Lossless audio formats / Digital audio in repects to LPCM, DTS-HD-MA, and Dolby True

[bnieman]

Hello,

As I continue my search to incorporate the new lossless surround sound codecs into my home theater, I stopped in at my dealer to get his advice and possibly test drive a couple of units.

We ran into a disagreement about the nature of lossless formats. It is my understanding that bitstreaming a DTS-HD Master Audio signal is exactly the same (bit for bit) as a DTS-HD Master Audio signal decoded by the player and sent via LPCM (multi-channel PCM) to the preamp/processor.

For example, the PS3 cannot bitstream the new formats but it can decode them within the player and then send them via LPCM to a receiver via HDMI. After much research, it seemed the consensus was this is every bit as good as bitstreaming the audio since you are going from one lossless format to another without analog processing.

On the other hand my dealer's understanding is if the player is decoding the signal on board and converting it to LPCM, that you are at the mercy of how accurate the on board decoder is.

Am I understanding this correctly? Is the signal exactly the same? Or can coloration be added when the audio is converted to LPCM inside the player prior to being sent to the preamp/processor?

Some might be asking how this relates to Rotel... well I have been considering the RSP-1069 because it can handle LPCM via HDMI whereas it cannot handle bitstreaming of the new formats.

Has anyone done blind A/B testing with bitstream vs. LPCM?

Cheers!

[tpirovol]

Hi,

Last week I did this exact comparison between the PS3 which I own via LPCM and a Sony 550. To me there is a huge difference with bitstream over LPCM but I have the 1560.

But thats my 2cents!

Thanks terry

[bnieman]

Terry,

Thanks for the response. Have you compared LPCM vs. Bitstream strictly out of the Sony 550?

Best Regards,
Bryan

[hifiguymi]

There are a lot of different things in play so it's not a simple answer as to which one is better. Theoretically it should be the same but in my experience it's not. I spent quite a bit of time comparing bitstreaming and player decoding with a Denon AVR-3808CI and a Denon DVD-3800BDCI because they have the same surround decoder. With every thing I tried the player doing the decoding sounded better. The soundstage was bigger, the bass had better definition, and the dialog/vocal was more intelligible. Some of the things that I think contributed to that is the player has Denon's Advanced AL24 upsampling circuit on all channels and the AVR-3808CI has AL24+ (it's a different DSP circuit and not as good) only on the front left and right channels. Along with that the player has a really good digital clock in it so it's jitter output is probably low so the multi channel PCM streams are in really good shape when they get to the receiver (or preamp).

When I put a Sony BDP-S500 in place of the DVD-3800BDCI the system sounded better bitstreaming than having the player do the decoding. Since bitsreaming isn't linear like a PCM signal it's not affected by jitter like a PCM signal is. The Sony doesn't apply any DSP (that I know of) to a decoded signal and I doubt it's jitter is as low as the Denon's. That, I'm assuming, is why the bitstreaming sounded better.

I didn't have both players available at the same time to compare them head to head with bitstreaming but I do now so I'll give it a try this weekend if I have time and report back. I don't have a PS3 so I can't do any comparisons with it.

Eric

Edit: The Sony player was a BDP-S500 not a BDP-S550 as I oringinally typed. I must of had the BDP-S550 on the brain since it was mentioned in an earlier post and it was late at night when I was typing.

[bnieman]

Thanks for all the information hifiguymi.

I am eagerly anticipating your head-to-head bitstreaming comparison between the Sony player and the Denon player (although theoretically it should yield the same results).

I can wrap my head around the PCM signal sounding better on the Denon since the signal passes through AL24 processing before being sent to your preamp/processor.

I do not understand why the Sony sounds better bitstreaming as opposed to LPCM.

Is it possible/likely the Sony player incorrectly decoded the bitstream? I would think it would be easy to convert a bitstream to PCM without any errors. And even if it did create errors, how could it have THAT much of an impact on the sound? I always thought digital errors resulted in pops and clicks... not a different timbre characteristics.

I remain baffled

[hifiguymi]

Since the Denon AVR-3808CI does do some DSP (like the AL24+) to a signal that is decoded internally I think that is why it sounds better bitstreamed with the Sony player. I don't know of any receiver or preamp that adds any post processing to a multi channel PCM input. That is why you can't add things like PLIIx to the rear channels if you have a 7.1 system and the disc has a 5.1 soundtrack. That only happens if the receiver or preamp is doing the decoding. It's the same with "upsampling" the digital PCM streams, it's not done. Since the Sony doesn't apply any DSP to the signal if it does the decoding it just goes straight to the DACs (after delay and bass management is applied).

By definition the decoders in both a given player and given receiver, or preamp, have to decode thing correctly. If they didn't, things would be all messed up. The precision of a decoder may or may not have an affect on the sound quality. The only way to know would be to have a the same preamp and swap out the decoder. If you have decoders in different pieces of hardware then there are other variables so how could you tell? Besides, you should know by now that digital is digital and it all sounds the same. :W

Trust your ears and you will find what works best for you.

Eric

[bnieman]

Interesting points Eric. I think I am starting to get a grasp on the underlying concepts.

Most preamp/processors process bitstreamed audio which includes processes such as upsampling. On the other hand, most preamp/processors to not add any extra processing to multi-channel PCM asside from converting them from digital to analog. This would then indicate that the LPCM version of the signal out of most ordinary Blu-Ray players that do not do any additional processing is the purest form of the digital signal.

When people say they prefer the bitstreaming signal over the LPCM signal this is most likely because they like the coloration that their preamp/processor adds to the signal.

I've always been a believer in digital is digital and it all sounds the same (until of course it goes through a digital to analog converter in which case it's no longer digital).

Regarding PCM post processing: My RSP-1068 will process a stereo PCM signal if the sample rate is 48kHz or below. I tested this with my computer by changing the sampling rate and at 48kHz Dolby PLII was enabled. At 96kHz I couldn't add any processing and came through as 96kHz PCM stereo.

Thanks again,
Bryan

[hifiguymi]

I suppose you could call it a coloration when a component does some form of DSP to an audio signal and converts a 16 bit signal to a 20 or 24 bit signal and a 44.1kHz sampling rate to an 88.2 or 176.4kHz sampling rate. If it sounds better than that is not a bad thing in my mind. The same goes for video scalers. The question is where is it done and how well it's done.

As far as digital being digital and it all sounds the same, when it comes to PCM that is not the case. Because things like jitter will adversely affect a linear signal not all things sound the same when using a digital connection. If you are talking about a signal like a Dolby or dts stream since they are not linear in transmission they don't suffer the same problems like jitter. That information is sent in packets. Try hooking up two digital sources, like a CD player and a DVD player, to your RSP-1068, make two copies of a CD (so you eliminate the original vs copy debate), use the same interface (either Toslink or coax digital), and compare the two. They won't sound the same. Depending on the components used the difference might not be huge, but there will be differences.

Eric

[bnieman]

OK, I am confused again.

What other things aside from jitter will affect the sound? And what does jitter do aside from add audible pops and clicks in the audio?

It is my understanding that jitter is a digital signal not being read properly resulting in bits not being read at the right time or not being read at all and perhaps even doubling up on a subsequent bits to catch up. The result being audible pops or clicks. I do not understand how jitter could result in the sound having a different overall "quality". When I say quality I am referring to the characteristics our ears distinguish when we listen to audio. Things like the presence, sound staging, how clean is, how fast is, how airy it sounds... you know all those non-scientific terms we use to describe what we hear.

What is happening in the digital realm to make it sound different from source to source?

I will take your advice and do some testing, but even if I conclude they sound different I still won't understand why. :roll:

[Legairre]

Hi,

Last week I did this exact comparison between the PS3 which I own via LPCM and a Sony 550. To me there is a huge difference with bitstream over LPCM but I have the 1560.

But thats my 2cents!

Thanks terry

You have to raise the PS3's internal volume to +2 to get it to the same level as as a receiver's(default is 0). I tried this in the past as well and thought the receiver sounded better bitstreamed. Once I raised the PS3 internal volume to +2 there was no difference at all between bitstream and LPCM. It's just a level matching issue and louder sounds better.

Give it a try. Just hit triangle on the PS3 controller and toggle over to volume and increase it to +2.

[tpirovol]

Hi just an update from me I compared bitstream from the Sony 550 and there was large difference. I then tried lpcm and there was a very small difference compared to the ps3.

Thanks Terry

[bnieman]

Thanks for the update Terry!

Given what we have discussed, at this time I would conclude the 1560 must do some post-processing on the bitstream whereas with the LPCM it just sent it to the digital to analog converter.

[hifiguymi]

OK, I am confused again.

What other things aside from jitter will affect the sound? And what does jitter do aside from add audible pops and clicks in the audio?

It is my understanding that jitter is a digital signal not being read properly resulting in bits not being read at the right time or not being read at all and perhaps even doubling up on a subsequent bits to catch up. The result being audible pops or clicks. I do not understand how jitter could result in the sound having a different overall "quality". When I say quality I am referring to the characteristics our ears distinguish when we listen to audio. Things like the presence, sound staging, how clean is, how fast is, how airy it sounds... you know all those non-scientific terms we use to describe what we hear.

What is happening in the digital realm to make it sound different from source to source?

I will take your advice and do some testing, but even if I conclude they sound different I still won't understand why. :roll:

Jitter is not heard as ticks and pops. You are correct that jitter is small timing errors but there is so much info being processed that those small timing errors cloud detail and make things sound harsh and thin. The ticks and pops that you would hear are more than likely skips and read errors not jitter.

What is changing from digital source to digital source is jitter primarily. I personally would say the digital cable as well but that is a discussion for another place. All I will say is they don't all sound the same, especially coax digital.

Give that experiment a try and let me know what you find.

Eric

[bnieman]

Eric,

I plan to do the testing this weekend and am eager to report back with my results.

I still don't understand how small timing errors could make a significant change in the overall quality of the music. In order for something to sound thin I would imagine the dynamic range of the wave form would need to be significantly reduced which I can't comprehend how timing errors could do this enough to effect the entire sound.

Furthermore, for something to sound harsh I would guess that the signal would have to start clipping (produce values above and beyond the capability of the hardware). This theoretically could occur if the samples keep trying to double up on one another as they attempt to catch up with the timing. This would definitely be audible if it happened a lot, but if only happened a few times during the 44,100 samples in one second of CD audio, I cannot see how we as mere humans could pick up on it.

Sometimes I wish I was an audio engineer just so I could grasp some of the complicated topics in the art of recorded music

[hifiguymi]

Here is an article that Stereophile did on jitter. Give it a read and it may answer some questions you have about the effect jitter has on an audio signal.

http://www.stereophile.com/features/1208jitter/index.html

Eric

[bnieman]

Very interesting link Eric, I read the main article and plan on reading some of the older articles as well posted at the beginning of the article.

A lot of it went over my head, but perhaps further research will help my understanding.

Analyzing the waveforms reminds me of seeing microscopic photos of average bed pillows showing millions of tiny dust mites... something you never knew was there and now that you know you wish you didn't

Again, thanks for all the information Eric!