I have sometimes found that the ORIIGNAL CD releases are properly handled/mastered. But material like ‘greatest hits’, and even re-issues would be missing DolbyA decoding (giving the harsher, undecoded, ‘digital’ sound.) Later on, in the middle 1990s, it seems like a h*ll broke loose, and they brought out the surplus FM broadcast audio processors, and cranked them up to ‘11’. There is truly a 6-8dB loudness difference between a properly mastered (or decoded) ABBA disk and the recordings on ‘The Complete Studio Recordings’.
“All hell broke loose” 
Wow. There is a lot of wrong information here. First off, most of the best CD masterings are indeed from the mid 80”s to mid 90’s. Very often it is from the first CD releases and often from Japan. The Dolby problem is not a wide spread issue. In fact, it is a rare issue. Pre-emphasis is a bigger issue than Dolby.
Spend some time on the Steve Hoffman forum searching for opinions on the best masterings of any CD and you will see what I mean.
Most of my CDs ARE DolbyA encoded. Please refer to the Carpenters example… ABBA, Queen, etc.
It isn’t just EQ… Most people cannot even do a decode to verify (even thought that isn’t 100% accurate either.) The spectrum can also give a hint, and finding the HF compression (which is insidiously difficult to detect on much material.)
I can very easily decode DolbyA material (sometimes even doing a fair job on non-DolbyA – which is another source of confusion.)
I have a Nena CD that has both EQ (the old CD preemphais) & DolbyA.
John
Surprized no one has mentioned the Kate Bush CD2 Remaster boxset?
I bought this at Christmas, without hearing it, assuming it would be an improvement on the originals.
I was especially after Aerial and the 12 inch single remixes, i.e. running up that hill and the later remixed 2012 version for the London Olympics.
The original albums tended to be quite ‘trebley’ with little bass, guess what they did - took all the treble off and made it very bassy. sucked the life out of the recordings, made the albums dull and un-interesting.
Had to sell the box set, couldn’t listen to it.
Examples of good remastering - the Cure, latest one is Mixed Up, sounds great and Disintegration, also great. The Blue Nile remasters and the Beatles remixes are great too.
About the Kate Bush thing – I cannot claim that it was leaked DolbyA without listening/testing, but if the symptoms included a flattened spatial (stereo) sense, where the stereo depth is suppressed, then in all liklihood you had a leaked DolbyA copy. It is INSIDEIOUSLY difficult to 100% determine if something is leaked or not. I have a bunch of ABBA – where I can decode my leaked DolbyA copies to produce almost the same as the original Polar releases. I can even detect where one of the vinyl releases of ‘RingRing’ had screwed up EQ or something like that…
I have to admit that the ‘leak’ problem isn’t 100%, and might not even be 50% of the 1960’s through early 1990’s RE-ISSUES, but it did happen, and not just time-to-time.
Nowadays, the loudness wars problem makes the leaked DolbyA and pre-emphasis problems seem like non-problems.
I added a Nena example, original CD (noDEEMPH), CD plus deemphasis (yesDEEMPH), and the decoded version. The yesDEEMPH vs. the decoded version, it should be possible to hear the improvement in spatial depth (at least a little bit.) I didn’t do the exact precise calibration, so it might be off a little bit.
Demo:
1 Like
Steve Hoffman and Doug Sclar on the stevehoffman.tv don’t agree with your opinion…which is all it is. Look, CDs mastered with undecoded Dolby A master tapes would sound horrible without decoding. Yes, there are some CDs out there mastered with Dolby A undecoded. But they are few and far between. You are the only one that thinks it is as prevalent as you claim…
I do know a recording engineer also – and he has recognized that it has happened. I have lots of material as existence proof. The existence of the material has to trump opinion. I know EXACTLY what DolbyA material sounds like – there are some examples at the site that I pointed to in previous messages. Got lots more also…
In some cases, I did the research and found legacy copies when someone complained about lousy sound, and other times, I happened into the copies/CDs/digital downloads.
I’m wondering if this is also at play on some Led Zep remasters I bought recently. They almost made my ears bleed. If only I still had the original CDs which someone seems to have permanently “borrowed” from me, they weren’t perfect but they sounded a lot better.
Any time you think that it (the DolbyA leak) might have happened, make arrangements with me to send 60seconds (has to be 60 GOOD seconds) of the material, and about 90% of the time, I can determine if it is DolbyA.
It is kind of insidious – esp on POP material, where there is already enough compression from the studio that it is difficult to distinguish the DolbyA compression happening between about -20dB and -40dB – and it is really fast also. That kind of compression (DolbyA encoding) hides really well in some material. SOMETIMES, it just sounds like ‘more hiss’ because there might not be much material between 3kHz->20+kHz, and then other times one can hear the compression effects. Also, pure encoded material which has been encoded from another tape has some distinguishing characteristics on the spectogram. After considering all of the various data items, sometimes needing to listen also, I can be FAIRLY accurate whether or not the material is DolbyA.
Then, finally I might do an actual decoding operation – listening for a small amount of gating or a kind of over enhancement of the high frequencies (a kind of overly harsh HF) – and if the gating or harshness happens, then it is not likely DolbyA encoded.
The thing about leaked DolbyA – sometimes one cannot describe the problem – sounds harsh, sounds compressed, sounds shrill. A lot of the defects are hidden (not removed) by doing a bit of EQ instead of decoding.
I am NOT being absolute about any one recording being DolbyA encoded unless I analyze it.
I AM being absolute that there has been a significant amount of DolbyA encoded material leaked – often hidden behind -3dB or -6dB at 3kHz/Q=0.707. Some of the ABBA Gold releases (1992) didn’t even have the EQ – so many of those disks are pretty much pure DolbyA.
ADD ON: One more thing – a lot of recent remasters are a victim of the ‘loudness wars’ which is a different problem than leaked DolbyA all together. I have some material that is literally 6dB-8dB louder at the same peak level as the original – thanks to broadcast style compression… The loudness wars are UGLY.
John
Sure, It has happened. But, no one on the Hoffman forum in the recording business agrees with you in regards to how commonly you say it happened.
My engineer friend didn’t believe it either – until I proved it. Not all of the material is passed through the hands of recording engineers per se. From what I can tell – the recording engineer community does not like the fact, but the existence of the material IS the proof.
I did make an error on the ‘Day Before You Came’ demo – it is now correct. (I left the CD deemphasis for the Nena demo for the Day Before You Came’ demos. Now corrected.) Got 100’s more. Listen for yourself, and I didn’t have to hunt for my examples.
It isn’t a matter of ‘my daddy is bigger than yours’, but it is a matter of existence proof.
If you really need proof, I can make available perhaps 300-500 30second snippets of DolbyA encoded recordings… If you don’t care – then that is okay also. I have been very successfully using the material for testing, and it works very well. Non-DolbyA material isn’t as useful to me, and I found the material just collecting over the years.
John
1 Like
Fully agree! So what about ‘Brand X’ ?
Genesis was great until Peter Gabriel left.
2 Likes
Because the Japanese like that sound? Or because their perfectionism dictated that they make the copy as close to the original (= the master)? Who knows? Nevertheless, japanese vinyl was the holy grail in Germany back in the days when I was plagued with that format…
Interesting comment about the ‘perfectionism’. A master tape, properly handled/processed, shouldn’t sound a lot different than the CD (again, IF properly handled.) Sometimes, the mastering applies too much ‘love’ on the recording, but IMO the best mastering is the least reasonable mastering. The vinyl should also sound very similar to the master (maybe a bit less LF and/or some attention paid to HF tracking.)
If the material sounds a lot brighter or different from other versions – it might be purposeful.
Most of my DolbyA copies of ABBA come from Japanese releases, and a few even have zero (none, nada) EQ. Some of the Japanese versions DO sound brighter than the Polar versions – not just a little difference.
I have some material that is 1) a representation of the DolbyA ‘master’, 2) a representation of CD made from same original source, 3) a representation of vinyl made from same original source.
(Of course, neither the CD nor the vinyl were made directly from the same source, I am just claiming that all three come from the same source.)
All three sound ‘similar’ (the DolbyA version AFTER decoding), except one case that I have sounds VERY different, might actually have been a mistake in mastering the vinyl? (A vinyl copy of ABBA Ring Ring – really, far, far off.)
1 Like
BTW – with all of the talk of ‘bright sounding’ Japanese releases. I just brought out (onto the demo site) a recording that I purchased from HDtracks – the Carpenters ‘Singles’ album, and it pretty clearly IS DolbyA encoded (or maybe SR? – DolbyA is supposed to be able to partially decode SR, if a full SR decode isnt’ available – at least what I have heard.) I am not claiming DolbyA specifically because of the timeframe of the album creation.
Anyway, the material is DEFINITELY encoded with something that is somewhat DolbyA compatible. I just did a decode of about 45seconds of a song on the album (Yesterday Once More), providing both the encoded and decoded versions. The difference in stereo depth should be obvious, and the frequency balance should be much better in the decoded version.
To me, the undecoded version is ‘thin and picky’, while the decoded version is full and natural (well, considering the processing that they did on poor Karen’s voice.)
- Part of the reason why I am giving this example is to inform about what DolbyA encoding ACTUALLY sounds like. No need to listen to conjecture or other potentially filtered information – this is for actually listening for oneself*
The filenames start with ‘Carp-OnlyYesterday-Singles-fromHDtracks’, then decoded or undecoded.
I also changed the name of the Dropbox to be more friendly to all of the forums that I am chatting on.
3 Likes
I hate to reply to myself – but I thought for a real trip, I’d show another comparison to you.
A recording engineer and me (I am an EE/analog, DSP, and operating systems, C/C++ developer), we have been developing a decoder that is really good at decoding DolbyA material. This is the reason why I can do decoding so easily – no old-fashioned HW needed. Up until now, there hasn’t been an accurate SW DolbyA decoder, and it has taken me about 2-3yrs to complete this one. It is one in a series of NR decoding projects, DolbyA decoding was the first (and likely most difficult other than DolbySR.)
There is a significant difference between my decoder (which really IS accurate), and the actual, bona-fide DolbyA HW (we have done comparisons). The DA decoder (DHNRDS) is MORE accurate, and produces cleaner results because of the IMD suppression. The true DolbyA HW didn’t have IMD suppression (it would have been difficult in HW – refer to Orban Patent US 6205225 for a SIMPLE version of what the DHNRDS does.) (Actually, the algorithm is different, but some of the principles are the same.) Also, the feedback design that supposedly decodes the encoded material in the actual HW version is slightly flawed (otherwise TOTALLY GENIUS design by R Dolby.) There is a natural delay in the feedback loop, which makes the decoding incomplete. Also, the gain control feedback is necessarily fast – because it is ALSO a feedback loop. If delays (for distortion reduction) are added to that loop, very likely it will become unstable. The DHNRDS design, by virtue of incredibly tedious unfolding of the feedback design produces the same result as a DolbyA HW (within reason), but I have also added IMD remediation. In the lowest quality (fastest) mode, the DHNRDS sounds much more similar to an actual DolbyA decode on the difficult material. The DHNRDS is still a little cleaner because of the lack of delay in the loop.
On the same demo site, I uploaded a copy of Polar music’s original release of ABBA Take A Chance On Me, and the DHNRDS version. ABBA is almost a worst case test, not because the music is bad, but the mixed upper midrange vocals (lots of 3k-10k density) create lots of IMD, which like analog tape IMD, sounds ‘soft’ and smooths almost like compression (but it is NONLINEAR compression.)
If you compare the two – you can hear the individual voices more clearly in the DHNRDS version – not because it is faster in some ways than the Dolby HW (actually it IS faster in response times, but the attack/release trajectory is slower for the DHNRDS – the time being the same.) Also, the distortion sidebands are cancelled out when the waveform is high amplitude. So, the DHNRDS has two (actually three) methods to remediate the ‘fuzz’ relative to the HW decoder.
Just thought that it might be interesting – a lot of material has that slight muffled sound – ABBA often shows the most extreme partially because of legacy decoding technology and also the characteristics of their sound. Other material is less obvious – like symphonic music and a lot of less dense pop music. There is much less difference on less complex (less dense) music.
I have some jazz master tapes (real ones, with tones) for testing purposes, and even in that case, the difference is existant (cymbals are much cleaner/transparent), but the general sound is more similar.
The DHNRDS matches the gain curves, the attack/release times (very dynamic attack/release – constant RC time constants need not apply), but purposefully produces less distortion in the high quality modes.
Just thought that it might be fun – showing that the music on the master tape CAN be even more clean, more clear and more accurate than decoding properly with a DolbyA. Currently, I am working on Telcom C4 for some European archives. We have interest in the DA decoder from very big name archives – being able to recover old recordings more cleanly AND more convieniently (finally an accurate decoder in software.) It is NOT a Dolby decoder per-se however, it just happens to be able to decode DolbyA material very well. 
(We have already talked to Dolby about it – and don’t have clearance to advertize using their name, but they don’t plan to compete or develop any more DolbyA HW.) I am not advertising here, and only developed the software and like to talk technical. I do not do marketing, just like I do not do GUI software.
Again, thought that it might be interesting… I like to play with complicated toys :-). The decoder was NOT simple – the effort was well over 2000 hours of a very highly paid software/EE/DSP/OS developers’ time. No way to EVER pay back the time investment – but the purpose was to make old music more available and more accurate.
3 Likes
That is what I think, should be the case. And what I would try do do, if I was the mastering engineer. By the way, I would never transfer a DolbyA-encoded tape WITHOUT the proper decoding to a digital medium, it is beyond me, how that can be a standard practice (unless I misunderstood, what you wrote earlier)…
IMHO, there have been times, when “bright” was en vogue (maybe excluding the audiophile community), I think, when the CD was introduced in the early 80s, that was probably the case. So, everything got remastered to match that taste (I remember the Zappa ‘Old Masters’ boxes on vinyl, I found them extremely “bright” and didn’t like them at all… I had some of the originals for comparison, but not the precious early ones, that were re-released in the first box)… Soon later the opposite was en vogue: a warm, round, smooth sound… (you know what I mean…)… and everything got remastered again, according to the new ideal. And so on… the industry can just make endless money out of the same material by adapting to new trends.
Regarding the archiving procedures being direct from analog to digital tape. I promise that I’ll hunt for the document on the procedure for backing up analog tape onto digital formats. It is somewhere in my archives, just very disorganized.
However, here is a logic game that explains what happened…
…
When you DolbyA encode material, you DO NOT want to encode/decode/encode/decode, even though multiple cycles aren’t that damaging, it still causes damage.
When backing up material, generally one doesn’t decode before saving off onto another tape machine, usually one depends on the DolbyA tones for calibration, and just copy the raw material.
Since just copying the raw material has been the procedure, the raw analog was copied to the digital media – just the same as if it was analog.
Most often, the final DolbyA decode (after/during final mastering) would be done late in the vinyl preparation process.
…
Later on, maybe decades later, there is a bunch of DolbyA encoded material, but it is a royal pain in the b*utt to do a DolbyA decode in the digital domain. Answer: do a -3 or -6dB at 3kHz – sure, the public doesn’t care… The public is ‘stupid’, right? They’ll accept the inferior compressed material, I mean, who hears compression between -20 to -40dB anyway, right? We all know that just a bit of EQ will hide our(their) sins, right?
…
About 3yrs ago, I started a software DolbyA decoder project. I kept getting feedback similar to ‘cant be done’, or ‘impossible’… Don’t do that to me, because I will try to fight the impossibility. Crazy me.
Come about 1yr after start, I got connected to a recording engineer (more of a restoration engineer), and he started working with me/helping with details that I could not discern as a software/DSP/analog hw person.
It has taken 2 more years (3yrs total) to produce a decoder, that does a very good job on DolbyA material, sounding identical on ‘mellow’ material, and vastly superior on ‘difficult’ material.
…
Up until about 2-3mos ago, there wasn’t a really good decoder for DolbyA that runs entirely on a computer. When I designed/wrote the decoder, I added all kinds of embellishments that basically get rid of the fuzzing that a true DolbyA creates. You can hear the ‘fuzzing’ very distinctly on the ‘SOS’ example, where the Polar music version has the ‘fuzz’, and the DHNRDS version doesn’t have very much fuzz.
…
The basic story seems to be – that it has been expensive for the distributors to do the DolbyA conversion for DIGITAL material, but vinyl has always had the DolbyA decoding process in place. Even when using a digital archive to produce a vinyl copy, the final DolbyA decoding is still operative.
The above is as much of the story as I know… I will look for the archiving procedure, and it is (to me) outrageous – but maybe that is in hindsight. At least, the DolbyA issue has given me a problem to solve, right? :-).
John
2 Likes