192k, but really suggest using it between 48k-96k. The newer versions don't have the sample rate/quality tradeoff to the extent of the free consumer version. *NO MATTER WHAT, DOLBYA ENCODING DOES DISTORT AND IS WELL KNOWN BY PEOPLE WHO USE IT. The tradeoff is lower noise. But for intense, HF material it can be problematical (like ABBA, my basket-case test case), but for orchestral/symphonic type stuff even normal DolbyA decoding is okay. (A lot of that fuzzed-out ABBA sound is DolbyA, btw.) Carpenters sounds a LOT better with DolbyA than ABBA for example, but Carpenters doesn't naturally create as much intermod either.

The 'consumer' decoder will generally work/sound better than a real DolbyA, while the new one is better practically all of the time (haven't found a case where the real DolbyA is nearly as good.)
The caveats are: command line only, linux or win64 builds only, the DolbyA level is currently not calibrated on any version, but one can use the tone level as measured as a reference, then add about 1.5dB to that for a threshold setting on the command line. It can be tuned by ear, but that is a current disadvantage for every version. Usually tuning within 0.2dB is just fine, but sometimes it really needs 0.1dB or better -- but it is usually possible to narrow it down. I have some material that I had troubles tuning, and then found that it wasn't encoded (really just did have too much HF.)

The decoder is not just an approximation, but does deviate from a real DolbyA on purpose in important and useful ways. The anti-intermod is heroic, but intermod is the bane of any fast compressor/expander. Evey possible calculation to actually reduce intermod without being insanely costly is being employed (that is why the decoder only runs at realtime rather than 10-20X realtime.) A simple decoder that produces lots of intermod would easily run 10-20X faster, but there is a LOT of work in this new (both versions) decoder.

About the last minute note -- you can hear the improvement by comparing the file MamaMia-orig.mp3 vs. MamaMia-decoderc.mp3, where the 'orig' is from a real Polar disk, and 'decoderc' is using the decoder. If the source used for the decoder is used with a normal DolbyA -- then the quality will resemble very closely the sound of the 'orig' version. If you know ABBA sound quality, the -demo examples might knock your socks off - best that it has sounded since the studio, REALLY. The comparisons reside on a different site: https://spaces.hightail.com/space/xghqJodgrj .

So, if you find that the 'consumer' decoder comes close, but not quite -- then contact me, and I'll see what I can do to make you a temporary version of the full blown one -- will be okay to use for 6mos to a year - and by that time the commercial/archvies version will really be avalable, and we'll hopefully find someone who would like to write a plugin version also.

BTW -- this project really sucked me in. Initially, it was just a way FOR ME to get rid of the 'digital HF excess' so common on recordings from the late '60s through the early 1990s. I found that maybe 1/4 or so of the recordings were NOT just HF boosted, but were left DolbyA encoded. I wrote a super simple decoder (the level that one would normally purchase elsewhere and be disappointed), but it did help the sound quite a bit. A recording pro found out about my project -- got some real help from him with specific information, and now we have effectively a perfect decoder. I got more detailed curves (we developed) of the actual DolbyA units, and I invested a LOT more work in the anti-intermod side of things. There are anti-intermod techniques that I'll bet haven't been known or used by anything currently available until now.

Again -- give it a try if you want, and if insufficient, just get in touch with me..

John
 
Last edited:

astrotoy

VIP/Donor
May 24, 2010
1,547
1,017
1,715
SF Bay Area
Thanks, John. I'm not quite following this. So I need to ask a simple question to make it clear for me (not a real computer person). Do I have to digitize the tape or can I attach the dolby A and just play the tape through it, like a real dolby A decoder? Does the program digitize the signal from the tape recorder? or does it just process it in analogue and gives out the analogue output? My experience is that when I play a dolby B tape, I just plug it into the dolby B decoder and hook up the output of the dolby B into my preamp. I don't think it digitizes the signal as it processes it, but I could be wrong. My dolby B decoder is a modified version of the Dolby 330 Series (I think probably from the 1970's?).

Larry
 

John Dyson

Member
Jul 2, 2018
41
1
13
Thanks, John. I'm not quite following this. So I need to ask a simple question to make it clear for me (not a real computer person). Do I have to digitize the tape or can I attach the dolby A and just play the tape through it, like a real dolby A decoder? Does the program digitize the signal from the tape recorder? or does it just process it in analogue and gives out the analogue output? My experience is that when I play a dolby B tape, I just plug it into the dolby B decoder and hook up the output of the dolby B into my preamp. I don't think it digitizes the signal as it processes it, but I could be wrong. My dolby B decoder is a modified version of the Dolby 330 Series (I think probably from the 1970's?).

Larry

Straightforward answer about the decoder -- it takes a DolbyA encoded .wav file in, and produces a decoded .wav file. There is a very signficant amount of commercial and archival material that is DolbyA encoded, and believe it or not, some consumer material has leaked out into distribution which is DolbyA encoded.
So, it takes a .wav file in, then in the command line, one specifies the 'threshold' level which is roughly equivalent to the tone level. The program can measure the tone level, and then you can use that measurement to specify the command line parameter.
In the case of material which does NOT have tones (e.g. The Hollywood records version of Queen's greatest hits, singificant amount of Carpenters' stuff, some Olivia NJ material, some copies of the Brasil'66 stuff, Petula Clark, Simon & Garfunkel, etc.) I have found a way to detect SOME of the consumer DolbyA material, but it requires some skill, is approximate and DOES NOT require listening as much. If you look at the spectographs, you can notice that there is a decay after transients. Material which has been dolbyA encoded/decoded and encoded again (like some of the consumer leaks), the noise decay curves on the spectogram are extended from a normal noise decay curve on decoded DolbyA material. (I know this seems confusing -- but there is a way to detect SOME of the material.)

Much consumer material has already gone through an encode/decode cycle -- also material which hasn't, especially new material, will have a smooth noise behavior on the spectorgraph. It is difficult to detect 'encoded' materal per se, but relatively new recordings with hiss (stuff after middle 1960s) and doesn't have an obvious DolbyA decoding noise curve, is a VERY GOOD candidate for being DolbyA encoded (and esp if the high end is intense.) Even a 15dB boost at the highest frequencies doesn't sound all that horrid (bad, but not horrid), and an EQ to correct the high end vs. dolbyA decode -- both generally work... However, the DolbyA decode is superior if it applicable (removal of HF compression artifacts, etc.)
THe HF compression artifacts actually become a bit more noticeable to me if I simply do an EQ of DolbyA material, and that is another hint that it probably is encoded.
The advantage of doing the full decode instead of just EQ is 1) Much better stereo image depth, 2) better noise reduction with more accurate freq response corrrection, 3) removal of compression artifacts, and probably more.

So, the decoder can take the DolbyA encoded material and then produce a the best version of the original, pre-DolbyA encoded material other than the source master itself.

I have been working on improving my decoding of the ABBA stuff (and improving the achival quality decoder itself.) I like ABBA, but don't like listening to it as much as I have to :). Anyway, the current decoded versions are better than even yesterday, and provide an almost impossible recovery of the original master tape (I haven't heard the master tape without encoding, but it sounded INFINITELY clearer than any other verson avaiable even a month ago, yesterday it was amazing, and today it has me confounded it is so very good. I just came up with a new scheme a few hours ago, and have had no time to implement yet -- the next should be even better. Call me the intermod king :).

Just do me a favor -- listen to the CURRENT (as of about 1HR ago) version of my repository, and compare with any other copy of ABBA that you might have... Amazing -- let me know what you think: (there are only a few complete examples)

Repository location https://spaces.hightail.com/space/xghqJodgrj

John
 

John Dyson

Member
Jul 2, 2018
41
1
13
A super major breaktrhough on the DolbyA compatible decoder. This version has a full solutoin to the intermodulation problem -- plagues real DolbyA HW units even worse. The solution is pretty much complete, and resolves the intermod issue practically down to the theoretical limtis. If you want to hear the resutls -- there are two example files on my repository -- Mia-orig.mp3 and Mia-decodere.mp3. Mia-orig.mp3 comes from a Polar CD -- the ABBA standard source for the 'best.' The Mia-decodere.mp3 comes from a DolbyA near-master tape that I acquired a few decades ago, but has been decoded by the latest version of the DolbyA compatible decoder. When you listen to the two -- I'll bet you'll think that Mia-orig is more compressed -- but it is the other way around. There is a slight amount of additional compression in the decodere version. The 'compressed' sound of Mia-orig comes from the severe intermod caused by the female voices mixing with the gain control. The problem is solved in my decoder now. The results are NOT pristine, but as good as can be done nowadays. When you listen to the other examples on the repository, you are hearing as good as possible other than if you listened to the material BEFORE DolbyA encoding in the studio. I am NOT exaggerating about this -- the decoder is NOW that good. It will be coming out soon -- the resutls are frighteningly good.
Location for the demos: https://spaces.hightail.com/space/xghqJodgrj

John
 

About us

  • What’s Best Forum is THE forum for high end audio, product reviews, advice and sharing experiences on the best of everything else. This is THE place where audiophiles and audio companies discuss vintage, contemporary and new audio products, music servers, music streamers, computer audio, digital-to-analog converters, turntables, phono stages, cartridges, reel-to-reel tape machines, speakers, headphones and tube and solid-state amplification. Founded in 2010 What’s Best Forum invites intelligent and courteous people of all interests and backgrounds to describe and discuss the best of everything. From beginners to life-long hobbyists to industry professionals, we enjoy learning about new things and meeting new people, and participating in spirited debates.

Quick Navigation

User Menu

Steve Williams
Site Founder | Site Owner | Administrator
Ron Resnick
Site Co-Owner | Administrator
Julian (The Fixer)
Website Build | Marketing Managersing