Dolby TrueHD Bitrate Question

[ClaytonMG]

Alright, so upon looking at the bitrate for the Dolby TrueHD track on 300, I noticed it actually drops down to around 1.2mbps in some spots... Which is actually lower than a DTS track and a Dolby Digital-Plus track. How can it do that if it's supposed to be lossless? I'm not saying that it isn't lossless, I just find it interesting and at the same time, it's making me want to stick with PCM's constant 4.6mbps bitrate.

[Helicon]

Some scenes may not require the bit bitrate to be higher than 1200kbps.

With PCM 16-bit 48kHz all channels runs at a constant bit-rate of 768kb, which gives you a total of 4608kbps.

But imagine a scene with dialog only, no other speaker than the center channel has to play. Then you can cut down the bitrate for the other channels.

[Sir Terrence]

Originally Posted by ClaytonMG

Alright, so upon looking at the bitrate for the Dolby TrueHD track on 300, I noticed it actually drops down to around 1.2mbps in some spots... Which is actually lower than a DTS track and a Dolby Digital-Plus track. How can it do that if it's supposed to be lossless? I'm not saying that it isn't lossless, I just find it interesting and at the same time, it's making me want to stick with PCM's constant 4.6mbps bitrate.

I assure you it is lossless. It is variable bit rate lossless, in that it only allocated the bits when called on to do so. This is part of what makes both Dts HD MA and Dolby TrueHD more efficient at packing data on the disc.

PCM is constant bit rate in that no matter how many bits are actually needed to code the audio, there will always be more than necessary at any given time, and enough available when needed. Not the most efficient way, but I prefer it because it is widely used, and there are no additional processes to get it to playback other than D/A converters.

[WickyWoo]

The moral of the story is "Some parts squish better than others"

[dialog_gvf]

Originally Posted by ClaytonMG

Alright, so upon looking at the bitrate for the Dolby TrueHD track on 300, I noticed it actually drops down to around 1.2mbps in some spots... Which is actually lower than a DTS track and a Dolby Digital-Plus track. How can it do that if it's supposed to be lossless? I'm not saying that it isn't lossless, I just find it interesting and at the same time, it's making me want to stick with PCM's constant 4.6mbps bitrate.

DTS is Constant Bit-rate (CBR). It is always (for example) 1.5Mbps.

It doesn't always have to be. When it is absolutely silent, you could get by with something approaching 0Mbps. Yet, DTS is still using those 1.5Mbps (and LPCM 4.6Mbps).

One way to look at it is that DTS is lossless whenever the TrueHD is below 1.5Mbps. And gets progressively lossy for sound causing TrueHD to go above 1.5Mbps.

Gary

[ClaytonMG]

Originally Posted by dialog_gvf

One way to look at it is that DTS is lossless whenever the TrueHD is below 1.5Mbps. And gets progressively lossy for sound causing TrueHD to go above 1.5Mbps.

Gary

Yeah that's kind of how I've been looking at it, which is why I'll select a PCM track over Dolby TrueHD. It seems there's less chance of an error. Does anyone know if DTS-HDMA works the exact same way?

[ClaytonMG]

Actually... now that I think about it, what is the birt rate on a DTS-HD High Resolution track? Since Dolby TrueHD only stays around 2.0mbps (it stays below 2.0 a lot of the time), I am curious to know if DTS-HD High Resolution is more lossless than Dolby TrueHD.

[WickyWoo]

Actually... now that I think about it, what is the birt rate on a DTS-HD High Resolution track? Since Dolby TrueHD only stays around 2.0mbps (it stays below 2.0 a lot of the time), I am curious to know if DTS-HD High Resolution is more lossless than Dolby TrueHD.

You can't be "more lossless". Either you're lossless or you're lossy. DTS-HD is lossy, DTS-HDMA-lossless. TrueHD-lossless PCM-lossless. Dolby Digital-lossy, DTS, lossy

[ClaytonMG]

But if Dolby TrueHD dips down past the bitrate of a DTS track, then the DTS becomes "more lossless" then the TrueHD track. And my other question still stands, since the bitrate of a DTS-HD High Res track is probably higher than the bitrate of a Dolby TrueHD track, wouldn't the high res track be closer to lossless than the TrueHD one?

[Gary_DolbyGuy]

So lets be clear about lossless codecs. In the Dolby TrueHD case-the bit rate will vary depending on the content being encoded losslessly. It will attain the lowest bit rate it is capable of and still maintain lossless quality. The Dolby & DTS codecs are different and therefore can achieve lossless results with different bit rates.

[ZoetMB]

Originally Posted by ClaytonMG

But if Dolby TrueHD dips down past the bitrate of a DTS track, then the DTS becomes "more lossless" then the TrueHD track. And my other question still stands, since the bitrate of a DTS-HD High Res track is probably higher than the bitrate of a Dolby TrueHD track, wouldn't the high res track be closer to lossless than the TrueHD one?

You can't be "more lossless". That's like being "a little bit pregnant". Lossless is lossless. Lossless means that the resulting file, when uncompressed, is a bit-for-bit copy of the original. They're just getting better at figuring out algorithms to encode the data to smaller file sizes and still wind up with a lossless file.

[slimdude]

It could be the fact that Dolby uses Dialogue Normalization, (a feature that other codecs don't utilizes) which may effect its overall bitrat,e causing variable readings. The DN signal decreases the dynamic range of the soundtrack.

[crackinhedz]

Originally Posted by slimdude

It could be the fact that Dolby uses Dialogue Normalization, which may effect its overall bitrate. The DN signal decreases the dynamic range of the soundtrack.

dialnorm does not affect dynamic range. This is a common misconception.

Sticky: Understanding Dialog Normalization

[slimdude]

Originally Posted by crackinhedz

dialnorm does not affect dynamic range. This is a common misconception.

Sticky: Understanding Dialog Normalization

Well I prefer soundtracks without DN applied! I don't know if its's just me but Dolby always sounded a bit softer, and less robust than DTS, and PCM 5.1.

[Anthony P]

Originally Posted by ZoetMB

You can't be "more lossless". That's like being "a little bit pregnant". Lossless is lossless. Lossless means that the resulting file, when uncompressed, is a bit-for-bit copy of the original. They're just getting better at figuring out algorithms to encode the data to smaller file sizes and still wind up with a lossless file.

you would think that, and as related to Clayton's post it is right (bitrate does not matter if truly lossless), but there is more to it. Lossless referes to the original (LPCM) digital file. On the other hand no digital file is lossless compared to the original analogue or soundwave which in the end is what you want to have perfect. For example a CD is 16bit 44kHz uncompressed and so a 16bit 44kHz DTS-MA/DTHD of the soundtrack would be just as good, on the other hand if it was recorded as 24bit 192kHz it would be much better. That is why some of us would like to see BD-audio, and some bought into SACD & DVD-A even though CD was uncompressed (and so lossless compared to uncompressed) there was still room for improvement.

[singhcr]

Originally Posted by Anthony P

you would think that, and as related to Clayton's post it is right (bitrate does not matter if truly lossless), but there is more to it. Lossless referes to the original (LPCM) digital file. On the other hand no digital file is lossless compared to the original analogue or soundwave which in the end is what you want to have perfect. For example a CD is 16bit 44kHz uncompressed and so a 16bit 44kHz DTS-MA/DTHD of the soundtrack would be just as good, on the other hand if it was recorded as 24bit 192kHz it would be much better. That is why some of us would like to see BD-audio, and some bought into SACD & DVD-A even though CD was uncompressed (and so lossless compared to uncompressed) there was still room for improvement.

I totally agree!

You can record audio at 128kbps MP3 standards and store it as DTS Master Audio. It is lossless because it is not losing anything from the master copy, but the master itself could be better.

That is why the BD of Akira is so great. The original master analog tapes are of extremely high quality, far, far beyond CD quality. According to the production booklet, the 192 KHz, 24 bit 5.1 track on the BD is very close to the original tapes. I believe most movies these days are recorded at 96 Khz 24-bit (i.e. 96/24) but are presented on BD at 44.1/16, which is CD quality. Don't get me wrong, CD quality is pretty good, but we had that in the laserdisc days. If BD is capable of multichannel 192/24 audio, why not use it? At the least, increase the sample rate from 16 to 24 bits.

[ZoetMB]

Originally Posted by Anthony P

On the other hand no digital file is lossless compared to the original analogue or soundwave which in the end is what you want to have perfect. For example a CD is 16bit 44kHz uncompressed and so a 16bit 44kHz DTS-MA/DTHD of the soundtrack would be just as good, on the other hand if it was recorded as 24bit 192kHz it would be much better.

That's not how the term lossless is generally used and accepted in the industry - it means what I originally said - whether the output file is a bit for bit copy of the input file.

And yes, theoretically, 96/24 or 192/24 should sound better than 44.1/16, although I have a CD recorder that can optionally record at 96/24 (which is one of the reasons I bought it) but whenever I've recorded anything at 96/24 and compared it to 44.1/16, I could not perceive any difference whatsoever and I would defy anyone else to. That's not to say that I think pro recording shouldn't bother with 96/24 or higher.

Furthermore, I would defy anyone to pass a blind A/B test between an (analog) LP playing back and the CD-R of that LP, even at 44.1/16, regardless of whether the LP was made from an analog recording/master or a digital recording/master. I've conducted that test many times in my home and no one has ever gotten it right better than 50% of the time (which you can achieve by randomly guessing.)

[WiWavelength]

Originally Posted by singhcr

You can record audio at 128kbps MP3 standards and store it as DTS Master Audio.

The first half of your point is inaccurate. You can store "audio at 128kbps MP3 standards," just as you can "store it as DTS Master Audio." But digital audio is not recorded as MP3; it is recorded as LPCM. (More accurately, most modern A/D converters utilize very high sample rate, low bit sigma delta conversion that is immediately decimated to LPCM of desired bit depth and sample rate.) Recorded LPCM can then be converted to lossy/lossless compression formats for storage/transmission.

Originally Posted by singhcr

I believe most movies these days are recorded at 96 Khz 24-bit (i.e. 96/24)...

Nope. The current output standard for motion picture audio is typically 24 bit 48 kHz. (Internally, while audio remains within mixers' work stations, the format may be 32-64 bit floating point at 48 kHz sample rate.)

Originally Posted by singhcr

...but are presented on BD at 44.1/16, which is CD quality.

Again, nope. You will find few, if any 16 bit 44.1 kHz tracks on BD. Nearly all are 16 bit 48 kHz or 24 bit 48 kHz.

Originally Posted by singhcr

If BD is capable of multichannel 192/24 audio, why not use it?

Very few sources are recorded/transferred at 192 kHz -- certainly not the "sausage filler" of library and location audio sources that make up much of motion picture audio. Those few sources that are recorded at 192 kHz tend to be classical music recordings.

Originally Posted by singhcr

At the least, increase the sample rate from 16 to 24 bits.

Many BDs already utilize 24 bit tracks. See above.

AJ

[Anthony P]

Originally Posted by ZoetMB

That's not how the term lossless is generally used and accepted in the industry - it means what I originally said - whether the output file is a bit for bit copy of the input file.

did I state otherwise? That is why I started with Lossless refers to the original (LPCM) digital file that was used. But as someone that wants to listen to something (let's say a song at a concert to keep it simple) I care about how it compares to the original sound (being at the concert) and not what the digital recording of that sound sounds like. Let's say I take an CD (to keep it simple), use DD, DTS, MP3 (not important)... and compress the hell out of it. Then I take that compressed file and turn it back to LPCM (let's call the original CD MA and this one MB). Now let's say one person used a lossless compression DTHD DTS-MA) on MA to make CA and the other used MB which we would call CB. Will CA be a lossless compression? How about CB? obviously both are lossless since it was stated in the problem. Will CA and CB sound the same and will they have the same values once decompressed? No because CB was created from MB that was at some point overly compressed. Would it make sense to describe CA as "more lossless"? yes because in the end CB represents MB but MB should represent MA and so ultimately CB should be representing MA and CA is much better at that.

And yes, theoretically, 96/24 or 192/24 should sound better than 44.1/16, although I have a CD recorder that can optionally record at 96/24 (which is one of the reasons I bought it) but whenever I've recorded anything at 96/24 and compared it to 44.1/16, I could not perceive any difference whatsoever and I would defy anyone else to. That's not to say that I think pro recording shouldn't bother with 96/24 or higher.

Furthermore, I would defy anyone to pass a blind A/B test between an (analog) LP playing back and the CD-R of that LP, even at 44.1/16, regardless of whether the LP was made from an analog recording/master or a digital recording/master. I've conducted that test many times in my home and no one has ever gotten it right better than 50% of the time (which you can achieve by randomly guessing.)

Have you ever though there is something wrong with your set-up or wrong with your ears or maybe it is influenced by you not caring in the first place so you just dismiss the difference which is measurably there and others can hear going through a double blind A/B test?

[WiWavelength]

Originally Posted by Anthony P

did I state otherwise? That is why I started with Lossless refers to the original (LPCM) digital file that was used. But as someone that wants to listen to something (let's say a song at a concert to keep it simple) I care about how it compares to the original sound (being at the concert) and not what the digital recording of that sound sounds like.

It is almost pointless to speak of a recording -- analog or digital -- as "lossless" to the original event. Short of a theoretically perfect binaural recording, any other recording will "lose" some of the spatial/directional characteristics of a reverberant sound field. Hence, "lossless," in this context, correctly refers only to a digital compression stage applied to a digital recording or transfer.

Originally Posted by Anthony P

Let's say I take an CD (to keep it simple), use DD, DTS, MP3 (not important)... and compress the hell out of it. Then I take that compressed file and turn it back to LPCM (let's call the original CD MA and this one MB). Now let's say one person used a lossless compression DTHD DTS-MA) on MA to make CA and the other used MB which we would call CB. Will CA be a lossless compression? How about CB? obviously both are lossless since it was stated in the problem. Will CA and CB sound the same and will they have the same values once decompressed? No because CB was created from MB that was at some point overly compressed. Would it make sense to describe CA as "more lossless"? yes because in the end CB represents MB but MB should represent MA and so ultimately CB should be representing MA and CA is much better at that.

Please edit this paragraph. Your writing is difficult to follow, your abbreviations indecipherable.

Originally Posted by Anthony P

Have you ever though there is something wrong with your set-up or wrong with your ears or maybe it is influenced by you not caring in the first place so you just dismiss the difference which is measurably there and others can hear going through a double blind A/B test?

If you are certain that greater bit depths and/or higher sample rates result in differences audible to you, have you personally performed and passed ABX tests?

AJ