This writeup contains lots of numbers, acronyms and abbreviations. This is unavoidable. You have been warned.
...in the annals of historic face-offs, perhaps none has been as heated, controversial and just plain emotional as the one between Dolby Digital and DTS.
I've heard that Dolby Digital isn't as good as some competing systems. Is this true?
...coder technologies such as Dolby Digital can sound just as good as others like DTS which use from two to four times as much data.
- excerpt from Dolby's 'Cinema Sound FAQ'
What is the main difference between DTS and Dolby Digital?
Dolby uses more compression than DTS. Dolby Digital sounds good, but DTS delivers more of the clarity and dynamics of the original master soundtrack.
- excerpt from DTS' FAQs
In late 2000 and early 2001 a series of short papers were published by both Dolby and DTS, each essentially a critique of the other company's product. The first was published by Dolby as an 'evaluation' of DTS, cleverly titled "Dolby Evaluates DTS". Those that followed were DTS' rebuttal, Dolby's rebuttal to DTS' rebuttal and finally DTS' rebuttal to Dolby's rebuttal of their rebuttal. This writeup explains and comments on these exchanges.
First, a quick primer. Dolby Digital (hereafter referred to as DD) is an audio compression and encoding scheme capable of encoding up to six discrete channels (commonly referred to as 5.1 channels, since the sixth channel - the low frequency channel - does not use full signal bandwidth). The soundtrack is printed onto one side of the film roll itself, read by hardware on the projector and subsequently fed to the theatre sound system. It was introduced in 1992 with the film Batman Returns and several years later became the first discrete multichannel sound system available to consumers (at the time it was - and occasionally still is - referred to as AC-3, for Audio Codec 3). The most recent revision to the codec (THX-EX, aka Dolby Surround EX or Dolby Digital EX) added a seventh 'back' channel, matrixed from the rear channels much like Dolby Pro Logic derives a centre channel from the front channels.
DTS - Digital Theatre Systems - is another compression and encoding system that debuted in 1993 with the film Jurassic Park, with the personal endorsement (and financial backing) of director Stephen Spielberg; all of his subsequent films and most of his previous films are commercially available with DTS soundtracks. DTS was the first sound system to be completely digital from the encoding stage: for theatre applications the soundtrack is recorded onto two CDs that are played on hardware in the theatre, synchronised with the image using a time code along one edge of the film. Because the soundtrack is stored on compact discs (much like the music on DTS music discs) far less compression is needed; DTS uses a data rate almost four times that of Dolby Digital which, DTS claims, produces higher sound quality. A recent iteration of this codec added a seventh, discrete 'back' channel (DTS-ES) while remaining compatible with older 5.1 DTS systems.
The debate over which of these systems is superior is endless, spanning many thousands of usenet and web forum posts. My personal view is still not fully formed, for several reasons:
- I have never heard the same film at full-rate Dolby Digital and full-rate DTS, on the same hardware. Comparatively few films have had separate DD and DTS releases for home use, but it is these over which the majority of debate ranges. For example, many DTS aficionados cite releases like the Saving Private Ryan DVD - which is available in separate full-rate DD and DTS releases - as evidence of the superiority of DTS. While I do own said DTS DVD and will instantly cite it as the most impressive surround track I have ever heard, I have never heard the Dolby Digital version (though note the comments about the same DVD in above w/us).
- Hardware limitations. My home cinema system has speakers that cannot reproduce the full bandwidth signals of both systems. While I might notice differences in clarity, effects placement or consistency between DD and DTS soundtracks I can't fairly compare dynamics.
- Since DTS has 1/4 the compression ratio of DD, a DTS soundtrack takes up far more space on a DVD disc than a DD soundtrack. This is at least part of the reason for the limited adoption of DTS on DVD releases (virtually all major films are released with prints that support all three encoding schemes - DD, DTS and SDDS); the DTS soundtrack leaves little room for additional content. The DVDs of Apollo 13 and 12 Monkeys, for example, don't even have subtitles. Because of the high storage cost of full-rate DTS, any DVDs that include 5.1 or 6.1 channel soundtracks encoded with both DD and DTS, use the 'half-rate' version of the DTS compression scheme. This has a bitrate of 768Kbps rather than the 1509Kbps of the original scheme. In my - as I said, limited - opinion, trying to compare this version of DTS with DD is pointless.
Update: I recently obtained a copy of the full-rate DTS laserdisc of Jurassic Park, which pretty comprehensively trounces the later, half-rate DVD of the same. I'm tempted to use this to illustrate the superiority of full-rate over half-rate DTS, but it is possible the mixes are different (at that period of time DTS movie releases were encoded, possibly mixed by DTS themselves - these days production companies purchase DTS encoding hardware and do their own mastering) so I won't. I will simply say it is a stunningly powerful DTS track, and if you have the playback facilities you should definitely try to acquire a copy of this disc, as scarce as it is.
Different encoding schemes. Due to the aforementioned DVD capacity issues, DTS came up with a codec that operated at a reduced data rate, but claimed it lost no quality over full-rate compression. For reasons I don't know there are also two encoding rates for DD, but the difference in data rate of the two is far less than the difference between DTS' two encoding schemes, so one wonders why they exist. Furthermore it is difficult to tell what version of the DD encoder any given soundtrack employs, and their existence muddies the waters further when trying to make qualitative comparisons.
It is commonly held that DTS' higher encoding rate certainly gives it the potential to sound better than DD, although it's obviously not as clear-cut as that or the usenet would be a much quieter place. I do have a soft spot for DTS, but it probably has similar elitist roots to my acquisition of laserdisc hardware back when most people were using VHS (although the difference in that instance certainly is clear-cut). This writeup will doubtless solve no disputes but I do think the 'views' of the two companies deserve a writeup.
Episode 1 - "Dolby Evaluates DTS" (November 2000)
Dolby states their aim as resolving the differences reported by listeners when comparing Dolby Digital and DTS soundtracks of the same source - for example a DVD with both Dolby Digital and DTS soundtracks. Dolby opens by stating they have performed a qualitative comparison of the DD and DTS codecs, now that DTS' CAE-4 encoder is on public sale. They also state that the differences highlighted in the paper, because they were found in a studio environment, would be "small or insignificant to most listeners, and in many home playback systems...might go unnoticed altogether."
Dolby tested the two systems together using both encoding rates available on each; 448kbps and 384kbps DD against 1509kbps and 754kbps DTS. Without elaborating Dolby states they tested three different versions of DTS' CAE-4 encoder, with identical results from each.
Quoting their findings Dolby gives short shrift to full-rate DTS, apparently more interested in giving paragraph space to criticise half-rate DTS. Their measurements show a bandwidth ceiling of 15kHz for 754kbps DTS, and a gradual decrease in LFE (subwoofer) channel response across the board, starting from 0dB at 20Hz, to 1dB at 50Hz to 3dB at 90Hz (*hairline recedes visibly*). A frequency plot is shown to support this. In comparison Dolby states DD exhibits full 20kHz bandwidth at 448kbps and 18kHz at 384kbps, with less than a 0.1dB reduction in LFE channel response right down to 120Hz.
Dolby are saying that half-rate DTS cannot encode frequencies above 15kHz at all (typical adult hearing ranges top out at 20kHz, dropping gradually with age), and that DTS as a whole has trouble reproducing frequencies at the upper end of the scale that the subwoofer channel deals with. Predictably, Dolby goes on to say that DD encodes these frequencies flawlessly.
Continuing, Dolby states both DD and DTS-encoded material exhibited noticeable artefacts at all encoding rates; the tests were performed with 'coder killer' material: soundtracks that by nature are difficult to compress and encode effectively. Dolby finishes by saying that half-rate DTS performed worst, exhibiting the most artefacts.
Codecs like DD and DTS are called 'adaptive', or 'lossy' compression schemes. Some stuff has to get thrown away; that's the nature of compression. Compressed audio will never be identical to the original, but audio codecs like DD, DTS and Sony's ATRAC (what is used for Minidisc and SDDS, Sony's digital surround system) are designed to try only to throw away stuff you won't notice and retain the essence of the source material to the maximum extent possible. Much like some image compression schemes (like JPEG) can introduce artefacts to the encoded image, compressing and encoding sound can also have minor undesired side effects. Some codecs are better at minimising this than others, though results are often subjective. Dolby says that in this respect, half-rate DTS is the worst in this group test.
The next section of the paper concerns listening tests conducted in three studios, though Dolby do not record any of the test conditions. This section contains rankings for tests performed in three studios, with 448kbps DD getting first place on two of them and 1509kbps DTS getting first on the third. Second and third place are a mix of both DD schemes and 1509kbps DTS (except in the third case), but 754kbps DTS comes last in all three. Dolby even adds that half-rate DTS was "distinctly below the other three options."
The section finishes with criticism of full-rate DTS:
"DTS full rate imparts subtle audible colorations compared to the source. These may be summarized as an emphasis in the 800Hz-1kHz region; a dulling of the top octave, a sense of compression on percussive transients, and an emphasis in the frequency region of 8-10kHz. Listeners also described a loss of ambience or "air" and spatial dimensionality compared to the source."
...continuing on to criticise LFE channel response, citing their earlier findings. Again Dolby states their codecs do not suffer these types of problem.
The next section concerns source material for film soundtracks and mastering, attempting to explain the so-called 'differences' heard by many, that had not been noticed under Dolby's "carefully controlled conditions."
Dolby concludes there are three reasons for these differences. Firstly, disparity in overall volume level between DD and DTS software: DD uses a feature called 'dialogue normalisation' which usually reduces overall volume by 3dB, in an effort to standardise volume across all Dolby software. DTS does not use this feature; according to Dolby most listeners would misinterpret the increased volume on DTS software as sound that has more "punch or dynamics compared with the source." I am unsure why Dolby would note this, or why it would assert that level-matched comparisons may be flawed for this reason, since it undermines the listening tests' obvious purpose as demonstrating DD's superiority to DTS.
Secondly, differences in soundtrack mastering: Dolby claims that film soundtracks are frequently remastered or have their mix altered for different releases (obviously pointing at DTS releases of films, which at the time were frequently released some time after the DD version). That differences listeners notice in sound quality are down to changes made to the digital master, rather than differences in codec efficiency between DD and DTS. This also seems to undermine Dolby's earlier assertion of DD's superiority.
Finally, and similar to the second point, Dolby claims differences between DD and DTS material may also be explained by remixes of the source material. The DVD for The Haunting is mentioned as having an 'original' Surround EX soundtrack, but that the DTS ES version was remixed to add the extra discrete channel. Dolby then mentions another DVD that had its soundtrack remixed to add Surround EX to it; again I wonder why it would mention that this may be applied to DTS and DD masters.
The paper concludes that the issues Dolby report (the three points above) show that differences in quality between DD and DTS are not necessarily due to codec quality, then apparently and somewhat incongruously uses the same findings to illustrate that "Dolby Digital not only equals DTS, but in fact outperforms it". The last paragraph admits Dolby's bias (duh), but encourages others to carry out the same tests as they have.
Episode 2 - "DTS Position on "DOLBY EVALUATES DTS"" (November 21, 2000)
I read this paper before the others, sometime in early 2001. It is solely concerned with attempting to refute Dolby's claims about the performance of the DTS CAE-4 encoder (though DTS do take a couple of swipes as well) and contains some priceless statements among the corrections. My own bias probably colours that perception to a degree somewhere between 'slightly' and 'considerably'.
The first paragraph establishes DTS' contempt for criticism from Dolby:
"[Dolby's] paper, as might be expected, asserts the superiority of Dolby Digital over the DTS Digital Surround system. Such a finding from a principal competitor is suspect on its face. It appears that the paper is being appropriately discounted by informed observers,"
DTS corrects Dolby's statement of 754kbps DTS' bandwidth limitations, stating that it is actually 19kHz, not 15kHz as Dolby reported. DTS goes on to note a 24kHz bandwidth for 1.5Mbps DTS at a 48kHz sampling rate. There are also corrections of Dolby's various statements about the frequency response of DTS' LFE channel. Where Dolby claims it to roll off 1dB at 50Hz and 3db at 90Hz, DTS claims that their codec's LFE channel has a flat response up to 100hz, increasing up to -6dB at 125Hz.
DTS then amplifies Dolby's statement about the performance of the DD codec, that being that it does indeed reproduce all frequencies up to 20kHz at 448kbps and up to 18kHz at 384kbps. DTS goes on to claim that DD combines channels above 15kHz at 448kbps and above 10kHz at 384kbps.
The three paragraphs above are corrections on the performance of DTS' two encoding schemes. DTS first states the CAE-4 can effectively encode frequencies up to 19kHz at half-rate, and 24kHz at full rate. Secondly, DTS adds to Dolby's statement about the 20kHz bandwidth of the DD codec, claiming that DD has no channel separation for frequencies above 10kHz or 15kHz, for the low and high encoding rate respectively. In other words sounds above those frequencies would be heard equally from all channels with no sound placement. Note that the highest frequency the average adult can pick out is about 16kHz. Finally, DTS states the CAE-4 encoder reproduces low frequencies up to 100hz without distortion, increasing in response up to 125Hz.
DTS make a rather peculiar point next, because it appears to quote a section of Dolby's paper that does not exist. It seems likely Dolby later edited this out of their paper, since Dolby make no reference in their subsequent paper to DTS commenting on a statement Dolby never made. DTS quotes Dolby as saying "The DTS CAE-4 encoder is a nearly featureless product -- it does not support downmixing and dynamic range control, two important features that DTS claims to support." DTS' response is that the audio quality of the CAE-4 is its primary "feature", that it is intended to produce a true representation of the original master without change and that other encoders are available with the features Dolby mentions.
DTS dismisses Dolby's listening tests as farcical since no details about them are disclosed, combined with the Dolby's bias as the moderator of the tests:
"If Dolby wishes to assert that DTS at 1.5Mbit/s is perceptibly different from the unencoded original, it would behoove them to prove it."
DTS challenges Dolby to have a blind test conducted by an impartial third party, comparing a DTS-encoded soundtrack to the unencoded master and see what the results are...
"There would be no reason to include Dolby Digital in the test, since its lack of transparency is a matter of scientific record (cf. Soulodre, et al., Journal of the Audio Engineering Society, Vol 46, #3 ["Subjective Evaluation of State-of-the-Art Two-Channel Audio Codecs"])"
The remainder of the paper deals with Dolby's statements point by point. DTS discounts Dolby's statements about dialogue normalisation (saying that DTS encoders used to add 0.6dB to audio, but that at time of writing the feature had been removed from the CAE-4) as contradictory, saying that Dolby admits to never having been able to test the CAE-4 before now, so it could not possibly have been aware of such a change. DTS further denies that changes were ever made to the CAE-4.
DTS continues, dismissing as nonsense Dolby's writing about consumer perception of the lack of dialogue normalisation in DTS soundtracks ("most listeners...would describe the louder sound as having more 'punch' and 'dynamics'"). They also agree with Dolby's point about different masters, saying that they would not make for a fair comparison of the two encoding schemes. DTS also writes that if this were the sole reason for one soundtrack being preferred over the other, that the balance of consumer preference would be roughly equal. However, they claim, "DTS is typically preferred." A quick skip around various web forums and the usenet does seem to support this statement.
The next point regards Dolby's reasoning that changes to soundtracks during mastering may explain some people's preference for DTS. DTS' response is that they always recommend against such changes. Part of Dolby's statements to this effect included testimony that the DTS laserdisc of Jurassic Park sounded very different to the later DVD (I personally found the sound of the DTS DVD very disappointing compared to the DTS laserdisc). DTS responds that they encoded the laserdisc themselves, whereas Universal (the film's distributors) encoded the DVD; they deny altering the masters, as Dolby seems to imply.
Further comments by Dolby specifically cited two music DVDs (Steely Dan - Two Against Nature and Dave Grusin - West Side Story), claiming that the sound was artificially "sweetened" during the encoding process. DTS denies this was the case for the former, referring to the publisher's statements that both DD and DTS versions came from the same master. In the latter case, DTS admit to changing the mix but states it was at the request of the record company.
Finally DTS addresses Dolby's various statements of dubious purpose about specific DVDs (The Haunting and Terminator 2) that were remixed to add a DTS ES soundtrack. In TH's case, the original master was mixed in DD EX. As DTS states the back channel in DD EX is matrixed (like the centre channel in Dolby Pro Logic), whereas DTS ES can provide a discrete back channel. Therefore a remix was essential for DTS ES encoding - this was performed under the direction of the original mix engineers, "to provide optimum separation without disturbing artistic intent." DTS states this procedure is always followed for remixes such as this. In T2's case, a remix was necessary for both DTS ES and DD EX (when T2 was originally released, Dolby SR was standard), so it doesn't help Dolby's case to have mentioned it.
The paper sums up in a slightly disenchanted tone, urging comparative tests but to compare codecs to masters rather than compare codecs with one another, and to have tests conducted by impartial parties. DTS goes on to recommend DD for restricted-bandwidth situations, but states DTS to be superior when accurate source representation is critical.
Episode 3 - "Dolby Reviews DTS's Position" (March 2001)
The main purpose of this paper is for Dolby to clarify its original statements, though a couple of concessions are made. Dolby begins by criticising DTS for not backing up any of the claims in their paper of November 2000. The fact Dolby gave no details about its listening tests is not mentioned.
Dolby amplifies their original statement about 754kbps DTS' frequency response up to 15kHz, saying that frequency response does indeed reach 19kHz (as DTS states), but that the response from 15-19kHz is reduced by 4dB+/- . A plot of frequency response is given to support this, though no mention is made of measuring equipment or test material. Dolby also maintains its position regarding the rolloff in frequency response of DTS' LFE channel, stating the same response was also found when measuring actual DTS content (track 13 of Steely Dan's Two Against Nature DVD). Frequency plots are given to support both of these statements.
The next issue addressed is Dolby's earlier statements regarding the adding of 0.6dB to DTS content when compared to the same content encoded in DD. Dolby repeats their assertion in the face of DTS' denial and criticism, saying that the measurements were made with two DVDs (Twister and Interview with the Vampire - one encoded by the film studio and the other by DTS itself), that no DTS encoder was necessary to make them and that anyone can make the same measurements. Dolby admits the CAE-4 encoder it purchased does not exhibit this characteristic, but the study of DTS content seems to support its claims. Dolby criticises DTS for not backing up their assertion that a 0.6dB difference in level would be "inaudible in an A/B listening session."
Further, Dolby claims that the studio that made both DVDs involved in the test was asked to check the calibration on their encoding equipment. Using a 1kHz test tone a 0.6dB difference was found between the DD and DTS encoders. The studio is said to have reported this to DTS, who then "updated the code in the unit," removing the gain. The studio involved is not named, though given the two DVDs mentioned it would have to be Warner Home Video.
Dolby continues to the issue of 'remixed' DTS versions of film soundtracks, focusing as before on the DVD of The Haunting. This film was originally released encoded with a Dolby Surround EX soundtrack. Any change to this, Dolby states, apparently even if done under the direction and approval of the original mixing engineers (as DTS claims is always the case for soundtracks that are remastered for DTS encoding) is a "compromise", with a "clear possibility that the two will sound different." According to Dolby, film soundtracks that are remixed for home use do not fall under this category, since according to Dolby they remain "exactly representative of [the video company's] wishes."
In the final section Dolby spends two pages restating its position on artificial sweetening of DTS musical recordings and DTS LFE channel performance. Apparently DTS was correct in its response that no sweetening had occurred for the music DVD Dolby mentioned. Dolby investigated further and found that the LFE filter for the Two Against Nature DVD's DD encoder had been turned off, stating that "the LFE low pass filter in Dolby Digital encoders normally cuts off steeply above 120Hz." A frequency spectrum plot of the DTS and DD LFE channels together is shown to support this. This becomes Dolby's explanation for the difference in bass response between the two versions of the Two Against Nature DVD.
Dolby goes on to say that the previously-mentioned sine wave plot of DTS' LFE channel response was compared to the result of the same measurement taken from the Two Against Nature DVD, showing the two as identical (the combined plots were previously shown separately in the paper). Dolby states, "this proves that the entire rolloff seen in the DTS version's LFE channel is attributable solely to the DTS encoder's LFE filter." Dolby further states that the original source master "obviously contained audibly more bass than the DTS version delivered", supported by the fact the DD version did indeed contain significantly more bass than the DTS version did.
Dolby slightly confusingly finishes this section off by saying that the DD version of the DVD had a sound that was closer to the original, that DTS was unable to reproduce the full spectrum of the LFE track in the 6-channel master, but that DD "should not be asked to carry such wideband signals, as it can lead to aliasing components." Dolby then writes that to get best results the LFE channel should be omitted altogether!
In conclusion, Dolby does not reassert DD's superiority to DTS but does state that differences between soundtracks from the same master are not necessarily due to differences in encoding, again implying that DTS soundtracks are taken from different masters to DD soundtracks.
Episode 4 - "DTS Position on "DOLBY REVIEWS DTS...""
As before, DTS is concerned with refuting Dolby's statements in this paper, as well as repeating some of the challenges they made. Beginning with the criticism of 754kbps DTS, DTS agrees with Dolby's measurements and that DD has a slightly higher frequency response than 754kbps DTS, but states this behaviour is "characteristic of the implementation in [the CAD-4] decoder, not the encoder". DTS notes Dolby's lack of response to the point it made in response, that DD combines channels at high frequencies while 754kbps DTS does not.
DTS then addresses the thorny issue of LFE channel response, restating their original figures on the performance of the DTS encoder. They then raise an issue potentially damaging to DD, that of "group delay", or the phase of the LFE channel compared to the main channels. DTS states that their encoder/decoder set produces an LFE channel that leads the main channels by about 1.6ms between 20Hz and 120Hz, whereas DD's DP569 encoder and DP562 decoder exhibits a group delay of up to 7.8ms, or 225 degrees between 60Hz and 80Hz. DTS states DD's LFE channel lags behind the main channels by 180 degrees at 80Hz and provides a plot of main channel response against LFE channel response to back this up. If true DD's bass reproduction could be seriously affected, particularly if the user has employed bass redirection circuitry to send bass of the main channels to the subwoofer.
Think of group delay in terms of the famous (if an AV geek) 'lip-sync' problem that plagued some early DVD players and DVD discs. An actor would be talking and the sound would not match up with their mouth moving. According to Jim Taylor's DVD FAQ this is/was caused by:
Humans can typically detect a 1 frame, or 42ms, difference between sound and image. However the delay DTS is talking about is so small it relates more to how sound is heard rather than when it is heard.
- Improper sync in audio/video encoding or DVD-Video formatting.
- Poor sync during film production or editing (especially post-dubbing or looping).
- Loose sync tolerances in the player.
- Delay in the external decoder/receiver.
Phase is the other issue important to understanding this. Think of a speaker cone. It vibrates - moves in and out (very quickly) - to make sound. Whether it moves in or out at a given point depends on the phase of the signal it is receiving. If a speaker cone 'pushes' when given a certain signal, it would 'pull' if the same signal were sent in opposite phase. Swapping over the cable connections on the speaker (the positive and negative terminals) would have this effect. If you were to take two loudspeakers, reverse the wiring on one of them and place the two directly facing one another a few millimetres apart, they would cancel out virtually all sound they produced. When one speaker is pushing, the other is pulling (note - this is a useful way of 'running in' or softening up new loudspeakers).
This is what DTS is talking about when it mentions the 180-degree phase difference of DD's LFE signal with the main channel signals. To say nothing of the presence of delay - undesirable in itself - if the user has invoked bass redirection circuitry (common on upper-medium range equipment and above), then the bass from the main channels will be redirected to the subwoofer channel, which is already receiving the same signal in opposite phase. This means that if the subwoofer receives signals from the main channels and the LFE channel on the same frequency, at the same instant, they will cancel each other out. If DTS is correct and this does occur, DD could have quite seriously impaired bass reproduction.
Later note: I have subsequently realised that if this problem exists it is unlikely to cause problems with sound reproduction. The LFE channel generally only handles frequencies below those of the main channels (that's the whole point, to handle frequencies below the capabilities of the main speakers), so it is unlikely that the main channels and the LFE channel would ever receive signals on the same frequency.
DTS then addresses the supposed 0.6dB gain that Dolby mentions in its second paper. It questions the validity of a statement, attributed to an unnamed studio, which contradicts DTS' own statement that the CAE-4 encoder has not been modified and contradicts the fact that Dolby's own CAE-4 does not exhibit this gain. Dolby previously stated that DTS had not proved this difference to be inaudible; DTS responds that "inaudibility cannot be proved, only supported by statistical data. Audibility, in contrast, can be proved, so the burden is Dolby's -- any valid test being subject to standard protocols and absent participants with vested interests."
The final sections of the paper deal with different mixes and "sweetening" again. DTS expresses disbelief at Dolby's arguments for the retention of a matrix soundtrack in preference to a discrete soundtrack. DTS also repeat their claim that there is no loss of artistic intent when any such remix (for re-encoding in discrete form) is performed under the supervision of the original mixing engineer(s). Regarding the Steely Dan DVD, DTS highlights the fact an error was made during the DD encoding of the soundtrack and Dolby's lack of apology for their false allegation about DTS altering the DVD's soundtrack prior to encoding.
The paper ends by referring the reader to DTS' earlier points and the repeated sentiment that the two companies' time would be better spent developing their respective technologies. Apparently, Dolby agreed this time.
The DTS compression scheme seems to have significant support among AV enthusiasts. The battle lines between Dolby and DTS, however, are somewhat different to the conventional view of one product vs. another: the arguments are generally not between lovers and haters of DTS but between people who say omg DTS is teh bestest thng eva!11!1!!one roflmao and people who think it sounds no different to DD. Even doing a quick search on the usenet for the phrase "DTS sucks" does not reveal the expected results, with very little criticism about sound quality. Substituting "Dolby Digital" for "DTS" in the same search phrase does show some criticism but again, the volume is far less than I expected it to be. My personal view (as far as I can discern - see my earlier comments) is that DTS is better, but it's the difference between a Porsche and a Ferrari, not the difference between the London Hilton and your local B&B.
Dolby's statements about DTS' LFE channel response are highly interesting, certainly since bass response is a factor often cited in DVD reviews that favour the DTS soundtrack over the DD soundtrack of a given film. DTS do admit to variations in the performance of its compression scheme's LFE channel (although they are gains, not losses as Dolby claims) and even seem to try a little misdirection in the face of repeated criticism from Dolby (by raising the 'group gain' issue). Dolby's claims about DD's perfect-across-the-board LFE performance are par for the course, though not exactly consistent with their recommendation against its use, or with the delay that DTS claims occurs.
I don't quite understand Dolby's banging on about mixes being altered resulting in a different sound since hey, that's kind of the idea chaps. The specific issue Dolby cites is DTS remixing a 5.1 channel soundtrack to produce a 6.1 channel master for encoding in DTS ES. This would of course be essential in such a case; however, remixing or "sweetening" a soundtrack for a 5.1 channel DTS soundtrack that is already available in DD, for example, would be plain cheating. If, as DTS claims, DD encoding compromises the sound of the master and DTS encoding does not, then the DTS-encoded version will require no "sweetening" to sound different to the DD-encoded version. DTS denies that they have ever remixed soundtracks and states they recommend against anyone else doing it. It seems clear a film studio has done it at least once but DTS hardly deserves criticism for this.
It makes little sense to introduce a product that has identical capabilities to those of a well-established competitor. On paper DTS' compression scheme certainly has more potential than DD and more versatility, since it allows discrete 6.1 encoding rather than the matrixed 5.1 of Dolby Surround EX.
A surround sound encoding scheme that allows discrete encoding of all its channels can only be preferable to one which employs matrix encoding (even if only for one of its six channels), otherwise Dolby would still be marketing Pro Logic and would never have developed Dolby Digital. With this in mind, Dolby's criticism of DTS for remixing Surround EX soundtracks for encoding to DTS ES seems even more baseless and as DTS states in its latter paper, "it is difficult to believe Dolby advances this argument with any conviction." Dolby states that differences between the two encoders are so slight they will likely go unnoticed outside a studio environment, yet somehow its first paper reaches the conclusion that DD tangibly outperforms DTS.
However, much as I hate to admit it, Dolby does at least attempt to provide some data to support its claims (which do seem slightly like sour grapes in the face of increasing support for a direct competitor - after all, you don't see washing powder companies conducting public dissections of their competitors' products) unlike DTS which for the most part present facts as facts with no backup. According to DTS, Dolby isn't being completely forthcoming about the limitations of its own product but then, according to Dolby, neither is DTS. Nothing surprising there, however Dolby fails to challenge the issues DTS raises about the performance of DD (although one might argue that isn't the point, as the exchange started over DTS, not DD). DTS on the other hand, convincingly refutes most of the points raised by Dolby and concedes most of the remaining points with slight modification. So, that should solve all of the arguments.
The documents cited can be found in their entirety at:
- www.dolby.com/tech/mp.in.0103.DigitalVsDTS.pdf (both Dolby's papers)
- www.dtsonline.com/media/uploads/pdfs/dtsposition.pdf (DTS' first paper)
- www.dtsonline.com/media/uploads/pdfs/dolbyrvu.pdf (DTS' second paper)