[Sasha]

Has anyone compared BDP-1 to any of the new USB/SPDIF converters (off-ramp 4 or 5, Berkeley Alpha USB) and if so, in what environment using what equipment?
Here is my predicament.
Let’s assume that claims about more precise clocks in converters such as off-ramp are based on facts (and probably are, off-ramp clocks are most likely significantly better than the one used on ESI Juli@).
And let’s ignore the argument that I2S is the interface you want to use to connect to DAC in order to minimize the impact of jitter, let’s focus on SPDIF for those who do not have DACs with I2S.
Considering that multiple SPDIF conversions (on the side of USB/SPDIF converter or sound card and later on DAC side) will inevitably add jitter (usually it is measured in several hundreds of pS) that the DAC chip will be presented with, is there any benefit of using those USB/SPDIF converters with superb clocks IF you must use SPDIF?
My thinking is the following, let’s say you start with very low intrinsic jitter (~5pS), then transmitter adds several hundreds to place in on SPDIF output, then you add some line induced jitter, then add some on the receiver side, you end up with what, 500 - 700 pS on DAC? Do you benefit here if you start with 5pS vs. let's say 50pS?
Let’s leave “jitter rejection” techniques outside of discussion, they do not work, they only result in jitter artifacts embedded in the signal.
The dilemma here is, if you must use SPDIF do you see any real advantage with USB/SPDIF converters that use exceptional clocks? I do want to send as little jitter on SPDIF line to DAC, but I fail to see real advantage of superbly build USB converter like off-ramp over a well executed PC with good sound card, if SPDIF must be used?

[werd]

Noise trumps jitter in digital playback as a real audio issue. Converters like the Off-ramp are designed to be used in the vicinity of your system. Unlike xlr, RCA needs to be placed close to the dac.

You do not want your computer next to your playback system. IOW the first piece in playback that gets amplified should be a piece like the off ramp or a bdp.

When using xlrs the computer can be placed quite far and even off the immediate power grid feeding your playback system. This means that the first piece to get amplified is your dac.

Sasha you should be using xlr.

[Sasha]

Noise trumps jitter in digital playback as a real audio issue. Converters like the Off-ramp are designed to be used in the vicinity of your system. Unlike xlr, RCA needs to be placed close to the dac.

You do not want your computer next to your playback system. IOW the first piece in playback that gets amplified should be a piece like the off ramp or a bdp.

When using xlrs the computer can be placed quite far and even off the immediate power grid feeding your playback system. This means that the first piece to get amplified is your dac.

Sasha you should be using xlr.

What noise are you referring to?

What difference does it make where the computer is located? I am not talking here about your average PC into which to slap sound card without taking any care of power and electrical noise issues.

Why are you bringing into the discussion RCA and XLR, nowhere have I said anything about connectors?
And what it has to do with the subject of jitter introduced by SPDIF transmitter?

It seems you did not understand the subject of my post at all.

[werd]

What noise are you referring to?

What difference does it make where the computer is located? I am not talking here about your average PC into which to slap sound card without taking any care of power and electrical noise issues.

Why are you bringing into the discussion RCA and XLR, nowhere have I said anything about connectors?
And what it has to do with the subject of jitter introduced by SPDIF transmitter?

It seems you did not understand the subject of my post at all.

The noise that emanates from your hardware like fans. Also the noise that emanates from your speakers  measured in db like noise floor. Or the more elusive noise that presents itself as a high treble clamber that I associate with gating.

Maybe I don't understand the subject of your post.

[Sasha]

The noise that emanates from your hardware like fans. Also the noise that emanates from your speakers  measured in db like noise floor. Or the more elusive noise that presents itself as a high treble clamber that I associate with gating.

Maybe I don't understand the subject of your post.

Ok, we are not on the same page.
There are no fans, no any kind of spinning devices, and regardless that is not the subject.
The subject is jitter added by the process of framing and transmitting data on SPDIF.

[skunark]

Sasha,

The DAC you mention does reclock the spdif stream but you are essentially opening up a can of worms here.. but some points

Clock jitter is additive, since spdif is source synchronous, any reclocking/repeating the signal can impact the overall jitter.  This is something you pointed out which I feel you already know the answer here.... so yes 5ps is always better than 50ps.

Now with USB Audio Devices (DACs, SPDIF, etc) there's three types of implementations (points made with your limitations or an ideal environement):
Synchronous  -- SOF impacts the clock recovery, so jitter is additive (SOF (start of frame) packet is sent every millisecond.)
Asynchronous -- stream is buffered, clock is adjusted based on requirements, jitter is really just the quality of the clock in the USB DAC..
Adaptive -- stream is buffered, stream needs to be overclocked/upsample to match the usb dac, jitter is really just the quality of the clock in the USB DAC.

With all three types of USB Audio devices you have an extra layer of hardware and an extra software driver when compared to a sound card.    There are many other factors with can make either solution unsuccessful but a custom dedicated solution like the BDP (which is a computer) can limit the other factors.   As you might be able to make an USB SPDIF adapter sound as a good as a sound card, i would say the effort is higher.  (even given your requirement to ignore noise, reclocking etc.)

[Sasha]

Sasha,

The DAC you mention does reclock the spdif stream but you are essentially opening up a can of worms here.. but some points

Clock jitter is additive, since spdif is source synchronous, any reclocking/repeating the signal can impact the overall jitter.  This is something you pointed out which I feel you already know the answer here.... so yes 5ps is always better than 50ps.

Now with USB Audio Devices (DACs, SPDIF, etc) there's three types of implementations (points made with your limitations or an ideal environement):
Synchronous  -- SOF impacts the clock recovery, so jitter is additive (SOF (start of frame) packet is sent every millisecond.)
Asynchronous -- stream is buffered, clock is adjusted based on requirements, jitter is really just the quality of the clock in the USB DAC..
Adaptive -- stream is buffered, stream needs to be overclocked/upsample to match the usb dac, jitter is really just the quality of the clock in the USB DAC.

With all three types of USB Audio devices you have an extra layer of hardware and an extra software driver when compared to a sound card.    There are many other factors with can make either solution unsuccessful but a custom dedicated solution like the BDP (which is a computer) can limit the other factors.   As you might be able to make an USB SPDIF adapter sound as a good as a sound card, i would say the effort is higher.  (even given your requirement to ignore noise, reclocking etc.)

Thanks, but I am still not clear on a few things.
Let’s put USB DAC aside for now (DAC with USB interface) and focus on those USB/SPDIF converters.
Let’s say you do start with asynchronous USB transfer in USB/SPDIF converter, and let’s say you do have exceptional clocks for 2 groups of frequencies in it.
Now you have to generate SPDIF and send it to DAC.
Here is the question.
Let’s compare this approach to the approach with sound card (a very good one) in PC that sends SPDIF to the same DAC.
With USB/SPDIF converter you can have much better clocks than any sound card has, that is given I think.
But you still need to generate SPDIF in USB/SPDIF converter, so you are still facing the additive jitter in the process.
USB/SPDIF converter will certainly have an advantage over sound card since it will not be in noisy PC environment (its power can be cleaner, its data lines will not be polluted, etc.).
But since you still need to do SPDIF, all the effort you put into jitter minimization up to that point and all the advantages you had over sound card will start to melt away.
Will you really end up with significantly less jitter on SPDIF output of USB/SPDIF converter than on SPDIF output of sound card, and ultimately less jitter in DAC?
That is my question.

And finally, I am somewhat confused with your last sentence: “As you might be able to make an USB SPDIF adapter sound as a good as a sound card, i would say the effort is higher.”
It sounds as if you think sound card approach is advantageous, did I get it right?

[skunark]

A USB Audio device, whether it's a USB DAC, USB SPDIF, or even the USB Headphones all share the same solution using the USB isochronous transfers.  Where I mentioned USB DAC, your an swap it with the USB SPDIF in my previous post. 

You are making an incorrect assumption that USB SPDIF converters have better oscillators than a sound card.  The ESI Juli card has the two key clock oscillators that are a multiple of the 44.1 and 48khz.  Any good asynchronous USB SPDIF(or usb dac, cd, hdmi, etc) will also have a similar solution.   

Ignoring the impact (if any) the chipset can cause with USB when compared to PCI/PCIe, and also ignoring any issues with USB cables, USB drivers (if custom), the additional power supply for the USB SPDIF device, etc, the flow is really similar between a sound card and a usb spdif device.   The application has to convert the file to a PCM stream and then deliver to the buffer of the sound card or usb spdif device, which eventually is clocked out on the SPDIF outputs and the jitter begins.   Clock oscillators, impedance, isolation, etc impact the quality, and now the tradeoffs can be debated from the same starting point.   

BDP, as a computer, shines here with all the Bryston's customizations to the power supply, spdif outputs on the sound card, OS configuration, etc, and all without the need of a USB cable, external USB SPDIF convert and the additional power supply.   Anyone into computer audio will be working through the same issues for either solution as you are aware.    So to me the USB SPDIF device just increase the complexity unnecessary if you have a spare PCIe slot and if you can find a good sound card, computer and power supply, I would definitely consider that route first.   

Jim

[Sasha]

The application has to convert the file to a PCM stream and then deliver to the buffer of the sound card or usb spdif device, which eventually is clocked out on the SPDIF outputs and the jitter begins.   Clock oscillators, impedance, isolation, etc impact the quality, and now the tradeoffs can be debated from the same starting point.   

Thanks Jim,
That is exactly what I had in mind, probably did not express myself clearly. There is really no advantage of one approach to another, it is all in quality, execution and compromises one makes, in the end you have to face SPDIF and jitter.

[skunark]

Thanks Jim,
That is exactly what I had in mind, probably did not express myself clearly. There is really no advantage of one approach to another, it is all in quality, execution and compromises one makes, in the end you have to face SPDIF and jitter.

Just keep in mind that we narrowed down the scopes and potential problems and focused on the asynchronous solutions.   Synchronous USB SPDIF/DACs jitter starts with the 1ms start-of-frame (SOF) message that the usb audio driver sends.   The adaptive usb dacs requires up-sampling or down-sampling to the preferred rate(s) which the application must perform.    And not to mention there might be another cable and another power supply involved. 

[James Tanner]

Hi guys

Interesting discussion. Do you think Bryston should look at a USB to SPDIF converter product that would do 192/24.

James

[Sasha]

Hi guys

Interesting discussion. Do you think Bryston should look at a USB to SPDIF converter product that would do 192/24.

James

I think it should be done since this seems to be the direction the market is moving to.
Maybe even put USB interface on DAC to take advantage of bypassing SPDIF conversion.
But whatever is done it should be done very well what does not come cheap.

[James Tanner]

I think it should be done since this seems to be the direction the market is moving to.
Maybe even put USB interface on DAC to take advantage of bypassing SPDIF conversion.
But whatever is done it should be done very well what does not come cheap.

So what would you say would be required to be considered .... "done well"

james

[Sasha]

So what would you say would be required to be considered .... "done well"

james

Considering that I did not compare BDP-1, my PC transport with Lynx and a number of high performance USB/SPDIF converters on the market, and how they impact performance of various DACs, I can only go by a few comparisons I did (some of them I made posts about) and few meaningful reviews that included measurements (and I am absolutely convinced that measurements correspond to good sound, if a component, any component, measures well it has potential to sound great, if it does not measure well it is extremely unlikely that it sounds good, this has been confirmed in my experience many times, for this reason I ignore any component that does not measure well and do not bother to evaluate in my system).

Having said that, in terms of figures I would consider "done well" what Bryston achieved with BDP-1/BDA-1 using technology of that time, so if prospect Bryston USB/SPDIF converter would match or preferably exceed the performance of BDP-1 feeding AES/EBU on BDA-1 in terms of jitter figures and spectra, it would be successful product and make Bryston digital products presentable to much wider audience.
And it seems it would be feasible to exceed performance due to SPDIF conversion/transmission taking place outside of noisy PC environment, so if the same Bryston approach was applied, good clocking, isolation, power regulation, it should be successful I think.
Finally with proper USB interface on BDA and avoidance of SPDIF conversion and transmission altogether the performance level could in theory be even higher, so if future BDA with such interface shows less artifacts on its output than what it does today it would be once again a winner.
In other words, check Stereophile review and especially the measurements of BDP-1 and BDA-1, if future USB/SPDIF converter feeding BDA-1 shows less artifacts on analog output of BDA-1 then you got a winner.
And same goes for USB on future BDA if that is the direction Bryston takes rather than standalone USB/SPDIF converter.

[werd]

Hi guys

Interesting discussion. Do you think Bryston should look at a USB to SPDIF converter product that would do 192/24.

James

At first thought the bdp1 would have been a good place to put a converter. But after thinking about it I am glad you didn't since it didn't need to be any more expensive. I would not have used it any ways. An outboard converter is the way to go for those that want it.

[TomS]

James,

It would be really nice to see it in the BDA. I have the Audiophilleo 1 with Philip's new VLN battery supply, using it with the BDA1.  The difference vs the internal USB is not subtle and of course it does 24/192 with OSX, W7, and Linux standard.  Incredible device.

Tom

[skunark]

Hi guys

Interesting discussion. Do you think Bryston should look at a USB to SPDIF converter product that would do 192/24.

James

There's really no point in offering a USB SPDIF/DAC since you have the BDP.  You've really done all the work with customizing the OS, motherboard, power supply, sound card, etc.   Anyone wanting to go down the PC/USB SPDIF/DAC route won't even come close in matching the customizations the BDP provides even if they through more $$ at it.   The processor found in the BDP is mostly used for embedded or very custom solutions and will consume significantly less power than any computer out there today which lets Bryston focus on providing a better power supply to it's consumers.  It will be very hard for any consumer to achieve these results without this type of customization.

USB SPDIF is not an advancement in technology... just another spin of a "USB Audio Device" which has all the same issues as USB DACs with half the parts but now you add more jitter with the digital interconnect and not necessarily for half the price. 

As I agree that it is possible for an audio manufacture to build a complete system around a USB SPDIC/DAC that can sound as good as the BDP+BDA, I also see audio dealers struggle with the demo and users still not satisfied with their USB DACs.  The longstanding issue with USB Audio devices has been the setup.  You have to select the driver, the application and USB port and even minimizing USB devices (i.e. use FW/eSATA HDD avoiding USB).  There's just so much chatter about this I feel folks forget that this is actually a more complicated solution over a sound card.   Yes a sound card can have it's own driver issues, but you avoid any potential issues with USB.

If the goal is to reach an additional market for folks that don't understand the BDP, then sure offer them but charge more than the BDP to cover support issues.    I'm sure the BDP has been costly in that field, but that also could have been avoided/reduced with a user application to help synchronize files between one's computer and the BDP.

[skunark]

At first thought the bdp1 would have been a good place to put a converter. But after thinking about it I am glad you didn't since it didn't need to be any more expensive. I would not have used it any ways. An outboard converter is the way to go for those that want it.

BDP already provides a SPDIF output with the customized sound card .. the USB SPDIF would be redundant.

[skunark]

James,

It would be really nice to see it in the BDA. I have the Audiophilleo 1 with Philip's new VLN battery supply, using it with the BDA1.  The difference vs the internal USB is not subtle and of course it does 24/192 with OSX, W7, and Linux standard.  Incredible device.

Tom

I think you are asking for the BDA to have a higher quality USB input not another product like the USB SPDIF feature.   

[skunark]

Considering that I did not compare BDP-1, my PC transport with Lynx and a number of high performance USB/SPDIF converters on the market, and how they impact performance of various DACs, I can only go by a few comparisons I did (some of them I made posts about) and few meaningful reviews that included measurements (and I am absolutely convinced that measurements correspond to good sound, if a component, any component, measures well it has potential to sound great, if it does not measure well it is extremely unlikely that it sounds good, this has been confirmed in my experience many times, for this reason I ignore any component that does not measure well and do not bother to evaluate in my system).

Having said that, in terms of figures I would consider "done well" what Bryston achieved with BDP-1/BDA-1 using technology of that time, so if prospect Bryston USB/SPDIF converter would match or preferably exceed the performance of BDP-1 feeding AES/EBU on BDA-1 in terms of jitter figures and spectra, it would be successful product and make Bryston digital products presentable to much wider audience.
And it seems it would be feasible to exceed performance due to SPDIF conversion/transmission taking place outside of noisy PC environment, so if the same Bryston approach was applied, good clocking, isolation, power regulation, it should be successful I think.
Finally with proper USB interface on BDA and avoidance of SPDIF conversion and transmission altogether the performance level could in theory be even higher, so if future BDA with such interface shows less artifacts on its output than what it does today it would be once again a winner.
In other words, check Stereophile review and especially the measurements of BDP-1 and BDA-1, if future USB/SPDIF converter feeding BDA-1 shows less artifacts on analog output of BDA-1 then you got a winner.
And same goes for USB on future BDA if that is the direction Bryston takes rather than standalone USB/SPDIF converter.

Since there are issues with both SPDIF and USB DACs, I feel Bryston would be better serving it's customers by providing custom i2s (like) interface between the BDP/BCD and BDA and also a viable USB input.     Hopefully a custom i2s interface could be something shared across the industry.