Axiom Home Page Forums General Discussion Home Theater Digital Audio Format Discussion

While trying to understand the benefits of high resolution (or definition) audio found on HD-DVD and BR-ROM disks, I’ve been drafting a note to myself for reference and edit as I pick up more information and understanding. This is how I learn; I write technical papers to myself for reference. Not only with A/V, but mechanical practices, photography, and any other ‘thing’ I happen to have an interest in.

Considering that, I thought it might be both helpful and also educational for this on-line community to see what I’ve drafted to date, and post it for further edit and conversation because I know that many, if not most of you want to learn about this as well.

Please feel free to cut up and edit what I post, and please add your input with the ultimate goal of this thread being a handy reference for any who may be interested in this topic.

……….

Digital audio sampling:

Digital audio is, in a very generic description, a digital copy of analogue sound.

To reproduce sound in a digital environment, samples of that analogue sound are taken, recorded and stored as bits of electronic information that is converted back to analogue by an audio processor’s digital-to-analogue-converter (DAC). Currently, there is one widely recognized format and a second format recognized to a lesser degree. They are very different, with PCM (Pulse-Code Modulation) being the primary format and DSD (Direct Stream Digital) being the other. Both of these formats have some similarities in the method in which ‘samples’ are taken, but at the same time, very different. To understand each, sampling needs to be discussed.

A sample is a measurement of sound taken by a microphone. The sample has two aspects. One being the sampling “rate” and the other being the “resolution” of that rate. (I am having difficulty understanding this, but will muddle on anyway)

The “rate” is the number of times within a second that the sample is taken and it is expressed in kHz (times per second). For reference, CD PCM audio is TYPICALLY sampled at 44,100 times per second (44.1 kHz).

The “resolution” is the depth, or dynamic range of the sampling rate and is expressed in ‘bits’, or ‘bit’. This depth is a sampling of the dynamic Hz range sampled (db’s). For reference, CD bit resolution sampling is 16 bit which represents a dynamic range of 98.09 dB. 8 bit = 49.3 db, 20 bit = 122.17 db, 24 bit has useless random values, so no one really knows what this dynamic range is.

Sampling ‘rate’ and ‘resolution’ are used to calculate the total amount of electronic data in an audio stream. This is done by multiplying the rate by the resolution by the number of channels recorded. IE: Two channel (PCM) CD = 44.1k*16 = 705.6k bits per second *2 (channels) = 1.4Mb/s

The LFE channel also contains some information, but not quite as much (and I do not know how much that is either).

Now when you make calculations for a 5.1 mix, you can start to see just how much information is actually there. This helps to understand why there is a need for data compression. 44.1k*16*5 = 3.528 Mp/s * 60 seconds * 120 minutes (or thereabouts) = 25,401,600,000 bits + an unknown number of LFE bits (25+ gig)

To continue with PCM sampling in mind, and in particular, the frequency of those samples, there are additional sampling rates available including: 8 bit, 20 bit and 24 bit. It is arguable if anything about 16 is audible to a human being, but 8 bit introduces a higher noise level than 16 bit and the difference in sound quality between 8 and 16 bit is very noticeable. Additionally, there are many recordings taken at a higher sampling rate, up to 192 kHz/s.

Where SACD differs from CD, is that SACD uses a different sampling process (or standard) referred to earlier; DSD. DSD is somewhat difficult to understand and I most definitely do not have a good grasp on it. DSD samples at a much higher rate than PCM, but only at a frequency rate of 1 bit. The DSD sample rate is 2.8224 Mhz (64 times greater than the PCM/CD sample rate).

Obviously, the DSD sample frequency of 1 bit verses 16 bit leads one to assume that something is amiss. From my limited understanding of this oddity, DSD uses a proprietary noise shaping quantization technique to increase the dynamic Hz range of that single bit frequency to an inaudible ultrasonic level with a dynamic range of 120 dB. ( more than 20 db greater than the CD dynamic Hz range)

Compression:

As hinted to earlier, depending on the sampling technique used to capture an audio track for film (or any A/V media for that matter), there is a lot of audio data to put on a disk, and until recently with the introduction of HD-DVD and BR-Rom, impossible to store in its un-compressed state along with the video data. It is for that rational that compression enters the conversation. Dolby Digital and DTS are the current leaders in audio compression for film.

DD and DTS use proprietary algorithms to compress PCM audio files into a bit stream format. In a nutshell, this means that each one throws away what they have determined, useless data. After compression, this data is stored on the storage media (DVD) as a codec that is later decoded and reconstructed to as close to its original state as possible. Codec is simply a name for the compression method used (WMA, MP3, DD, DTS, DD+, DTS++, DTS-HD, DTS-HD:M, DD True HD). When the disk is played, the media player’s decoder looks for “flags” which tell it what codec the audio stream is stored as and then continues to decode and reconstruct the file to re-create the original PCM digital audio file. (think Zip file)

Here’s where things start to get confusing (if not already).

DD and DTS have a compression ratio of about 10 to one.

DVD-A uses a lossless compression techniques called MLP (Meridian Lossless Packing) which is a proprietary technique developed by Meridian Audio Ltd. The MLP compression ratio is 2 to 1 and when the decoder unpacks the file, it is restored to its original PCM state undamaged, whereas DD and DTS do their best to reconstruct the original PCM file, but not without loosing data in the process.

Dolby TrueHD also uses MLP as a compression format, but it adds additional information over DVD-A such as higher bit rate and resolution, metadata, additional channels and some other things like user menu functions and interactive audio files found on some HD and BR disks.

DTS-HD-M (master) uses their own (DTS) proprietary lossless compression codec. DTS-HD-M also contains a ‘core’ lossy compressed audio codec for players that can not decode the Master lossless DTS codec. The lossy version is a compressed file that is limited to 1.5 Mb/s.

DD TrueHD is a mandatory codec for HD DVD players to be able to decode, and is an optional codec for BR-Rom players to decode. For both formats, the limitations / capabilities of True HD is the same, which is up to 8 discrete channels of 24bit / 96 kHz with a max bit rate of 18 Mbit/s. All players can down-mix the audio track to just about any number of channels or bit rate and transmit that data stream in a bit stream (compressed format) for further decoding in the audio processor (depending on the processors capabilities). Or, TrueHD can be decoded and output from the player as LPCM up to a bit rate of 35 Mbit/s via HDMI 1.1. HDMI 1.3 is required to transmit the lossless compressed TrueHD bitstream audio file.

DTS-HD Master Audio is an optional codec for both HD-DVD and BR-Rom players. It has also been called DTS++ and DTS HD. DTS-HDMaster has unlimited discrete channel capabilities of up to 24 bit / 96 kHz each. All BR and HD players can decode an embedded “core” DTS soundtrack at 1.5 Mbit/s. Not all BR or HD players can decode the lossless codec into a LPCM sound file or transmit the lossless bitstream codec. HDMI 1.3 is required to transmit the lossless bitstream codec, but HDMI 1.1 will carry the unpacked LPCM sound file.

So what does all this mean?

Well, in summery, if you have a receiver that has HDMI 1.1 and it will process HDMI audio (not just switch HDMI audio), that’s really all you need to hear lossless sound or SACD. With the exception of DTS-HDMaster, all BlueRay and HD players will decode all codecs and output them as LPCM. (lossless PCM)

If you wish to hear SACD / DSD, you will need HDMI 1.2 or higher, but you will also need an SACD player that recognizes DSD and you will also need an audio processor that recognized DSD as well. Otherwise, the SACD player will be limited to its on board decoder which will decode the DSD data stream and output that via HDMI as LPCM. Or, if you use the multi channel analogue outputs, the SACD player will decode the DSD data stream and convert that digital signal to an analogue signal for the audio processor to amplify.

If you have an audio processor with HDMI 1.3, and it has a decoder that recognizes DD TrueHD and DTS-HDMaster, it will decode every surround format that the player you have recognizes and outputs.

If you do not have an audio processor that does not have any HDMI audio processing, you will have to rely on the player’s decoder to unpack the core DTS codec, or down mix DDTrueHD to a compressed DD audio stream and output those via the digital outputs. Or, you will have to use the analogue outputs, if, the player is equipped with them and recognized the codecs.

As side note, it is debatable to whether the player or the audio processor will do a better job at decoding the codecs. Some will argue that the audio processor has the edge, but I seriously doubt that that claim can be qualified without a true blind listening session.

Damn, I'm going to end up having a book just on Mike's material!

Thanks for yet another very informational post Mike, I'm going to end up learning something about this HD stuff yet.

Yes, thanks for the post. It's nice to have so much encapsulated in one place.

Basically what it is saying is that I get to upgrade my 1804 Denon for a new AVR that has HDMI 1.3 when I finally get around to buying a TV that will support HD and a HD player.

Once again thanks for sharing.

Thanks Mike. Great Stuff.

You have a talent there. You should consider a web page of your own for tech. articles or a part time position writing for one.

Thanks guys…my profession requires me to write tech papers that engineers and technicians can understand…

I’ll working on a summery that lists all the different media audio resolutions. I’ll add that later.

Tech writer?

No, I just work with engineering and technical (craft) work forces. I get to figure out how to make engineering solutions and design packages actually work.

It would be a good idea for you to post these up on a website somewhere(that is if you want to share them). I know Google now has free web hosting and it's very easy to set up pages.

Excellent and very informative post!

[quote=pgans]I know Google now has free web hosting and it's very easy to set up pages.
[quote]

Thanks for the info yourself. I hadn't noticed they are hosting now.