DSD through a processor

Hello,

I am building a very high end car system and I am not able to get an answer from the 12v community.

I am looking to take the analog out of the new Sony high resolution deck (w/DSD or 24/192) and go into a Helix DSP Pro signal processor. When doing so, the Helix will convert A/D and D/A. When it does, will there be any degradation? The Helix is 24/96 limited. Here is a link. HELIX DSP PRO
I am not looking to go digital out from the Sony, as I want the higher resolution.
TIA, Mike
 
I would say that you would definitely have a significant amount of degradation. I am assuming you are talking about the Sony RSX-GS9. Based on the picture, the Sony has many Audiophile level design elements. Split layer separates digital/analog sections and shields against digital noise. It also has very nice power supply sections for all the op amp analog stages. Going through the Helix would kill any of the benefits of this Sony deck as the Helix would do everything based on PCM. All the A/D, processing and D/A will lose a lot of the hi-res signature coming from the Sony. Really the only reason to put in a Helix (or something like the Audison Bit One) is if you absolutely have to have some sort of EQ or correction (such as taking signal from speaker wires or re-eq of factory radio systems). Also, the Helix will not have power supply and DAC/audio stage quality anywhere near the Sony (in my opinion).

Please be aware that subwoofer and some configuration settings may not work during DSD playback. I have not seen any product that can process DSD (i.e. crossover, eq, etc.). Systems normally convert DSD to PCM to do any processing because of the math involved.

Since the Sony already has a front/rear/sub preout and also has time alignment, I think the biggest issues with sound quality are already dealt with. Just look to really nice amps/speakers.

If you really need an EQ, I would look towards the analog equalizers, such as the AudioControl EQS/EQX/LCQ products.
 
Last edited:
Hello ainami, thanks so much for the response. Yes, I am looking at the Sony RSX-GS9, but I am wanting to utilize the DAC ( and to your point, other audio goodness), and native DSD, 24/192 ability. I was afraid that the Helix would pollute the Sony signal/downgrade the resolution. My system is a 3 way active set up, subs, mids, tweeters. The only other option I have considered is:
-Tablet w/external HD and audiophile hi-res app with 30 band EQ and memories
-Out of tablet, digital into the Sony deck (through the DAC input)
-Use the delay in the Sony deck for each channel
-Out of the Sony into a 3 way active crossover, like the AC 6xs

Clearly you have an understanding of my goals. What are your thoughts on this set up? Other options you can think of? Thanks again
 
I think that sounds like a great option given what you what to accomplish. I think the biggest choice you have is whether or not you absolutely have to do DSD. Don't get me wrong, DSD is the absolute best digital source medium for audio. However, it really cannot be processed (which means you really cannot apply DSP for eq, crossover, room correction). If you play DSD, you really want a pure 2-channel system that has no processing. You can still use analog crossover like AC 6xs, but there is no digital equalizer available. If you really want to run DSD and still have eq, I would recommend putting in a AC EQL, or even a Kicker 03KQ30 if you want 1/3 octave control. Put the EQ between the deck and crossover, such as:

Sony Deck ==> Analog EQ (ACL EQL) ==> AC 6xs

I see that the AC 6xs has the ability to drive up to a 13 volt signal. This is good and if you want to take advantage of this, I would mount the 6xs as close to the deck as possible. Then crank up the gain on the 6xs and turn down the input gain on your amps. That way, it will remove a large amount of "hiss" you get from natural gain. If you mount the 6xs further away from the deck, you're really not getting this benefit because you will be running low-gain signal interconnects between the deck and the 6xs.

As far as the tablet, go for a Windows 8 tablet. This should support either JRiver or HQPlayer as hi-res software app. I use Jriver here at home. I have heard good things about HQPlayer. Both are good, but they have different main sound quality features.

JRiver has an excellent DSP plugin which allows you to add 10-band standard eq or a completely customizable parametric eq. The parametric eq allows you to add unlimited fiters for either or both left/right channels. This allows you direct manual control and you can hear the change right away.

HQPlayer has a lot of upsampling/conversion filters that you can apply to the sound (like upsampling DSD or converting PCM-to-DSD. There is no manual eq/dsp available. Many like the performance of these filters.

Both allow the use of Convolution (room correction) files that can be engineered from software such as REW or Acourate. If you choose this direction, be prepared to spend hours and hours doing sound engineering in your car. When using DSP/Convolution in JRiver, it will not play DSD or you will need to configure it to convert DSD-to-PCM so that it can be processed. The HQPlayer author says he will process DSD using room correction, but he does not say exactly how he is doing this, just that it's a "magic" feature in his software. It is suspected that he is converting DSD-to-PCM to do the Convolution and then back to PCM-to-DSD. We know that it soaks an enormous amount of CPU to do this.

If you choose to use room correction methods or any of the HQPlayer advanced conversion filters, be prepared to buy the fastest tablet CPU you can afford because this stuff sucks CPU! If you just want to stream DSD/PCM "as-is" (or with minor eq in Jriver) to the Sony, then tablet CPU performance is not as important.

I have played with room correction systems (using Dirac software) and I have not had good experiences with it. It changes things in weird ways. Phasing becomes weird and the effect can be like hearing the sound in your head (like headphones) instead of out in front of you. Also, room correction can be aggressive in the bass correction area and may push your amps/speakers into peaking, which is really bad. I have found stuff sound better without room correction (but there will be many people who think room correction is great).

Also, when doing eq adjustments, I would advise to use eq conservatively. I would not make any adjustments more than about 5db. EQ is really meant to correct small environment defects and make the sound a little truer. It is not meant to drastically fix a bad environment. For example, say you have a hole from 72Hz to 80Hz where the bass drops out completely (like 20-30db). This is really from the bass waves being cancelled out within your environment. Trying to EQ this will cause the woofer to move too much. In the end, you are wasting amp power trying to correct an area that really cannot be fixed. This generally does not happen in cars, but it's just an example. It's better just to live with that small dropout/hole and focus on what you can fix. My own opinons.
 
Last edited:
oh, a quick note. You may have to play with configuration in the HQPlayer or JRiver app to get the tablet to interface with the Sony deck. I don't know that this particular scenario has been tried, so either of these software may or may not work, or may not work with all features. I would get the tablet and download trials of each. Then determine which you like. Also, it looks like the USB-b input on the deck is what you want to use. There is limited DSD in USB-a.

It's totally fine to do DSP in the HQPlayer or JRiver software because the information is still in the digital format domain. You only want to do a D/A once and you want to do it the best you can.
 
Last edited:
Wow, thanks again for such a detailed response, I really appreciate it. Clearly you have an understanding. I am intimately familiar with high end 12v, I spent years in the biz, just short on hi res knowledge, but very interested.

If I was to use the HQPlayer or JRiver, do either have an EQ that could be utilized while playing DSD? I will not be able to do an analog EQ after the Sony, as I want to use the delay for all (6) channels out of the Sony. Thanks again!
 
I understand what you are all saying here, but have you heard a properly tuned dsp based mobile system? I guarantee the benefits of dsp in a car far put weigh any negatives you get from the a/d - d/a conversion.

Just because you choose to send the signal through a bunch of analog processors doesn't mean you'll maintain signal integrity or purity. The reality is the Helix DSP Pro is a very powerful mobile dsp. Although I only use the regular Helix, it doesn't impart any noise or have any operational issues.

Having time alignment and a powerful eq section really help with the awful environment in the car. Honestly, give the Helix DSP a shot before you go down this other path. It's simpler and more powerful.
 
recca - you are absolutely correct. I'm sure there are many people who consider the Helix DSP completely acceptable and a fine piece of gear. I am probably in the top 3% of audio guys. For example, I can actually hear the sound quality difference between audio interconnect cables. I can definitely hear problems introduced by A/D and D/A stages of very high end home theater processors (where you have ultra high-end components and a very good +/-15V split power supply (where you actually have 30 volts for the audio circuit to use for doing audio stages). In car audio, you only have +12V, where the product manufacturer has to pay very close attention to what kind of op amps can operate well in that power supply voltage range. It's all relative. Precisionmike asked for opinions, and I gave mine.
 
I hear what you're saying, but what I'm trying to say is that the difference that you'll hear from an uncorrected listening environment is 100000x greater than the difference you'll hear from AD conversion. Its just choosing the lesser of 2 evils.

FWIW, looking at the manual for the GSX9, native DSD is not available when using any onboard processing besides sub crossover.
 
Precisionmike - I tried to explain above, and I know it can get confusing, but the short answer is "No", there is no hi-res app (like JRiver of HQPlayer) that will EQ while playing DSD. If you are using something to EQ a DSD song, then you really aren't hearing true DSD (because internally the DSD data is being converted to PCM 24/192 or 24/96 to do the "EQ" process). I see that the Sony deck has a "DSD Sub" feature which allows you to specify a subwoofer crossover frequency when playing DSD. I don't know how Sony is doing this. I doubt that they would develop a special digital algorithm to process the DSD, but I could be wrong. Or they could be internally converting to PCM to do this, or they could be using some of the DSP processing available on the actual ESS DAC chip to do this crossover (internally on the ESS, it could also be converting to PCM to do this processing). These are all questions I do not have answers for. Why not get both JRiver and HQPlayer? Since you are looking at stuff like the Sony and Helix, dropping a couple hundred bucks on software may not be a big deal. You can use JRiver when playing normal PCM and take advantage of the EQ, and then use HQPlayer when you want to do DSD and apply your choice of the DSD filters.

I don't know if you have got your amplifiers yet, but looking further into this, I would probably not use the AC 6xs. You can use the 6xs if you have to, and it's an excellent solution, but the signal is going through several more op amp audio stages within the crossover (these can degrade the audio signal a little as well, but it's still better than doing A/D - D/A in a Helix processor). I would run the output of the Sony deck directly to the amplifiers. I would look at something like the Audison TH Quattro and use the crossover module option in these amps to split the signal between mids/highs. The crossover modules in the Audison are just resistor networks that determine the crossover point directly in the audio stage of the amp. I would run the Audison in "A Class" mode for best sound quality (using a separate amp for sub/rear channels of course). I may be splitting hairs here, but this is how I would approach the system if I were designing strictly for best sound quality. You could then use the eq/dsp in JRiver to do your correction and settle for PCM files. This would also satify recca's opinion that corrected sound is better than un-corrected.

6xs internals:
http://forums.trinituner.com/upload/data/bf/Copy of DSC00567.JPG

I know that you want to use all 4 channels of the Sony output, but in my experience, the rear channels can actually degrade sound quality. I had a very nice system back in the day in a small hatchback and I found that when I disabled the rear channels, the audio quality actually became cleaner and clearer. Also, the imaging was raised up to where the instruments seemed to be directly in front of me instead of down below (I had a tweeter and 6.5" mounted in custom fibre glass kick panel enclosures). Even if you use time alignment and EQ, the sound from the rear speakers is going to reflect all over the place and will blur/smear the sound some. Ultimately, the left channel and right channel should each be coming from a single point source. I'm splitting hairs here as well, but I could hear the difference. So, you can use all 4 channels when you have people in the back, but then fade to the front for critical "driver only" listening.

I'm sure I don't have to say this, but this extends to mounting position of the speakers. Having a tweeter closer to your ear than the woofer is a really bad idea. You can pretty much throw out any attempt at time alignment if you do this (such as if a tweeter is 2 feet away in the upper door and the woofer is 3.5 feet away down below). At least try to get the tweeter and woofer exact same distance away from your ear. In a perfect world, both tweeter/woofer would be in the exact same spot to form a single point source (such as with a coaxial speaker). Unfortunately, this is not always optimal in a car situation.
 
Last edited:
Further research shows that JRiver can do processing at DSD level, but only if you force DSD output of all audio (which is not a bad thing here). DSD processing has to convert the DSD data to PCM before processing, but it seems to be okay if the PCM sampling is done at DSD data rates (which JRiver does do). You are going to use a lot of CPU here, so a very fast tablet is required, but I actually think this may be a viable option. Configure JRiver to output everything as DSD, as described in the link here:

NEW: Real-time DSD output

Then you can add EQ filters in the JRiver DSP section and this will apply to any song being played from JRiver. All songs will be output to the Sony using DSD (reglardless if it's DSD or PCM). JRiver uses some very high sampling rate 64-bit data algorithms to process DSD, as below:

DSD and Digital Volume Control / Digital Crossovers??

Using the DSD Sub crossover in the Sony may not be that bad. It seems that someone has already developed a DSD Crossover (they also convert to very high rate PCM -- this is much higher than 24/192 .. probably something like 24/768 or higher).

http://www.diyaudio.com/forums/pc-based/253702-dsd-sacd-fir-based-software-crossover.html
 
Thanks again for the response, fellas. My overall goal is to be able to have the highest resolution, but also will require some processing. The configuration is a bit unique and since you both have a high level of understanding, I will explain a bit further. Again, I truly appreciate your input here.

I spent about 20 year in 12v retail and was an avid competitor for about 10 years, a few of world finals trophies under my belt, but hi res is all new to me. I have heard it, and I want it. Back "in the day", I heard all the top car, which I believe rival most anything out there today, from what I have seen. Of all the great systems, there was one that was in a different class, IMO. That car was owned by Earl Zausmer. Earl Zausmer - 1994 BMW 540 - 1995 CA&E. I knew some day I would have "that sound". The car I am building is the same body style (BMW E34), and I am building the system much the same way, but with some improvements. Earl and I are now very good friends, and he is assisting with the design. I am using the same mid/tweeters (which took me the past 10 years to find). At this point, I am working on the dash, see pics attached.

My goal is 30 band EQ (w/memories), TA and active 3 way x-over.

I realize what each of you are saying regarding the Helix, and the strength of the processing. Perhaps the only way to know which is best is to test both options. Thanks again.
 
6.6.2016 001.jpg

6.6.2016 004.jpg

6.6.2016 011.jpg

6.6.2016 012.jpg
 
Wow! That looks like some awesome fiberglass work. And great choice on the B&W drivers! Yeah, the Silver Signature drivers are rare and really hard to find. I actually use B&W Diamond series here for my home theater speakers. They have an excellent full-bodied midrange.

The Helix samples at 24/96 PCM and will only process at 24/96 PCM. At this point, you have lost the accuracy of the DSD waveform (which essentially samples at 2.884 Mhz minimum). JRiver may not process DSD at that sample rate, but it's much better than 24/192. It probably does sampling/process of DSD at somewhere around 24/768. I would try this choice first.

JRiver has two Parametric EQ in its DSP stack. You can define unlimited filters for each of these Parametric EQs, though I don't think you will actually need to define 30 individual bands. These two Parametric EQs can be used together or individually and can be enabled/disabled through the JRiver UI at any time. There is also a normal 10-band eq that you can use as sort of a "third memory" if you wish. I would try the tablet/JRiver option first before going to something like a Helix.

https://wiki.jriver.com/index.php/DSP

In my 20 years of audiophile listening, I have really never heard a tube amp that did it for me. They just have never been exciting enough. I have even heard a megabucks $100k set of tube monoblocks going into some Avalon ceramic/diamond speakers. It sounded "nice", but it just did not reach out and grab me. The tube amps just don't have the excitement that a solid state amp does.

Now, Class A solid state gets interesting. I have spent the last year rebuilding a home audio dac, going through various capacitors, regulators and op amps. When I used Class A fully discrete op amps, the result gave me everything I wanted. I posted my comments here:

http://www.diyaudio.com/forums/digi...te-opamp-output-my-cd-player.html#post4741445

I believe the Class A nature of the slew rate curve is the same thing that happens inside tubes. The Class A circuits still give me a lot of excitement and have excellent high-frequency response. And nothing is harsh/bright. That's why I would personally lean toward using an Audison in Class A mode. However, if you really love the Milbert tube amp signature, you should definitely go for that. You will have to use an active crossover in this situation, though (like the 6xs you mentioned). Thinking about this, the tube amps may help tame any brightness/harshness that the B&W metal dome tweeter may produce. It may compensate and end up being pretty good. Of course, in my experience, the less audio stages a signal has to pass through the better.
 
Thank you very much for the compliment. The dash is very rough in those pics, I have been filling/sanding since. The combination of the B&W drivers and Milbert amps is amazing. Very warm, detailed, full with amazing dynamics, especially with the driver placement. I find the on axis greatly reduced reflections and preserves the sound. As you can see, the drivers are positioned directly at the driver's ears (with a laser pointer), hence the need for delay.

If I understand the Sony manual, I think I can still use delay on all (6) output channels when in DSD, correct? I realize the crossover selection is limited. I was even considering using an active x-over on the subs (2xs), using the Sony HP x-over on the mids, and (good, custom) passive between the top of the mid and tweeter. Just a thought; that way the upper end signal is not going though the additional electronics of the 6xs.

Thanks again for the Jriver info. For the tablet/Sony/6XS set up, that looks like just the ticket.

Great chatting with you
 
Ugh, god, nothing is perfect, isn't it? I just re-read that section of the Sony manual and I may have to take back the idea of running crossover from the Sony. It's great if you're just running PCM, but as soon as you get into DSD, everything falls apart. The DSD Sub is limited to 150Hz low pass and you will lose a lot of the tight midbass that the B&W will do in the 80-125Hz area. Also, the time alignment for a subwoofer is actually not available when playing DSD from USB-DAC in the Sony deck.

In my opinion, time alignment for the subwoofer is not as critical as it is for mids/highs. If you feel this way, there are a couple of option I can think of. You could use the Milbert TC-2r tube crossover. This will separate the sub from the rest of the range. You could also start swapping out the tubes in this crossover to help tune in the sonic signature that you like. You could then use high-quality passive to do high-pass on the tweeter and low-pass on the mid (driving the mids/highs from either the same amp or separate amps if you want). This is really not a bad scenario. Please remember that ultra-high end home audiophile speakers all use passive crossovers 99% of the time. You can use Parametric EQ in JRiver to correct for any bumps/drops from the passive crossover points. Engineering a 100% accurate passive crossover is extremely difficult as you will have to execute lab tests for frequency response and impedance response from each driver in their respective enclosures and environment and then engineer your passive circuit using audio circuit software. Much easier just to correct in JRiver Parametric. Passive crossover components can get really pricey. Remember that you are limited to a 6db or 12db per octave crossover point. This is generally not a bad thing for mids/highs. Two sites that can be useful if you choose to go this way:

Humble Homemade Hifi - Cap Test

Parts ConneXion - The authority on Hi-Fi DIY parts and components

Alternatively, you could use a single 6cx to do all three (sub/mid/high). It has a much more aggressive 24db/octave crossover slope. May sound a little different. Blending between mid/high may be a little different, but the steeper crossover slope can help with reducing phase errors.
 
If you really want to do time alignment for the subwoofer as well, I see a couple different scenarios. You have to use two separate crossover units:

Scenario 1:
Sony Front outputs ==> Milbert Tube Crossover ==> 1 or 2 amps using passives on mids/highs
Sony Rear outputs ==> Audiocontrol 6xs ==> subwoofer amp

Scenario 2:
Sony Front Outputs ==> First Audiocontrol 6xs ==> separate amps for mids/highs
Sony Rear Outputs ==> Second Audiocontrol 6xs ==> subwoofer amp


Time alignment will work completely in the Sony because all channels are driven full range. There is no real "processing" for doing time alignment for DSD bitstream data. The Sony will just hold the bitstream data in cache memory for a certain amount of milliseconds before sending it on to the DAC chip.

Using 6xs for the mids/highs will help prevent a lot of the passive crossover engineering errors, but you will lose some of that tube signature that the Milbert crossover can provide. You could try Scenario 2 first because it's cheaper. You could always switch over to Scenario 1 with passives if you still feel you are missing some of that tube richness. I actually have no recommendation on which of these options to try first. They are just options.
 
Last edited:
Wow, thanks again for all of the information!

I like the idea of either the 6xs, or passives, to preserve the DSD or 24/192 quality. I think I can use the Sony sub out also, correct? I plan on using similar x-over points to what Earl used:
Subs- LP @ about 300hz, preferably 24db/oct
Mids- HP @ about 300hz, LP @ about 3khz preferably 24db/oct
Tweeters- HP @ about 3khz preferably 24db/oct

As far as delay goes, I could live with no delay on the subs. In a perfect world, I would only be using delay on the drivers side, as the passenger sub will be the furthest driver away, everything else delayed to match, to the driver's seat.

So as I see it, these are the options:

OPTION #1-
Tablet w/JRiver and EQ, external HD for additional storage
Into the Sony through the Micro DAC (DSD) input
Use delay and output levels only (no other processing) in Sony, F,R,SUB out
R,F,SUB into 6xs, 6xs handles all crossover, 3 way active. Potential upgrades from 6xs? Pro audio?
Output to amps- (2) tube for mid/tweeter, (2) solid state for subs

OPTION #2
Tablet w/JRiver and EQ, external HD for additional storage
Into the Sony through the Micro DAC (DSD) input
Use delay and output levels and HP for mids out of Sony
SUB LP- 24xs
Output to amps- (2) tube, (2) solid state for subs
Mid LP and tweeter HP- custom passive

Make sense? Thanks again!
 
Here's a link to the page w the manual: https://esupport.sony.com/US/p/model-home.pl?mdl=RSXGS9&LOC=3#/manualsTab

I might have been wrong about what dsp setting are available during playback.

1. There's a dsd sub setting which allows for a 150hz lpf. Obviously not ideal.

2. The POS (position) setting says it's available for any input which would be huge. TA is the hardest part to find and analog option.

3. This unit is more interesting than I first thought. I can be used in a full USB DAC mode which would mean you could potentially use an Android device with an OTG cable.