PCM to DTS 1.5mbit Transcoder?

[consultant]

I'm trying to find a happy medium between filesize and audio quality for my Blu-Ray rips. I did some reading and based on real-world blind listening tests, regular DTS 1.5mbit (non-HD, but not the 768kbit on a lot of DVDs) is supposedly 'transparent' to human listeners - in other words people can't tell the difference between DTS 1.5mbit and the lossless source. (I know there are audiophiles that will swear they can tell a difference between TrueHD/DTS-HD and regular DTS at the full 1.5mbit rate, but I call that the 'tech spec placebo effect'. I actually think putting PCM tracks and HD tracks on Blu-Rays was primarily a marketing hype move, and a move to make pirating movies without transcoding more difficult due to the storage requirements -- not because there was any significant benefit in audio quality - but that's really another thread.)

Anyhoo, when the Blu-Ray has DTS 1.5mbit or DTS-HD, great no transcoding just pass it thru or in the case of HD, extract the core. But for PCM and TruHD, if people want quality better than AC3 640kbit but filesizes less than PCM/TrueHD/DTS-HD, what are you doing, transcoding to FLAC? Does PCM transcoded to FLAC end up with a filesize smaller than the same audio in TrueHD or DTS-HD? I just want something better than AC3 640kbit but filesizes less than TrueHD and DTS-HD that will stream to my PS3 and to the latest Popcorn hour. In some cases I may want to burn a disc to take with me on vacation or something but I suppose the FLAC can be transcoded to AC3 640kbit for that purpose. I'm guessing FLAC is my only option?

If I could, I'd transcode to DTS 1.5mbit since as far as hardware players go, it is more compatible than FLAC, I think. But I don't think there are any free DTS encoders? So as far as filesizes 'in between' HD formats and AC3 640kbit, aside from DTS 1.5mbit and FLAC'd PCM (asuming it is smaller than TruHD and DTS-HD), are there any other options?

[Atak_Snajpera]

Don't bother with stupid inefficient dts! AC3 reaches transparency at 640 kbps . DTS requires 1,5 Mbps!

[tebasuna51]

1) First of all, there are some test than prove the AC3 640 Kb/s is more transparent than DTS 1.5 Mb/s. Only with a very good audio equipment, equipped room and good ears you can distinguish from the source uncompressed audio.

2) Don't exist free DTS encoders, the free AC3 encoder (Aften) isn't Dolby certified (certified implies pay) and don't know test to prove than Aften offers the same quality than certified ones.

3) About filesizes I make a simple test with an audio track fragment (to be BluRay compliant a TrueHD must have a included AC3 inside):
File Size %
----------------------- --------- ------
T.wav_.ac3 24076800 13.89
T.wav_.flac 49012532 28.28
T.wav_.dts 56790756 32.76
TrueHD_5.1_16.truehd 58244568 33.60 (original)
T.wav_.dtshd 77363092 44.63
T.wav_.thd+ac3 82321316 47.49
T.wav 173338628 100
Like you can see the FLAC track can be smaller than DTS, with some tracks, the problem is the hardware players compatibility.

I always use AC3 640 Kb/s for my backups because I can't ear differences.

[Atak_Snajpera]

Like you can see the FLAC track can be smaller than DTS, with some tracks,
I've noticed the same with 'Son of Rainbow' movie. FLAC was smaller than DTS core :)

[consultant]

I completely disagree that AC3 640kbit is more transparent than 1.5mbit DTS.

First off, let me point out that every post in any forum I've seen of someone claiming something like this always says something like "its been proven" or "blind tests show" and then they can't provide a link to "the test" or "the study"

The bottom line is, that most people make the mistake of trying to compare the two as apples to apples. The reason is that there are underlying parameters of the encoding that are not always the same and the algorithms are different. Actually, AC3 has the potential to be encoded with more inferior parameters. Here's what I mean:

Dolby Digital (AC3) is always encoded with 16-bits.
DTS is encoded with minimum 16-bits but can be encoded with 24-bits.
Dolby Digital sampling rate is 32, 44.1 or 48 khz. Although all DVDs and Blu-Rays will probably always use 48.
Sampling rate on DTS is minimum 48 or, optionally a much higher 96 khz.
Dolby Digital can have a compression ratio ranging from 4-1 to 12-1
DTS only goes up to 10-1

You can go here to verify these claims:

http://www.tomsguide.com/us/high-def-audio-hdmi,review-1088-3.html

So when you compare the two they have to have the same parameters. Although we can assume I think anyone making a broad claim is assuming the bit depth, sampling rate, and compression ratio are the same. With that being the case then the only differentiating factor (aside from the underlying algorithms - which is what we want to compare in the first place) becomes the bitrate.

Essentially Dolby Digital has put statements out that AC3 640kbit achieves the same transparency as DTS. This is a claim not being made by an independent third party, so while it may or may not be true, you can at least say the claim is suspect because of who is making it. Also not they don't say "at the same bit depth, sampling rate, and compression ratio" Because of course you could easily make an AC3 640kbit sound as good or even better than DTS if you compared AC3, 16-bit, 48 khz, 4-1 compression to DTS, 16-bit, 48 khz, 10-1 compression.

What you see on most web pages that posters have read is the regurgitation of the information on the DTS Wikipedia entry which states:

"technical experts rank DTS as perceptually transparent for most audio program material (i.e., indistinguishable to the uncoded source in a double blind test)."

but also includes Dolby's statement:

"Dolby claims its competing AC-3 codec achieves similar transparency at its highest coded bitrate (640 kbit/s)."

Note the word "similar" not "better" not "identical" not "almost", but a very vague "similar."

http://en.wikipedia.org/wiki/DTS_%28sound_system%29

My opinion is that people stating AC3 640kbit is the same or better than DTS 1.5mbit, with the same bit depth, sampling rate, and compression ratio, without providing reference to a credible double-blind listening test, are essentially deciding to "drink the Dolby Kool-Aid" and take what they say for face value.

If you are interested in the difference regarding the algorithms, this is a good write-up I've seen on it. But really, why do we care, all we care about is how does it sound. However I this this statement is spot-on:

"Reviewer opinions are varied, however the range is from DTS-is-better to they're-pretty-much-the-same. "

http://www.hifi-writer.com/he/dolbydts/dolbydts.htm

If you REALLY are interested in comparing audio codecs, this is an excellent study. It doesn't directly compare DTS 1500 with AC3 640, but if you look on page 21, out of all the codecs tested DTS 1500 came out on top. BUT, the runner-up is DD+ 448! DD+ (EAC3) is a better codec that DD (AC3) but I would venture to guess EAC3 448 is no better than AC3 640. So I would extrapolate from THIS study there is no perceptible difference between AC3 640 and DTS 1500.

This is my favorite write-up on the subject:

http://www.hemagazine.com/node/Dolby_TrueHD_DTS-MA_versus_Uncompressed_PCM?page=0%2C1

"Whatever acoustic elements were removed in the code/decode process were clearly superfluous"

This listening test essentially concludes that the difference between AC3 640 and DTS 1500 to the lossless PCM for listeners is either indistinguishable or so tiny that it is superfluous. My toungue-in-cheek translation: If you care about a superfluous, difference then your life priorites/energies are probably being focused on the wrong things. Of course I am then setting myself up to sound like a hypocrite later in this novel.

If you are a "Golden Ears Morrison" I'm sure you would argue my point that it a 'superfluous' difference is insignificant. I know the definition of an Audiophile.

I did note that on the DTS they said the bitrate was 768kbit OR 1.5mbit. I'm disappointed they didn't say if they could tell a difference between the two and was the statement "the compressed version differed slightly only in barely noticeable presence" are they referring to 768kbit, 1.5mbit or both?

Here's the Bottom Line FOR ME:

There's probably no perceptible difference, all things being equal, but, I prefer DTS over AC3, WHEN I HAVE THE CHOICE, because the spec has a higher minimum sampling rate (minimum 58khz) and a lower maximum compression ration (10 to 1 instead of 12 to 1). How often does AC3 go below 48khz and above 10-1 compression, probably not often. The storage requirement on DTS 1.5mbit is so much smaller than TrueHD, DTS-HD or raw PCM, that I don't care if I'm wasting a little space for not getting any perceptible gain in audio quality over AC3 640. But I take comfort in knowing if there IS a perceptible difference, I'm getting it.

If all I have available to me is AC3 640, and don't have raw PCM I can convert to FLAC - I'm happy with it. Because to me the bottom line is that all these HD audio formats are really just storage pigs and were created like a lot of new things in this world just to for marketing hype just to say "LOOK! THE NUMBERS ARE BETTER! IT HAS TO BE BETTER!" I still firmly believe studios are puting raw PCM and DTS-HD MA on discs in order to make the filesizes as large as possible to discourage consumers from ripping the discs to storage since transcoding is SLOW on the average CPU. We all know you can get a Blu-Ray main movie down to 1/3 the size or less than it is on the disc with no perceptible loss in video or audio quality. You can fit 62 8GB movies on a 500GB drive but only 16 if the filesize is 30GB. In addition you can burn an 8GB file to a cheap DVD9.

So in my mind the more interesting argument is not which is better AC3 or DTS, but is there any significant benefit of TrueHD and DTS-HD, aside from having 7.1 instead of 5.1 capabilities?

In conclusion it really seems like RAW compressed to FLAC is really the best way to go when possible. It is kind of nuts though that Dolby and DTS spend all this time and money and all the receivers are certified, but FLAC beats them all as far as the quality/filesize ratio.

[7ekno]

I prefer DTS over AC3, WHEN I HAVE THE CHOICE, because the spec has a higher minimum sampling rate (minimum 58khz) and a lower maximum compression ration (10 to 1 instead of 12 to 1).

That, my friend, is called the placebo affect ;)

If DTS is so superior in your mind, pay the DTS encoder fee and be done with it ...

I can not hear the difference on a $10k theater setup (and I have blindly tested a group of 11 friends, none of which could pick the DTS source tracks with an accuracy of greater than 72%, P=0.28), and given no free encoders and the massive waste of bits, I will stick with 640kbs AC3 :P

and don't have raw PCM I can convert to FLAC - I'm happy with it.

But using your own logic, FLAC is inferior to AC3 because in it's specs, the sample rate can be as low as 8KHz!!!! Either your logic is wrong, or your reason for the use of FLAC is wrong ;)

7ek

[shon3i]

Dolby Digital (AC3) is always encoded with 16-bits.Hmm, there is no bit-depth in lossy encoders such DTS and AC3 they are both operate at floating point, so this statament is completly no sense. Anyway you want prove, so here's is one of most ABX document that shows AC3 448 have tied scores with DTS 1.5 over 90% which is overkill for most listeners http://www.ebu.ch/CMSimages/en/tec_doc_t3324-2007_tcm6-53801.pdf Btw I'm not one of those who rely on it, i do my pesonal ABX primary with Sennheiser HD 650 headphones, i compare channel by channel, from lossless source.

FLAC in this case is usless for me, because is not standard, and is not popular in most standalones. Beside is most popular in stereo world of arhiving and lossless.

For multichannel i would go with AC3 or E-AC3 (7.1) and AAC ofcourse.

[rik1138]

You know, a lot of these arguments could be avoided if people would just answer the questions being asked, rather than trying to convince the original poster that he's wrong in his choice of codec (like the second post in this thread).

The OP asked about encoding DTS (or other alternatives besides AC3), he didn't want opinions on whether AC3 is better than DTS, has asked about encoding DTS. There is absolutely no reason to post a response with anything other than discussing ways of encoding DTS...

There is no study or research on the planet that's going to change most people's opinions of what audio codec they like better (look at the people that argue DTS HD MA sounds better than TrueHD or vice-versa... :rolleyes:). It's an OPINION, nothing more. If you can't answer the question being asked, don't post in the thread...

So far, the only response that should have been posted here is tebasuna51's

[shon3i]

He asked what we doing if we want less filesizes than lossles/lossy sources (PCM, TrueHD, DTS-MA, DTS-Hi Res) and want better quality than AC3 640kbps, and asked is there other options. So DTS 1.5 is not that, and is not better AC3 640kbps, and we all answer same "AC3 @ 640kbps". So it's basicly clear situation.

[Blue_MiSfit]

Indeed. IMO, there is absolutely no place for transcoding to DTS for BluRay backups. If there's a DTS track, fine. If not, there's no reason to create one!

A properly encoded AC3 at 640kbps will be quite transparent. In fact 640kbps for transparency is quite pathetic. AAC can do much better ;)

I usually do AAC at Q=.5, which creates 5.1ch files at ~ 300-400kbps. I have an E-MU 0202 USB DAC and Sennheiser HD-580 cans, and can tell NO difference between this encode and even a PCM source! In fact, I usually can't tell even lower than this, but I just do q=.5 since it's the eac3to default, and I like to leave in plenty of margin for error!

DTS is very old, and it's just plain bad at compressing audio. If you need 1.5mbps to make 5.1ch sound transparent ... well... that's just not up to par with the better standards available!

Don't get me wrong, it was a fantastic format when it came out. The idea of putting perceptually lossless 5.1ch audio at 16/48 PCM bitrates was quite revolutionary! But, these days... meh :)

~MiSfit

[kolak]

AC3 even at 640kbs is quite far away (based on mathematical comparison) from DTS at 1.5Mbit (quite close to 754kbs DTS). I've read a lot, but finally made a test by myself and don't need anyone to convince me that it's different. Audibly they will sound the same for most of the movie sources, but not for good classical music recording.

DD and DTS have almost the same compression efficiency

DTS is better because of 1.5Mbit mode (Dolby stops at 640kbs).

DD is very often good enough for movies, but for music titles DTS is better and with good equipment, there is quite audible difference between DD 448kbs and 1.5Mbit DTS.
As you said - you need very good ears to hear difference between 1.5Mbit DTS and PCM.

If your Blu-ray has DTS-MA- extract core and use this for backup. If it has TrueHD, extract core and use this for backup- there is no really reason for encoding to DTS unless this is music/concert Blu-ray.

Andrew

[kolak]

Don't bother with stupid inefficient dts! AC3 reaches transparency at 640 kbps . DTS requires 1,5 Mbps!


Both DTS and DD have almost the same efficiency, which is much worse compared to new formats (but none of them is a standard yet).

If you think that DD at 640kbs reaches transparency, DTS will do the same.
Don't read people opinion- just do test by yourself.

Andrew

[kolak]

1) First of all, there are some test than prove the AC3 640 Kb/s is more transparent than DTS 1.5 Mb/s. Only with a very good audio equipment, equipped room and good ears you can distinguish from the source uncompressed audio.

2) Don't exist free DTS encoders, the free AC3 encoder (Aften) isn't Dolby certified (certified implies pay) and don't know test to prove than Aften offers the same quality than certified ones.



Where are these tests? If someone claims this than he is based on marketing/placebo effect.

There is no free DD encoder on the market as there is no DTS one!

There is no such a thing like non certified DD encoder- if there is one, Dolby can suit this company, because you can't name your encoder Dolby Digital when it didn't pass and have Dolby certification. Dolby requires certification for all of their products.


Andrew

[tebasuna51]

AC3 even at 640kbs is quite far away (based on mathematical comparison) from DTS at 1.5Mbit
Isn't a valid method to compare.
DD and DTS have almost the same compression efficiency
Same efficiency? False, point.
If someone claims this than he is based on marketing/placebo effect.
Just the inverse.
There is no free DD encoder on the market as there is no DTS one!
I never say DD encoder, I say AC3 encoder (based in A52 specs).

BTW, everyone is free to select the desired format.
Remember the rule 12 in this forum:
"Do not ask "what's best" because this question cannot be answered objectively. Each and everyone has their own view about what's best in a certain area. The best is what works best for you!"

Please, stop this old discussion many times repeated.

[kolak]

Isn't a valid method to compare.

Same efficiency? False, point.

Just the inverse.

I never say DD encoder, I say AC3 encoder (based in A52 specs).

BTW, everyone is free to select the desired format.
Remember the rule 12 in this forum:
"Do not ask "what's best" because this question cannot be answered objectively. Each and everyone has their own view about what's best in a certain area. The best is what works best for you!"

Please, stop this old discussion many times repeated.


Where is link/reference to "these" testes?

Why mathematical comparision is not valid- the more encoded file is similar to original the better- that's all. Even if you count human nature of sound perception it still valid. All encoders are based on math :)

Other way- DD at 640kbs will sound almost the same as DTS at 754kbs. DD at 1.5mbit would sound probably the same as DTS at 1.5Mbit, but there is no such a mode for DD.


Aften is not Dolby certified, so there is no guarantee that it will work with hardware players, so it' definitely not good for archiving movies. I've look at the thread about it and there seams to be many issue with it, including nos as good as certified DD encoders quality.

Do you see word "best" in my posts? I havent't started this discussion it was actually YOU. I'm just trying to show that you're spreading spam. I can prove by maths that DTS and DD have similar efficiency- you make statement based on nothing.


Andrew

[Midzuki]

Audibly they (AC3 and DTS) will sound the same for most of the movie sources, but not for good classical music recording.

Serious audiophiles would say, NO type of lossy compression is suitable for music, in general, and for classical music, in particular. :)

[nurbs]

Where is link/reference to "these" testes?
Other way- DD at 640kbs will sound almost the same as DTS at 754kbs. DD at 1.5mbit would sound probably the same as DTS at 1.5Mbit, but there is no such a mode for DD.
http://www.ebu.ch/CMSimages/en/tec_doc_t3324-2007_tcm6-53801.pdf
DTS seems to run into noticable problems when you encode at 448 kbps so I don't think you can claim it has the same compression efficiency as DD at all bitrates.
From the link:
Dolby Digital requires 448 kbit/s and DTS still requires around 1.5 Mbit/s for "Excellent" quality.

Why mathematical comparision is not valid- the more encoded file is similar to original the better- that's all. Even if you count human nature of sound perception it still valid. All encoders are based on math
The problem with audio and video comparisons based on metrics is that you are guaranteed to run into situations where the encode with the lower score or the worse looking frequency plot actually looks or sounds better due to a better psy model.

edit:
Didn't see that shon3i already posted the link :o

[kolak]

Serious audiophiles would say, NO type of lossy compression is suitable for music, in general, and for classical music, in particular. :)

Serious audiophile use speaker cables worth fortune, but when one of them made a test and connected speakers with clothes hanger none of them was able to tell a difference :) They are a bit maniac :P
Going off topic :)

Andrew

[kolak]

http://www.ebu.ch/CMSimages/en/tec_doc_t3324-2007_tcm6-53801.pdf
From the link:
Dolby Digital requires 448 kbit/s and DTS still requires around 1.5 Mbit/s for "Excellent" quality.

The problem with audio and video comparisons based on metrics is that you are guaranteed to run into situations where the encode with the lower score or the worse looking frequency plot actually looks or sounds better due to a better psy model.

Yes, but psy in audio are quite well known and all encoders use them in the same main way. And NO- if you frequency response is far away form original than it's not going to sound good. Psy can't be over done- there is always a point when it doesn't work anymore- just make source sonud/look artificial.

First of all DTS was designed to give very high transparency and it definitely does.

Look at the graph carefully- DTS requires 448Kbs to be in excellent category- don't just read conclusion made buy someone, who didn't think :)
It also proves that 1.5Mbit DTS is "very" transparent together with all new codes, which can have much lower bitrate.
Look at DTS 448kbs, DD 448kbs (why is it in phase 2?) and DD 384kbs- as I said they are almost the same- DD may be slightly more efficient and have better optimization for low bitrates.
1.5Mbit DTS is quite overkill for an average consumer, that's why it's used mostly for music titles for people with good equipment.


Andrew

[Atak_Snajpera]

Aften is not Dolby certified, so there is no guarantee that it will work with hardware players, so it' definitely not good for archiving movies.
Have you ever seen hardware decoder which had problems with files encoded by aften ??? Dolby certified ? Who cares. If you have to much money so buy certified encoder.

[kolak]

Have you ever seen hardware decoder which had problems with files encoded by aften ??? Dolby certified ? Who cares. If you have to much money so buy certified encoder.

I don't use it so I don't know, but there will be few times more than with certified encoders.
How many hardware players did you test: 1 or 2?

Andrew

[nurbs]

Yes, but psy in audio are quite well known and all encoders use them in the same main way. And NO- if you frequency response is far away form original than it's not going to sound good. Psy can't be over done- there is always a point when it doesn't work anymore- just make source sonud/look artificial.

Ever seen a frequency plot for HE-AAC? Doesn't sound as bad as it looks. And then there is Blade, the MP3 encoder without psy-model that looks good but sounds bad.
This Head-Fi thread (http://www.head-fi.org/forums/f15/lossy-audio-codecs-comparison-huge-amount-pics-itunes-update-p-7-a-225356/) is probably a good example on how to post pretty pictures without saying anything about quality at all.


By the way:
Aften is not Dolby certified, so there is no guarantee that it will work with hardware players
LAME is not an MP3 encoder, so there is no guarantee that it will work with hardware players :devil:

BD-Rebuilder uses aften and I haven't noticed any complaints about the audio not working. The AC3 spec (http://www.atsc.org/standards/a_52b.pdf), like most video and audio codecs, only specifies the decoding process, and aften is widely used with different hardware, so I think you are worrying too much about compatibility.

Look at the graph carefully- DTS requires 448Kbs to be in excellent category- don't just read conclusion made buy someone, who didn't think
The requirement for landing in the excellent category is that the "mean value of worst case" doesn't fall under 80. DTS at 448 kbps goes down to 55. Maybe you shouldn't be so quick to assert that the authors of a study aren't thinking.

[kolak]

The requirement for landing in the excellent category is that the worst sample doesn't fall under 80. DTS at 448 kbps goes down to 55. Maybe you shouldn't be so quick to assert that the authors of a study aren't thinking.

Sorry- yes, but did they test 754kbs DTS mode (or other)? Author said that DTS needs 1.5Mbit to be in excellent group- really (maybe 754kbs)?
I bet you DD 448kbs in the Phase1 test files would also fall lower- some files in phase 1 had to be difficult to encode, otherwise DD at 384kbs should do much better.
Anyway- it's good document- slightly "not complete" (the most interesting graph missing) and done quite long time ago with some special/beta version of the codecs- this should not be allowed :)
It shows how modern codecs outperform old ones- which is the way how it should be :)

Andrew

[nurbs]

Sorry- yes, but did they test 754kbs DTS mode (or others)?No.
Author said that DTS needs 1.5Mbit to be in excellent group- really (maybe 754kbs)?You can't make conclusions for bitrates between the tested ones, so in the context of the paper he's correct.
I bet you DD 448kbs in the Phase1 test files would also fall lower- some files in phase 1 had to be difficult to encode, otherwise DD at 384kbs should do much better.I think you'd lose that bet.

The samples of the first phase would have to be on average harder to encode than the samples of the second phase for that to be true. Since there were ten samples in both phases and the samples had similar content (see the descriptions) I think that is doubtful. For the organ (Bach) and the harpsichord there is the same description in phase one and two. DD improved from sub 60 to over 80 points on those samples by raising the bitrate from 384 to 448. Those are the samples where DTS scored lowest at 448.

The overall worst sample, "Applause", was used in both phases and using that and the codecs used in both phases the authors think they can show that the results are comparable.
Anyway- it's good document- slightly "not complete" (the most interesting graph missing) and done quite long time ago with some special/beta version of the codecs- this should not be allowed Which graph is missing and to you really think there have been huge strides in the development of codecs that have been around as long as DD and DTS in the past three years? Personally I can't see what's wrong with their codec selection criteria.

[shon3i]

How many hardware players did you test: 1 or 2?You some kind of funny guy? Realy i can't understand you sense of humor.

otherwise by your story, speaking MPEG2 is much better than H264 in video area, "they are have same compression efficiency", but MPEG2 should be four times higher bitrate to achieve transparency.

[kolak]

No.
You can't make conclusions for bitrates between the tested ones, so in the context of the paper he's correct.
I think you'd lose that bet.

True- they should not make such a conclusion.
They should test DTS 754kbs.



The samples of the first phase would have to be on average harder to encode than the samples of the second phase for that to be true. Since there were ten samples in both phases and the samples had similar content (see the descriptions) I think that is doubtful. For the organ (Bach) and the harpsichord there is the same description in phase one and two. DD improved from sub 60 to over 80 points on those samples by raising the bitrate from 384 to 448. Those are the samples where DTS scored lowest at 448.

Here is the point where I can argue.
The same nature of file, similar content can be relatively easy or hard to encode- it depends how it was recorded. It's enough to have some bits of noise and it will make 2 files totaly different. I don't understand why they have 2 phases with different samples ("... due to one of the codecs bug"- makes no sense)- it's comparing apples to oranges (I assume clips where chosen carefully, but why did they need different clips for 2nd phase?)


The overall worst sample, "Applause", was used in both phases and using that and the codecs used in both phases the authors think they can show that the results are comparable.
Which graph is missing and to you really think there have been huge strides in the development of codecs that have been around as long as DD and DTS in the past three years? Personally I can't see what's wrong with their codec selection criteria.

Yes- but it was exactly the same clip - not different recording of applause- so they have to be comparable.
Applause was only one the same clip. Both 384kbs and 448kbs DD are doing well with it. DD 384kbs is poor for rock, which is not included in 2nd phase at all.

I'm not talking about DTS or DD, but new codecs (some beta/special builds for test).
There is nothing really wrong except some missing bitrates and mixing very high bitrates with very low in phase 1. Why DD 448kbs is in phase2, but DD 384kbs in in phase 1? There is no direct comparision between some of them- I thought that was the point.
I would like to see DD 448kbs in phase 1 and DTS 754kbs, which is the most common DTS mode.
These 2 phases were pointless- if you've chosen very representative 10 clips that's all- you compare all codecs on them. If they would compare all codec on phase1 clips than on phase 2 clips than they could show (if results would be comparable) that clips on both phases are representative and it's proper test.

I would like to see more these kind of tests- they're interesting:)

Andrew

[kolak]

You some kind of funny guy? Realy i can't understand you sense of humor.

otherwise by your story, speaking MPEG2 is much better than H264 in video area, "they are have same compression efficiency", but MPEG2 should be four times higher bitrate to achieve transparency.

Looking at DTS 448kbs and DD 384kbs used on the same test clips, for me they look to have about the same efficiency.

Author's summary about DTS 1.5Mbit and DD 448kbs is not very professional (at least there should be some comment with it that DTS has also different modes, but they were untested).

Andrew