[FritzS]

Hi,
why I can transform a SACD (or DVD Audio) surround imaging to the 2 channels of a headphone? Give it some experiences?

 

[DCameronMauch]

Difficult subject. There is something called Dolby Headphone. But that is licenses technology. Can be implemented in software on PC (like in PowerDVD) or in a DSP. Do a search on the web for HRTF (head related transfer function). There are details on the appropriate delay and frequency spectrum adjustments for left and right ears for any particular angle. Then based on the angles each of the six channels of your source are supposed to come in on, you can select right set of data. Then you just have to implement them in circuitry - analog circuitry preferrable. The crossfeed circuits out there now perform a very similar function. Just need to add another four channels. I do not know myself how to adjust these filters to create the appropriate delay and frequency characteristics. Can anyone jump in on that part?

I really enjoy having more advanced topics like this in the forum...

 

[DCameronMauch]

This site has lots of dvd-audio information. Figure 9a shows the standard layout of the 6 speakers. The angles are 0 degrees, +- 30 degrees, and +- 110 degrees. Subwoofer goes anywhere.

Please please don't let this thread disappear because it is too complicated. I think there is enough knowledge in this group to get something like this done. I know there are a lot of newbies here and I do not mean to be giving anyone crap but I am really tired of only seeing threads on real basic concepts.

 

[Won]

Here's some info on head related transfer functions

 

[Vdubjunkie]

Quote:

Originally posted by DCameronMauch I know there are a lot of newbies here and I do not mean to be giving anyone crap but I do not mean to be giving anyone crap but I am really tired of only seeing threads on real basic concepts.

I may be a noob, but I have to agree. This kind of thing is an exciting change from the norm. I'm anxious to see how this plays out.

 

[Dusty Chalk]

I'm wondering if there's some way of incorporating the concepts of crossfeed with the input of a 5.1 system. I'm sure there is, but I'm too lazy to do the math.

 

[thomas]

I experimented with surround sound headphones a while back- I think an entirely electrical system is impossible for DIY, but a combination electrical and acoustic system might work:


Take a look at this, from the headwize library:

http://headwize2.powerpill.org/proje...onnors_prj.htm


I played around with somthing like this many years ago- I think actually using seperate drivers for front and rear produces much more realistic surround effects than any DSP, since it actually uses your natural hearing system to produce the surround effect.

I would run the left and right channel signals into a crossfeed network, and mix it with the centre channel. Then run the two rear channels into their own crossfeed. MAybe some crossfeed would also be used between front and rear channels on each side. Each of those crossfeed circuits would be buffered, and their output connected to a 4 channel headphone amp.


Then you need to actually build the headphones, i followed the basic concept, but made some modifications. I'd recommend using Koss drivers from the KSC/S/portapro line, they're quite cheap and have really good sound for that price. These sound way better than cheap earbuds, esp when placed away from the ear. But they are too large to use the arrangement he showed, so i simply used one front and one rear driver per channel. The drivers should be mounted in a large circumaural headphone, I'm not sure if open or closed is better. Also, the frequency respsonse of the drivers seems to change drastically depending on the distance to the ear, it will take a lot of experimentation or even an equalizer to get a flat response.

You may also need a time delay for the rear channels- this is very hard to do in analog, and expensive to do in digital. I believe almost all analog delay line IC's are discontinued and impossible to find, and in even if you found one, most have a SNR below 60. Digital delay is also very difficult to implement, you'll need a DAC/ADC, and DSP (All 192/24 resolution if you're using SACD).

 

[thomas]

I don't think there's very much chance of DIY'selfing an entirely electrical circuit that can create the surround sound effect. (ie without using 4 channel headphones). Dolby headphone chips are made by several manufacturers BUT its basically impossible to get them if you're not a large corporation. I believe RickCR42 was working on a DH amp, but couldn't get the liscene for it. Also, It seems unlikely that a simple crossfeed-like system would work with normal headphones, since there's no way of recreating the front-rear imaging. For that you would need a DSP system taking into account the HTRF's described by DCameronMauch. I don't think too many DIYselvers have the skill or money to pull this off.... One last idea is to buy a cheap dolby headphone device, and canibalize it for the DH chip.

Anyways, DH and all HTRF processing i've heard still adds noticable colouration to redbook audio, and i don't see it having much use for a high-end SACD system.

 

[grinch]

wouldn't it be possible to set something up (according to that connors surround sound project page) where you:

had two drivers for each ear, and made a dual amp for them, or just use two amps. then you could use the actual decoders on the sacd player to plug the system up.. i.e.:

rear r/l jacks on sacd player -> "rear" cmoy -> "rear" drivers

front r/l jacks on sacd player -> "front" cmoy -> "front" drivers

the only problem is that the "rear delay" thomas talks about might be required. i'm not sure if sacd's have a built in delay for the rear channel, but as far as i understand wouldn't the rear channels be set about as far away from you as the front channels (in an ideal home theatre situation)? if so, i don't think these extremeties would be necessary as either they a) aren't needed or b) already set up via the sacd/dvd/dvd-a.

this is a very interesting subject and i definitely don't know what i'm talking about as far as the electrical aspect of this goes, this is just what seemed to make sense in my head.

 

[DCameronMauch]

Doing this DIY style with standard headphones is totally possible. It is only a matter of the correct filters for each input channel to each output channel (6 to 2). The delay and frequency response characteristics are well known. The existing crossfeed circuits are half the solution right there. They emulate the standard +- 30 degrees for the front left and right channels. So why would it be impossible to add two more input channels for +- 110 degrees and finally the center input channel? The low frequency (aka .1) input channel does not require any processing at all. This is totally do'able. Going digital and doing the DSP thing would yield more accurate results, but would obviously be a lot more involved. But why would this impossible for the DIY'er? Has no DIY'er ever implemented basic audio filters in a DSP? This is not nuclear physics.

Since it seems clear I will not be getting any help from this forum, I have already started working on this by myself. Making pretty good progress.

 

[thomas]

grinch- yeah, that was basically what i was describing, (i suck at explaining things) but some crossfeed between channels might make it work better, since when you're listening to speakers, there is some mixing between channels..

Again, i'm not sure about time delays, I haven't looked at this stuff in a few years, and back then i was experimenting with a DIY pro-logic style decoder. (ie extracting 4 channels from 2) That decoder definately needed a delay, but it may not be needed when you have discrete channels..... I wish i had a multichannel system to experiment with

Quote:

But why would this impossible for the DIY'er? Has no DIY'er ever implemented basic audio filters in a DSP? This is not nuclear physics. Since it seems clear I will not be getting any help from this forum, I have already started working on this by myself

Crossfeed only simulates the mixing of sound and time delays for the two channels, it does not do any HTRF processing. It can be implemented in very simple analog filters. Once you add HTRF processing, it becomes much more complicated- just because the function is well known does not mean its easy to implement. Try designing an analog filter with this response:

You will need 6 different filters like this, each with a totally different response, and you still need to take into account time delays and crossfeed levels (and possibly reverb). Even if it was possible to design such a circuit, it would totally destroy the sound quality.

In other words, you will have to go to the digital domain. If you're using SACD/DVD-a, you will need at least 96/24, possibly 192/24. Simply designing an SACD quality DAC/ADC will take enourmous amounts of time and money. Then you'll need to buy an expensive DSP development kit and program in all of those calculations. I would hardly call HTRF processing a "basic audio filter". This project is definately possible, i just didn't think a DIYselver would be willing to commit so much time and money to a project.

Even the best professional virtualizer systems on the market destroy the quality of redbook audio-i don't know i single person that uses dolby headphone or any other virtualizer for music. Your system will have to beat those by a large margin to maintain SACD quality. I feel the multiple driver solution produces much better sound quality and is much cheaper (even if you make the headphones with senn HD-580 drivers), but to each his own.

good luck on your project

 

[rickcr42]

Quote:

I believe RickCR42 was working on a DH amp, but couldn't get the liscene for it. Also, It seems unlikely that a simple crossfeed-like system would work with normal headphones, since there's no way of recreating the front-rear imaging. For that you would need a DSP system taking into account the HTRF's described by DCameronMauch. I don't think too many DIYselvers have the skill or money to pull this off.... One last idea is to buy a cheap dolby headphone device, and canibalize it for the DH chip.

yeah , a
Dolby Headphone liscence is not for the little fish , way too much incurred cost and many hoops to jump through .

The easy part would have been the implementation using an analog devices dsp chip-codes included ONLY once liscenced

but i have designed some approximations that can sound surprisingly good

best case scenario is main channels full bandwidth with a mild crossfeed

center channel limited to 75 hz-15khz plus a 2db bump around 300hz

rear channels band limited to 100hz-7khz with a 30ms delay

all mixed together for two channel operation using an inverting opamp headpone amp stage

the problem for me was always finding a delay chip that would operate with batteries and last through several movies.

every time i found such a chip it would be discontinued which was VERY discouraging

my point was to make a portable device that could be used with a laptop plus perform double duty as a standard headphone amp with some "ambience" on demand (center ch off , rear on)

I even did some experimenting with totally passive implementations and while not as good as the active delay in sound it still gave a taste of surround sound with headphones-sorta kinda a hafler type ambience extraction

worth fooling around with if you have the time and the skills

 

[DCameronMauch]

You bunch of wankers... I have already got something working pretty decently. And totally analog. It uses six dual op-amps and three output buffers. And I am far from being an EE god. What is with you people? It seems like you guys are so willing to argue why something is too difficult. Try being a little daring. Or are you content forever discussing which op-amp sounds the best? Geeze. Even a digital solution is not that bad. In fact, processing a signal through a known fixed impulse response, is the most basic possible filter. No reverb is necessary, unless you want to create a virtual room too. Not a good idea. You would just be recreating one of the major problems with speakers systems. You hear two different sets of ambient information and your brain gets confused. This is the third time I have tried starting up an advanced topic. And the second time I got nothing but negativity. One was just plain ignored. I just have been so disappointed at the total lack of ambition in this forum. Maybe there should be a separate forum for people who are actually willing to put forth a little effort to discuss and contribute positively to more advanced design ideas.

 

[rickcr42]

I am sorry we disappoint you with our lack of ambition but the point of the diy forum is to help beginners with our "limited" knowledge while at the same time discussing the nuances of some known designs

you want to add the discussion , something informative , fine.

but if we bore you you can always go somewhere else where the discussion will be more to your liking

 

[Dusty Chalk]

Sorry about the above "lazy" comment, but I'm not a DIY-er. It really would be mostly theoretical for me, and I figured that someone with more experience than I could do a better job of taking the concept and running with it.

You can also just use a mixer. Wouldn't have the same effect as crossfeed, but at least you would be able to hear the whole surround mix (e.g., the extra guitar during "On The Run" in Pink Floyd's Dark Side of the Moon).