Dolby Surround (2014) vs Dolby Pro Logic II

[MidiMagic]

Not true. The mono and 7Khz cut on the rears in DPL is done in the decoder. Neither of those apply in a DPL II decoder so the same content can play back differently in DPL II rears.

But does that change where the sound is located by the ears? To me the direction the sound seems to come from is more important. I want a system where the speakers disappear from the image. I do not like puddles of sound at the speakers.

 

[Owen Smith]

Results using disc 2 track 30 surround channel identification.

Caveat: this test track only expects mono rears so won't show anything about rear left/right separation.
Setup: this is a 5.0 system. Played from USB stick in Oppo 203 over HDMI to Arcam AVR31 amp.

Decoded in Arcam AVR31
===================
Dolby Surround: basically OK but obvious channel leakage, a little left front into left rear and a little right front into right rear, and a little centre into each of left and right and a little of each of left and right into centre.

Dolby Virtual: amp won't let me select this for 2 channel source material.

DTS Neural:X: very discrete, no leakage of any channel into any other. But between each channel the audio snaps back to playing from all speakers for a fraction of a second, which might be distracting and might not be depending on source material.

DTS Virtual:X: miserable failure. Left speaker plays left, centre and right with equal volume, as does right speaker. Centre and surround speakers only play their own channels.

Auro 2D Surround: miserable failure. Front 3 speakers work well in that Left is shared in left and centre, Centre appears in all three, Right is shared in right and centre. But Surround channel plays in all 5 speakers, and rear speakers quietly play all 4 channels. This mode might work well with stereo music with no matrix encoding and therefore nothing intended for the rears.

Auro Native: plays exactly the same as when I have Digital Stereo selected ie only plays from left and right speakers. I can't work out what this mode is supposed to do, maybe it only does anything with actual Auro encoded bitstreams but in that case why let me select it with 2 channel source material?

Decoded in Oppo 203
================
Player has settings for DTS Neo:6 in Cinema and Music modes. But the HDMI still outputs LPCM 2.0 so it looks like the player's Neo:6 processing can only be applied to the analogue outputs. I can't use those as the Arcam AVR31 does not have multi channel analogue inputs.

Conclusion
========
A worthwhile experiment.

On proper Dolby Stereo encoded material with the Arcam AVR31 decoding it is a straight choice between Dolby Surround if you want smooth with some channel bleed, and DTS Neural:X if you want a more discrete presentation. I will try both in future when I'm watching encoded material. I still feel DPL II Movie did a better job, being somewhere between the two though this is based on memory from a couple of months ago and using source material not test tracks.

On stereo music and non encoded material Auro 2D Surround doesn't mess up the stereo image and spreads it nicely between the front three speakers, while putting a little of all channels in the rears. Sadly DPL II Music did a considerably better job in my experience, spreading the stereo image into a horseshoe that included the rears which none of these codecs do.

I've read somewhere on QQ that Dolby Surround is supposed to have some setting that affects the centre width like DPL II Music had, though with less values you can set. EDIT: that appears to be a separate Dolby Audio Processing option, but setting it to any of Off, Music or Movie makes no difference to how the channel ID test is played. That could be a bug in the Arcam AVR31 since it has plenty.

I'd like test tracks that expect a DPL II decoder and therefore put different things in rear left and read right. In the absence of that there are some unanswered questions about these decoders.

 

[Owen Smith]

Listening to the Dolby Stereo encoded music tracks on disc 1 kind of seals the deal. DTS Neural:X produces the infamous effect of pools of sound around each speaker, whereas Dolby Surround does not. The small amount of channel leakage in Dolby Surround is enough to create a cohesive sound field. The fact that my centre speaker only has the same woofer but not tweeter as my left/right and hence is tonally slightly different doesn't help Neural:X, and my rears don't share either tweeter or woofer with any of the front three so differ tonally a bit more. They are all from the same manufacturer (Castle) and are the closest I could get at the time. I suspect I am not alone in having slight tonal differences between speakers. The fact remains that Dolby Surround masks this with a little channel bleed, whereas Neural:X tends to highlight it.

I have noticed on discrete sources that it is rare for them to be totally discrete and they often blend the channels a little, more so modern mixes than classic 1970s quad. But even so Neural:X takes it to extremes.

 

[Owen Smith]

Oh yes and DTS Neural:X appears to be boosting the bass, naughty thing! The low bass is louder in Neural:X, whereas plain stereo and Dolby Surround it's about the same. Many people may not notice, I have Castle Harlech S1 floorstanding left/right speakers. The speakers spec says 40Hz to 20KHz at -3db but the subwoofer sweep tests on disc 2 with someone saying the frequencies on this disc shows pretty much full response down to 30Hz and then a considerable roll off below 28Hz but still reproducing all the way down to 20Hz. Quite remarkable, though they are 1m tall floorstanders with dual quarter wave folded pipe bass loading. The Arcam AVR31 amp has class G (voltage rail switching) power amps, and bass response is better than with my previous amp which was class A/B. This is my 4th amp with these speakers, and it may be the first that really does them justice. It's certainly a much more expensive amp than my previous 3 and these 25 year old speakers have opened up very nicely with this amp.

 

[JohnN]

Tried almost all = DTS Neural:X is my flavour of the moment for the height speakers.

 

[Wunlow]

Replies to Owen

DTS Neural:X: very discrete, no leakage of any channel into any other. But between each channel the audio snaps back to playing from all speakers for a fraction of a second, which might be distracting and might not be depending on source material.

That's interesting as a Pioneer DPL-2 decoder I had tested before has this very same issue, in both the movie & music modes.

Oh yes and DTS Neural:X appears to be boosting the bass, naughty thing! The low bass is louder in Neural:X, whereas plain stereo and Dolby Surround it's about the same.

One thing with DTS Neural decoding ... or at least the iteration I had tested a few years back .. is that while it redirected low bass to the LFE channel it did not do any directional steering of low bass at all - meaning it simply left them in left and right front channels. I wonder if this was done to prevent logic steering issues, or an attempt to maintain bass consistency. In either case it did result in a bit of a bass boost in my system... and especially if my fronts are set to small. Bass management and 5.1 as most know can be a messy thing.

And speaking of lower frequences- I did send you the test tones via PM but forgot to mention as many of the tests are pink noise, depending on how each decoder handles the low end, you might end up with some bleed through / leakage in some metering tests that could be more of the low end not being fully decoded. Some versions of Neural do this and based on another detailed report of Dolby Surround (specifically the Dolby DSU) is it may do the same thing or at least compensate for it better than the current Neural-X version.

It's been interesting to hear your results so far! Thank you for sharing them.

 

[Owen Smith]

By the time I got your test tones I had listening fatigue from the first set from the CD. I'll get to it.

I have a 5.0 system, no room for a subwoofer and as I've been increasingly discovering my left/right pair have better than expected really low bass anyway particularly with my new amp (class G has more drive strength). So I'm not getting bass boost due to it being redirected to the sub and also still in the mains.

 

[JediJoker]

I have Castle Harlech S1 floorstanding left/right speakers. The speakers spec says 40Hz to 20KHz at -3db but the subwoofer sweep tests on disc 2 with someone saying the frequencies on this disc shows pretty much full response down to 30Hz and then a considerable roll off below 28Hz but still reproducing all the way down to 20Hz.

Likely down to the difference between an anechoic measurement and your in-room response.

 

[Owen Smith]

Likely down to the difference between an anechoic measurement and your in-room response.

Agreed that's very likely. There are definitely some volume peaks and troughs as the frequency descends, that subwoofer frequency sweep is slow enough to hear them where all others go too fast to spot them.

But my new amp has definitely strengthened the low bass, it was the first thing I noticed when I swapped to it. The class G amps are 100W into 8 ohms and 180W into 4 ohms so if speaker impedance drops at the lowest frequencies (it's never flat on any speaker) the new amp will drive it better. My previous Arcam AVR350 is class AB 100W into 8 ohms but no figure listed for 4 ohms which implies to me it's not much better, recent Arcam 100W class AB amps do 120W into 4 ohms.

 

[Owen Smith]

I tried the test tones Wunlow sent me by PM. Most of them repeat what I've already tried with the same results but crucially they included rear left and rear right encoded pink noise. Rear left and rear right give matching results as you'd expect except on opposite sides. All decoded with Arcam AVR31 on a 5.0 system with all speakers set to Large as before.

Dolby Surround: pink noise is heard in all 5 speakers but is considerably louder in the intended rear right or rear left speaker.

DTS Neural:X: total silence from the front 3 speakers. Minor bleed into the opposite rear speaker but less than with Dolby Surround. Strong output from the intended rear right or rear left speaker.

I'm not testing the other decode modes since they did such a poor job last time.

So the results are much as before, DTS Neural:X is a lot more discrete. But this does show that both decode DPL II stereo rears into the correct speaker.

I've been experimenting with the two modes when watching TV or films that only have 2.0 soundtracks:

When they're properly Dolby Stereo encoded with a mix that uses all speakers Neural:X sounds a lot better, Dolby Surround muddles the soundfield.

But when they're old films either genuinely in mono or just with very little in left/right, Neural: X sounds poor because everything is in the centre speaker and the soundfield collapses. Dolby Surround produces a more pleasing result here, producing a soundfield out of nothing. Even with mono content and my left/right speakers with a great phantom centre it manages to spread the sound across the front soundstage. Dolby Surround must be doing something with phase to achieve that.

With stuff that has plenty in left/right but not properly encoded, like a lot of domestically produced UK TV, it's hit and miss which one sounds better. Sometimes neither is better than the other, just different.

I've not reproduced the Neural:X pools of sound around each speaker except with the music on the Delos samples.

So it looks like we do have good replacements for DPL II Movie with these new decoders. What we're missing is a replacement for DPL II Music, none of them did a good job.

 

[Francis Drake]

Have been trying to make some sense out of this thread, but i think there may an incosistancy in the naming of standard dolby encoded material, Dolby Stereo and Dolby Surround?

As to a source of DPL II music, for some strange reason the music business didn't replace DS I with DPL II, i doubt too many would have noticed, and when equipment was updated it would have been more the worthwhile.

It does seem as though you'll need to get hold of the soundtrack from one of the DPL II encoded Playstation games, where it seems as though Sony were the only company to employ it. Lost oppertunity i think.

 

[ssully]

Replies to Owen

That's interesting as a Pioneer DPL-2 decoder I had tested before has this very same issue, in both the movie & music modes.

One thing with DTS Neural decoding ... or at least the iteration I had tested a few years back .. is that while it redirected low bass to the LFE channel it did not do any directional steering of low bass at all - meaning it simply left them in left and right front channels.

Not sure I understand this; do you mean, it copied low bass from front channels to the subwoofer, but also left low bass in the front channels (akin to the "LFE + Mains' option for bass management on some AVRs, effectively 'double bass')?

 

[Wunlow]

Not sure I understand this; do you mean, it copied low bass from front channels to the subwoofer, but also left low bass in the front channels (akin to the "LFE + Mains' option for bass management on some AVRs, effectively 'double bass')?

That is correct.

 

[Wunlow]

I tried the test tones Wunlow sent me by PM. Most of them repeat what I've already tried with the same results but crucially they included rear left and rear right encoded pink noise. Rear left and rear right give matching results as you'd expect except on opposite sides. All decoded with Arcam AVR31 on a 5.0 system with all speakers set to Large as before.

Dolby Surround: pink noise is heard in all 5 speakers but is considerably louder in the intended rear right or rear left speaker.

DTS Neural:X: total silence from the front 3 speakers. Minor bleed into the opposite rear speaker but less than with Dolby Surround. Strong output from the intended rear right or rear left speaker.

I'm not testing the other decode modes since they did such a poor job last time.

So the results are much as before, DTS Neural:X is a lot more discrete. But this does show that both decode DPL II stereo rears into the correct speaker.

I've been experimenting with the two modes when watching TV or films that only have 2.0 soundtracks:

When they're properly Dolby Stereo encoded with a mix that uses all speakers Neural:X sounds a lot better, Dolby Surround muddles the soundfield.

But when they're old films either genuinely in mono or just with very little in left/right, Neural: X sounds poor because everything is in the centre speaker and the soundfield collapses. Dolby Surround produces a more pleasing result here, producing a soundfield out of nothing. Even with mono content and my left/right speakers with a great phantom centre it manages to spread the sound across the front soundstage. Dolby Surround must be doing something with phase to achieve that.

With stuff that has plenty in left/right but not properly encoded, like a lot of domestically produced UK TV, it's hit and miss which one sounds better. Sometimes neither is better than the other, just different.

I've not reproduced the Neural:X pools of sound around each speaker except with the music on the Delos samples.

So it looks like we do have good replacements for DPL II Movie with these new decoders. What we're missing is a replacement for DPL II Music, none of them did a good job.

That was some great feedback Owen. Thank you! So we now have confirmation the Dolby Surround DSU does indeed decode DPL-1 correctly, and for the most part, also with DPL-2.

I'll add that with DPL-2 in the DPL-2 Movie or Music modes, those left rear and right rear decoding tests did fully decode. ...As in the left rear signal came out of just the left rear, and vise versa for the right rear.

Roger Dressler (who had worked at Dolby and worked with Jim Fosgate on the development of DPL-2) had also mentioned for DPL-2 Music the center channel had a 6ms delay added. While this slight mod is something I wish had been optional I could see reasonings behind that and how it could help.


It sounds like whenever I do upgrade to Atmos it will need to be a Marantz as they are one of the few manufacturers to still include a multichannel analog input. I could retain DPL-2 options with my current processor and also feed other multichannel sources / misc decoders in directly as well.

 

[Owen Smith]

That was some great feedback Owen. Thank you! So we now have confirmation the Dolby Surround DSU does indeed decode DPL-1 correctly, and for the most part, also with DPL-2.

With the caveat I have a 5.0 system which fits original DPL and DPL II quite well. If you have a 7.x system and/or some height speakers or main wides or all sorts of extras, Dolby Surround upmix will use them all as will DTS Neural:X. That is what they are meant to do, but I would argue at that point they aren't doing an approximation of DPL or DPL II decode any more as those old decoders wouldn't have done that unless you turned on 'x' or 'z' options.

And if I were to say one of them were decoding DPL and DPL II correctly I would say it is DTS Neural:X. Dolby Surround upmix smears things across multiple speakers and with genuine encoded content it doesn't sound great, though it's not awful either.

 

[Owen Smith]

I'll add that with DPL-2 in the DPL-2 Movie or Music modes, those left rear and right rear decoding tests did fully decode. ...As in the left rear signal came out of just the left rear, and vise versa for the right rear.

Sadly I never had those test tones when I had DPL and DPL II decoders in my old AV amp. Now where is that time machine?

Roger Dressler (who had worked at Dolby and worked with Jim Fosgate on the development of DPL-2) had also mentioned for DPL-2 Music the center channel had a 6ms delay added. While this slight mod is something I wish had been optional I could see reasonings behind that and how it could help.

I have a lot of time for Roger Dressler, he was much less bullshit and much more genuine advice and interest in making things better than most industry executives do. The interviews of him I've found on YouTube are great.

 

[Wunlow]

I had made those specific test tones with some input from Roger Dressler, although now I could make them a few different ways. I agree - He really is a great guy and the pairing of him with Fosgate could not have been better. (IMHO)

 

[Wunlow]

Have been trying to make some sense out of this thread, but i think there may an incosistancy in the naming of standard dolby encoded material, Dolby Stereo and Dolby Surround?

As to a source of DPL II music, for some strange reason the music business didn't replace DS I with DPL II, i doubt too many would have noticed, and when equipment was updated it would have been more the worthwhile.

It does seem as though you'll need to get hold of the soundtrack from one of the DPL II encoded Playstation games, where it seems as though Sony were the only company to employ it. Lost oppertunity i think.

Where you might be having some confusion with is what most of us originally had when Dolby introduced its latest matrix decoder.

Instead of naming their latest matrix decoding "Dolby ProLogic III, DPL-3", or "Pro-Logic Atmos", or something like that, they reverted back to an older term "Dolby Surround" which as most of know was the term originally used by Dolby (for decades) that essentially meant a bare bones non-logic decoder with crap separation. I get that they wanted to go generic but after decades of using "Dolby Surround" as referring to a process worse than the first Pro-Logic decoding, ...to me it was ripe for confusion. So I try to use the term "Dolby Surround DSU" when referring to their latest decoding. (Dolby Surround Upmixer DSU is part Dolby's new term when describing their new Dolby Surround matrix decoding)

And I agree it was a missed opportunity to not have DPL-2 encoded music but I think with discrete music platforms becoming possible at the time they didn't see the need for it. It's a shame though as DPL-2 encode/decode has wonderful separation results.

 

[kfbkfb]

Are there various versions of DTS Neural:X (depending on when the A/V receiver was built or when the software was last updated)?

Any particular adjustment setting for the DSU decoder/fake surround creator that results in good original Dolby Surround decoding?


Kirk Bayne

 

[Francis Drake]

Where you might be having some confusion with is what most of us originally had when Dolby introduced its latest matrix decoder.

Instead of naming their latest matrix decoding "Dolby ProLogic III, DPL-3", or "Pro-Logic Atmos", or something like that, they reverted back to an older term "Dolby Surround" which as most of know was the term originally used by Dolby (for decades) that essentially meant a bare bones non-logic decoder with crap separation. I get that they wanted to go generic but after decades of using "Dolby Surround" as referring to a process worse than the first Pro-Logic decoding, ...to me it was ripe for confusion. So I try to use the term "Dolby Surround DSU" when referring to their latest decoding. (Dolby Surround Upmixer DSU is part Dolby's new term when describing their new Dolby Surround matrix decoding)

And I agree it was a missed opportunity to not have DPL-2 encoded music but I think with discrete music platforms becoming possible at the time they didn't see the need for it. It's a shame though as DPL-2 encode/decode has wonderful separation results.

Thanks, helps to clear things up.