DTS 5.1 vs dolby digital 5.1 vs 24/192 master in pro tools (1 Viewer)

[Marti D.]

Greg K.
When mixing for Theatrical release the speakers are calibrated at 85 SPL in the front and 82 in the surrounds.
When mixing for home it's anywhere between78 to 82 all the way around.
See my link on "Room Calibration for Film and TV" here:
http://duc.digidesign.com/showflat.p...fpart=1#529464
Some mixers like to mix lower (78-80) to make sure all the words are heard better at home. Not all Movies are re-mixed for*DVD, only some. Of those who are re-mixed only some of those are actually remixed by the original Re-Recording Mixer. Part of the problem at home is that most of the speakers used by Home Theater enthusiasts are brighter than hell, measuring flat and sounding pretty bad!

 

[DaViD Boulet]

Amen.

I remember seeing Jurrasic park in the theater. My local cinema had flash-strobe lights installed to replicate flashes of "lightening" and they even had some fancy updates to the sound system so they could properly do the bass.

I've heard Jurrasic park on DTS laser and DTS DVD (both DVD DTS versions).

Guess which one of those three sounds most like the theatrical experience I remember...

Yep. The laserdisc.

 

[GregK]

Thanks for sharing the various calibration links Marti. Good stuff! My trusty 'Ratt Shack' analog SPL meter has served me well for a number of years, and can't imagine trying to set up a system without it.

 

[PeterTHX]

Actually that was a gimmick for the trailer of "The Lost World". Search on DTS' press releases them talking about using their timecode sync to trigger the flashes. It was used (AFAIK) ONLY for that one trailer in certain venues. The films themselves did not have this.

 

[PeterTHX]

Well isn't this where DD's dialog normalization comes into play?

 

[DaViD Boulet]

Peter,

thanks for clearing up about the strobe lights.

Regarding Dialog Normalization:

After hearing how much *better* the DTS was than the Dolby Digital was on Chicago and Pirates of the Caribbean I contacted the mastering engineer from Disney. He told me that on both DVDs the same mix was used for both the DTS and DD soundtracks...with no additional EQ applied. When I asked if there was any reason he could think of why the DTS sounded so much better (besides the compression codec) he said "oh, we applied dialogue normalization to the Dolby Digital".

Curiously, the Dolby Digital soundtracks that sound the most like "DTS" (Lion King DEHT mix) do NOT use dialogue normalization. And neither do the AC-3 mixes on laserdisc that *also* sound better than their DVD counterparts. I'll bet you that's the biggest reason why the AC-3 Phantom Menace sounds so much better on laser.

 

[Michael Reuben]

David, it's not a question of "using" or "not using" dialnorm, since dialnorm is part of DD. It's a question of the dialnorm setting. The encoder default is -27DBFs, which results in an overall volume reduction of -4db. That's the setting applied to the Chicago DVD (I just checked).

I never had The Phantom Menace LD, but every DD track I've ever checked on LD used the same encoder default. Not surprising, when you consider that most DD tracks on LD were created in the early days of consumer DD products, when a lot of engineers were still learning the format.

I too wondered how dialnorm might affect the differences people hear between LD vs. DD. But since the LD tracks I've checked all have a -4db cut, I suspect the "slam" that people hear on the LD versions has to be explained by other factors --probably the lack of remixing for near-field environments. For example, on the first Mission Impossible LD, if I leave the volume at what appears to be an appropriate level for most scenes, the final sequence on the train blows me out of my seat. That doesn't happen with the DVD -- and it didn't happen in the theater either.

M.

 

[GregK]

David,

Bjoern's report indicate the Phantom Menace LD and DVD have the SAME Dial Norm setting. When the levels are matched, his graphs show the dynamic range is the same.

Dial Norm does not change the original dynamic range. Never has. Never will. When you talked to the mastering engineer at Disney his reply was correct, but you appear to have misunderstood his answer.

What would make DTS sound so much better than Dolby Digital? One easy answer- Dial Norm. If you do not level match the two, the louder mix will sound like the better one.

 

[DaViD Boulet]

Yes, but even when you adjust the volume to compensate, the DTS still *sounds* more dynamic. Perhaps the adjusted center dialogue has an affect on the way one perceives the dyanmics of the overall mix? Science ought to not tell us that we're not perceiving what we are...but rather try to help explain why we *do* perceive what we do.

Would anyone who understands the details care to offer a laymen's explanation of exactly how dialog normalization works? does it decrease or increase the level of the center mix relative to the other channels? Or only key-freencies that are deemed "voice range"? Or what? Why is it even *there*...shouldn't a mix be...well..."mixed" just the way it's intended to sound without additional on-the-fly processing from a post-production decoding chip?

Regarding Phantom Menace on LD and DVD...there must be more than just simple dialog normalization going on there. Every listener with the laserdisc I've spoken to was smart enough to adjust the volume level of the two to compare and *still* thought that the laserdisc sounded better with much more "slam".

 

[Michael Reuben]

It's a lot simpler than most people make it out to be.

Dialnorm raises or lowers the overall volume by the amount (if any) dictated by the dialnorm setting. It does not differentiate among channels. It does not differentiate among frequencies. It does not anaylze, alter, filter or care about the mix.

Why is it there? Mostly for broadcast use where channels are being switched. Properly and consistently used, dialnorm would allow every source to be adjusted so that dialogue is always at roughly the same volume. Thus, switching channels shouldn't lead to wide variations in volume.

Is dialnorm actually used that way? No. I don't think anyone pays much attention to it, except in discussions like these.

I have no idea who chose the encoder default or why, but unless the default setting is changed, a DD track will be encoded with an instruction to the decoder to lower the total volume on all channels and for all frequencies by -4db. I have seen other settings, but they're the exception.

That's really all it is. That's really all it does.

M.

 

[DaViD Boulet]

Michael,

thanks for your easy-to-understand explanation of this feature. That's the most clearly I've ever heard it put!

Question...I can see how such a feature has some use when we're talking about audio streams from many different sources that need to be "averaged" together like mixing a movie and commerial audio on the same channel feed...

But when we're talking about a stand-alone movie disc...what's the point of simply "reducing" the volume? What's the reference point that's trying to be matched?




Also...and this is a biggie...HOW is the volume attenuation being accomplished? Is the audio signal being digitally recalculated prior to d/a conversion? If that's the case, then that means that the audio signal is being *degraded* prior to d/a conversion by additional processing to reduce the amplitude of the waveform. This would compromise the quality by both reducing the waveform's resolution *and* by artifacts introduced from averaging errors in the algorithm. Increasing volume on your pre-amp after the fact to restore to the original level would still result in degraded sound quality because of the digital decimation.

If the volume reduction is being accomplished by an analog process after d/a conversion is complete...then this would not be the case and increasing volume on your pre-amp ought to "sound the same".

I know when HDCD developed their algorithm they required that non-HDCD discs be attenuated by 6db so that they'd have "the same perceived loudness" as HDCD discs which used a wider 20-bit dynamic range.

Guess what...cheap HDCD DACs accomplished this by simply TRUNCATING the LSB of the CD data prior to D/A conversion!!! The result was that the HDCD filter *degraded* the performance of regular CD by their digital method of reducing playback volume.

Higher-end HDCD hardware applied the attenuation in the analog domain...after d/a conversion was complete. These DACs did *not* degrade the quality of standard CD playback. I actually modified the chip on my HDCD dac to bypass the LSB truncation for non-HDCD discs (which made them sound 6db louder...but I didn't care...I just adjusted my volume knob).

Things to explore!!!

 

[Michael Reuben]

Also a good question, and I don't know the answer. I suspect that it varies depending on the decoder design.

M.

 

[DaViD Boulet]

Michael,

we've got to find out the protocol with how this level-adjustment is being accomplished. If my suspicians are correct...and this is all digital-domain decimation/recalculation to keep costs low on decoder/DAC design (I'd assume since most DACs used in Processors aren't custom-tailored for any given decoding algorithm), then that would mean that virtually every Dolby Digital soundtrack in produciton on DVD is being routinely compromised by unecessary resolution reduction and digital-processing error. That's really a shame. And certainly when I *have* heard Dolby Digital soundtracks that did NOT use Dialog Normalization (Lion King) my perception was "Wow! That sounds as good as DTS!!".

If it turns out that DTS sounds "better" to many listeners' ears not because of any great feat of compression-codec magic...but rather by virtue that the decompressed LPCM bitstream is merely being provided directly to the DACs whilst the Dolby Digital counterpart is being forced to suffer additional resolution decimation and processing for no real reason...this is a BIG DEAL and every audiophile on the planet should be PISSED.



Keep in mind that all that glorious "lossless" high-res auido on HD media won't be worth a darn if Dolby Digital "features" like dialog normaliztion are then put in place to crap it up prior to d/a conversion!!!!



Someone needs to come out with a decoder that will allow a user to defeat the dialog-normalization instructions and render a "clean" playback of un-attenuated post-decompression Dolby Digital data.

Sounds like it would be easy to do. Has anyone ever heard of a processor that can do this?

I know I've spoken with many Electrical engineers in the industry (Proceed, Camelot, Onkyo, B&K) who have stated that "Dolby Digital decoder design mandates the use of filters that degrade the performance of the audio". I wonder if this is what they are talking about???

 

[Michael Reuben]

Can you ask them for specifics?

M.

 

[Mark Lucas]

I think the damage is already done in the encoding stage. It's like those awful dynamically compressed cds where the data is just simply lost.

I never understand the purpose of dialog normalization. DTS seems to get by quite fine without it. It also doesn't exist on audio cds, dvd-a, and sacds.

 

[Michael Reuben]

Dialnorm has nothing to do with compression or anything analogous.

M.

 

[DaViD Boulet]

Michael,

when I worked in high-end AV retail in 1997-1999 DVD was just coming to market and most high-end audiophile brands had yet to introduce multi-channel AV deocoders that incorporated DD or DTS. Some of our high-end brands sent a few of their technical staff to the store to discuss new products and HT issues and/or at times I was able to contact some of them via the phone and ask them questions about upcoming products (ie, not talking to the regular "sales reps").

I questioned some of the high-end audio guys (like Levinson etc.) about when they'd come out with some high-end audiophile multi-channel gear (amazing, just 8 years ago there was actually virtually *no* high-end surround gear!) and they told me that they were all "working on it" but many expressed mis-givings about considering anything that incorporated "Dolby Digital" to be "high end" at all. When I pressed them, they said it wasn't because of the compression codec, which they thought sounded pretty good. They said it was because of *compromises* that were applied in the decoding and d/a process that "mandated filtering which degraded the sound". At the time that's all they said and I didn't know any more about the architecture to question them further...but one engineer (I can't remember which product...might have been B&K) said that any Dolby Digital signal would sound inferior *NOT* because of the compression, but rather because of filters that the manufaturers were required to impliment in the decoder design to keep in Dolby Digital's spec.

The more I think about this...it sounds like we're putting the two sides of the picture together. Given how much better my 2-channel Audio Alchemy DAC sounded when I made the mod to stop truncating the LSB for red-book CD playback, I can only imagine what degradation is going on inside a DSP chip that's trying to recaluclate the amplitude of the waveform to attenuate the signal.

Lame guys! Dolby Digital...get with the program! If you want audiophiles to take your codec seriously...even your lossless codec...you need to let us get those bits to the DAC without additional signal processing that can't help but degrade sound quality.

And yet I ask...WHY IN THE WORLD would any soundtrack mastering engineer leave the dialog norm adjustment set to anything but 0? What in the world is he/she trying to "adjust" it to match?!?!?

If anything other than 0 basically forces your beautifully mastered soundtrack to go through a digital amplitute-attenuation filter then you've basically SCREWED your signal's chance of having any audiophile-playback-potential.

That alone could be responsible for the lack of "slam" or "atmosphere" or "naturalness" in one Dolby Digital soundtrack versus another "direct" DD or DTS recording...all other things being equal.

BTW, on D-VHS...what's the deal? Anyone know in particular about Moulin Rouge? That's another Dolby Digital soundtrack that had my head spinning saying "Wow...sounds like DTS!" Actually...in that case it sounded BETTER than DTS on DVD. Anyone know if that DD soundtrack was dialog-normalization free?



But I can state as fact that, thus far, whenever a Dolby Digital soundtrack has "wowed" me...and made me think "hey, that sounds as good as DTS"...when I've questioned the mastering engineer I've always been told "oh yeah, we didn't add any dialog normalization to that 5.1 mix" and whenever I thought "hmmm, the DTS mix just sounds more open, more dynamic, more natural" the mixing engineer said "oh yeah, we left dialog normalization on for that Dolby Digital mix."

There has got to be something to this.

 

[Mark Lucas]

I know.

Again if DTS, audio cds, dvd-a, and sacd can be great without dialog normalization why do we need it? And does anyone really understand it?

 

[Michael Reuben]

I do. And it seems to me that David does now. So at least there's two.

M.

 

[ChristopherDAC]

My Yamaha DDP-1 AC-3 decoder was released before there was any home Dolby Digital content; before, I think from reading the manual, the implementation was quite finalised.

Anyway, one of the available setting is "Dynamic Range: MAX/STANDARD". I think maybe this is a DIALNORM defeat switch, but I can't be sure. The description reads:


When "standard" is selected, it's possible to fine-tune the degree of low-level boost and high-level cut. So, does this have something to do with DIALNORM, or is it some other feature still?