iFi Audio Pro iDSD discussion thread

[iFi audio]

I would imagine so, as the audio signal passes through the SS or tubes on its way to any output, either the HP jack or RCA and XLR.

 
Hi,
 
To answer this and the previous question posed.
 
 
For the iDSD Pro we really went back to the drawing board after the first version. As a result its basic topology is very purist; which literally means just a passive lowpass filter, Volume Control and an Output Amplifier that could drive Headphones or Line Outputs with equal quality. This ultra simple signal path is one of the reasons that allowed the iDSD nano to offer such good sound quality on such modest budget. Its giant slayer status is largely attributed to this approach and we are adopting the same for the Pro iDSD.
 
Of course, for the iDSD Pro we have improved the quality and performance of these basic blocks massively. And we mean massively.
 
The passive low-pass filter incorporates inductors to form a LCRC filter with better filtering of unwanted digital noise while offering a wider audio bandwidth free of analogue phase-shift.
 
The Volume control is no longer a Chip, instead a balanced Alps Potentiometer (16mm Alps Black) is used. There is more to this Volume control which will be revealed at a later time.

For fixed level line outputs the volume control is fully bypassed via relays (just like you find on the AMR DP-777) for the purest possible sound, connect a headphone and the Volume Control is restored.
 
The Output Amplifier is our new all discrete “Pro-Stage” which does away not only with Op-Amp Chips themselves but even with traditional Op-Amp like structures implemented discretely (e.g. like all these “discrete Op-Amp replacements”. It should be quite unique. Instead the circuitry owes much to the discrete designs of both Neumann and TAB-Telefunken for the West-German radio & TV studio equipment and those of the RFZ in former East-Germany, either of which justly enjoy 'legendary' status in the Pro audio world. Integrating the option to switch the crucial and sound quality determining first gain stage between tube and a j-fet gives additional control over the sound quality.
 
The actual output stage of this circuit is a new Class A design with enough current delivery and available voltage (14V peak & 700mA peak) to drive any modern headphone and equally capable of driving even the most difficult line inputs (e.g. 600 Ohm matched transformer coupled Pro-Audio).
 
As such there is no Headphone Amplifier that could be removed, if a Headphone is connected the line outputs are disconnected and muted and the same single stage now drives the Headphones. Equally, the  choice between Transistor or Tube for the key amplifying stage applies equally to line outputs and headphones.
 
We hope this clarifies this part of the design somewhat. As said,it is a ground-up new design - it is unlikely to arrive before November. For the moment, this is as much as we can say - we are trying to walk the line between not getting you all too excited, yet answer some of the questions posed without giving too much away. Not easy.

 

[Mach3]

That's great news, in time for a nice Christmas present

 

[Franatic]

It's nice to hear the latest on the Pro. I am very excited by the new plan that came off the drawing board.
 
The original price estimate was  ~$1500. Are we still in that range or has that moved significantly with the redesign?
 
I would love to see it by the end of the year. Thanks for the update.

 

[EVOLVIST]

  That's great news, in time for a nice Christmas present

 
If you love yourself, more like Valentines Day, I'll bet.
 
Remember, even with a final design, it has to go into production, soooooo....save up by Valentine's Day and wait it out.

 

[JWahl]

Disclaimer : I'm not a Schiit fanboy at all, never owned any of their products. (but I'm gonna receive a Regen in the coming days)

The launch of the Yggdrasil a couple months back, promising (and apparently really offering) REAL TRUE NATIVE PCM conversion for 2k$ and change may have rattled a bit our favorite Englanders (and German born designer in chief).
USB interfaces have also progressed a lot these last few months (signal reclocking and impedance matching, see Uptone Regen for instance)

A full product reevaluation was to be done I guess. Wish the resulting products will be worth the wait!

I'm kind of in the same boat. I like iFi products and the high value they present, but I'm also sold on buying an Yggy for my personal preferences. It's seems like it's really setting the bar below $3k at least. Probably more comparable to the AMR stuff though different flavor.

I don't doubt for a moment that iFi is capable of designing something competitive. It's more of a question of their future business plans. Because if they make it "too good" it might encroach on their AMR gear. At least on sound quality, not construction. And I think that's what makes designing the Pro series difficult, making it good, without invalidating their traditional high end.

Though to be honest, I'm partly curious of what an Yggy and ICan pro combo would sound like.

 

[sdolezalek]

Actually it is the new Schiit multi-bit version of the Gungnir DAC at $1,249.00 that probably becomes the price competitive target for the Pro.  Two very different design approaches, however.  Cannot wait to see what the final iFi Audio Pro looks and sounds like -- but it sounds as though we will all have to wait a few more months...

 

[h1f1add1cted]

  Actually it is the new Schiit multi-bit version of the Gungnir DAC at $1,249.00 that probably becomes the price competitive target for the Pro.  Two very different design approaches, however.  Cannot wait to see what the final iFi Audio Pro looks and sounds like -- but it sounds as though we will all have to wait a few more months...

If I'm not completly wrong, since the micro iDSD or now the micro iDAC2 you can switch to Bit-Perfect filter (which will come with the Pro iDSD too), which means to "turn off oversampling" and act like multibit / R2R DACs in NOS mode, like the iDSD measurements:
 

 
If you compare with TDA, AD or PCM1704 multibit / R2R DACs the have the same treble roll-off like the iDSD or iDAC 2 if Bit-Perfect filter setting is on. Should mainly sounds pretty similar.

 

[EVOLVIST]

If I'm not completly wrong, since the micro iDSD or now the micro iDAC2 you can switch to Bit-Perfect filter (which will come with the Pro iDSD too), which means to "turn off oversampling" and act like multibit / R2R DACs in NOS mode, like the iDSD measurements:




If you compare with TDA, AD or PCM1704 multibit / R2R DACs the have the same treble roll-off like the iDSD or iDAC 2 if Bit-Perfect filter setting is on. Should mainly sounds pretty similar.

You know, this has always been my understanding of bit-perfect, but I've never thought about it in relation to R2R DACs.

I wonder if iFi can illuminate this.

 

[sdolezalek]

You have the additional complicating factor that if you upsample before the DAC, you can choose a less sloping filter as well as exactly where the cutoff starts (which theoretically is one potential benefit from DSD256 files).  The guys behind HQPlayer have done a lot of work with the iFi iDSD Micro and their upsampling software.    

 

[MLGrado]

 
If I'm not completly wrong, since the micro iDSD or now the micro iDAC2 you can switch to Bit-Perfect filter (which will come with the Pro iDSD too), which means to "turn off oversampling" and act like multibit / R2R DACs in NOS mode, like the iDSD measurements:
 

 
If you compare with TDA, AD or PCM1704 multibit / R2R DACs the have the same treble roll-off like the iDSD or iDAC 2 if Bit-Perfect filter setting is on. Should mainly sounds pretty similar.

 
 
Actually, in all cases the iDSD's Burr Brown dsd1793 is still partly a delta sigma DAC, so it is not like an R2R DAC  completely.  It DOES use discrete switches, though, for the upper 6 bits, much like an R2R DAC.  The rest, though, is handled by a third order delta sigma modulator, yes, even in bit perfect mode.  
 
In bit perfect mode, the 8x, 4x, 2x  oversampling circuit that precedes the segment dac is bypassed.  So, if you send, say 96 khz signal in standard or min.phase mode, it is oversampled or upsampled x4 to 384khz.  The 384 khz signal and accompanying clock rate are sent to the segment dac for conversion.  If you send 96 khz in bit-perfect mode, it stays untouched before the segment DAC.  So, instead of seeing 384 khz, the segment DAC will see the original 96khz signal.  
 
So in bit-perfect you eliminate any potential quality loss that may arise from use of the oversampling circuit.  Also, you do get true bit-perfect conversion then, for the top 6 bits of the digital signal.
 
It does sound very much like a 'true-PCM' DAC, but still, it isn't quite the same as a true multi-bit DAC.   iFi was looking for a DAC that provided real, DSP free DSD conversion, but were not willing to compromise on the sound of PCM.  The best DAC that met their requirements, that handled both formats as well as possible, was the Burr-Brown.  
 
 
But, I agree, even though technically not quite the same, it does sound quite a bit like true PCM DACs.  And that is a very, very good thing, in my book.  
 
---------------------
 
I would LOVE to see the new pro DAC available by Christmas.  That would be perfect.  But I am hoping at the latest early 2016.  I LOVE my iDSD Micro, and refuse to replace it with anything that isn't iFi.  I expect the new pro DAC to be an amazing product.  

 

[iFi audio]

You know, this has always been my understanding of bit-perfect, but I've never thought about it in relation to R2R DACs.

I wonder if iFi can illuminate this.

 
Hi,
 
A Short History of Chipset Euphoria (nod to JK Galbraith readers

)
 
Very fitting in the current economic climate:
 

 
 
In our book, firstly, R2R done properly takes a LOTTA work. Secondly, the statement 'Bit-Perfect Digital Filters' is a contradiction in terms; it only exists in sci-fi as it is a 'true oxymoron'¹.
 
Multi-Bit CAN be Bit-Perfect, but often oversampling and digital filtering is used, that cannot be defeated by the user. To us this not something we would implement.
 
Further, while the mainstay of CD standard Audio, neither Multi-Bit nor Delta Sigma on their own are great for wider dynamic range coupled with higher sample rates.
 
Multi-Bit systems struggle with distortion at low signal levels. The TDA1543 from Philips which is found in many Non-Oversampling Multi-Bit DACs for example is show to have at best 14.5 Bit linearity at low levels with worst case less than 13 Bit linearity. This is not even really good enough for CD.
 
By comparison, Delta Sigma DACs can have essentially distortion at low levels that are equal to the noise (read no extra distortion). But Delta Sigma has its own limitations due to supersonic noise, slew limiting and lack of resolution to handle large dynamic steps.
 
We feel that splitting the job between Muti-Bit for the high-level parts (where it is most capable) and Delta Sigma for low-levels is the best choice when attempting to go past the CD-Limit of dynamic range. If we need to play back CD, well, the TDA1541A or UDA1305T are still our preferred choice, but they are limited to 16 Bit performance.
 
 
Cheers.
 
 
¹ A 'true oxymoron' (as opposed to an 'oxymoron') - is "something that is surprisingly true, a paradox²."
² A paradox - the Greek roots translate to “contrary opinion,” and when two different opinions collide in one statement or action, that's paradoxical.

 

[EVOLVIST]

So, I'm guessing that the new iUSB 3.0 is going to pair nicely with the iDSD Pro.

Just my way of shamelessly bumping my favorite thread.

 

[technobear]

So, I'm guessing that the new iUSB 3.0 is going to pair nicely with the iDSD Pro.

Just my way of shamelessly bumping my favorite thread.

I'm guessing that the Pro won't need any such accessory because it's all in the box already.

 

[EVOLVIST]

I'm guessing that the Pro won't need any such accessory because it's all in the box already.

 
That would be great, yet it won't stop me from getting the iUSB 3.0 (and I suspect that would be the same for you). I imagine the end result, if you're correct, would be akin to chaining two REGENS together, yet without all of the fuss associated with the REGEN.
 
No telling when the iDSD Pro will come out, so might as well enjoy the iUSB 3.0 in the meantime.

 

[Franatic]

   
That would be great, yet it won't stop me from getting the iUSB 3.0 (and I suspect that would be the same for you). I imagine the end result, if you're correct, would be akin to chaining two REGENS together, yet without all of the fuss associated with the REGEN.
 
No telling when the iDSD Pro will come out, so might as well enjoy the iUSB 3.0 in the meantime.

You have not gotten your Regen yet, I gather. Did you cancel it? I can vouch that it is a very impressive upgrade for the micro. It has helped me to be patient in this drawn out development/regroup/redevelopment of the pro. I am curious whether the iusb 3.0 can be better. 
 
Is the iusb 3.0 already available?  http://www.ttvjaudio.com/iFi_iUSB_3_0_p/ifi0000032.htm
It doesn't mention availability............$400, not cheap.