Yes, USB 2.0 has more than sufficient bandwidth to supply uninterrupted data stream at 4 channels / 384 kHz / 32bit. Multichannel files work fine - at 192 kHz/32bit you can listen to 8 channels. You can find customer experiences here.
If I may... I hope Axa will not object that I like posting here
I feel need to update you on the latest I find out regarding the problem I outlined here before. The problem was that I could not find software crossover for OSX so I did it through the use of Jack and Logic. Crossover within Pure Music was fine but I needed something else in order to play streaming music or simply just to pop CD, since I was not able to do it in Pure Music.
Now life gets much simpler. Pure Music implemented play through option, which functions sort of like Jack. It allows you to select Pure Music as an output device. That way you could use any other player to play through Pure Music, where you might have set up crossover and Eq filters. That is much easier and simpler than what I was doing before with Jack and Logic. That way you could enjoy your digital files, streaming and CD all through the same set up on Pure Music player. We all do that, we enjoy our Flac files, but we also need simple background music as well.
So Axa, here is one more voice for OSX driver, hehe
AR2, thank you for the update. I will impingement your configuration in my test environment.
exa065
I read the thread, its an interesting product but so geared toward the latest S-D DAC's at the exclusion of the more popular. I know its I2S but not really at only 32 bit output.
A huge segment of DIY DAC builders concerned about highest SQ don't use S-D DAC's, even the most famous DAC the TDA1541 has no support with this USB-"I2S" device. I mean I2S basically came to fruition around the TDA1541 and you can't use the Exa with the most popular I2S input DAC ever built?
A huge segment of DIY DAC builders concerned about highest SQ don't use S-D DAC's, even the most famous DAC the TDA1541 has no support with this USB-"I2S" device. I mean I2S basically came to fruition around the TDA1541 and you can't use the Exa with the most popular I2S input DAC ever built?
Last edited:
As far as I know, the exa USB2I2S is not limited to only 32-bit output. Also, there are many converters which are not Sigma-Delta which work with I2S.
Are you certain that the TDA1541 does not work?
Hopefully I am missing something, but my copy of the TDA1541A datasheet says data must be 16 bit. Reading thru this thread and the EXA webpage says that the exa will only output 32 bit data, this makes sense as this is how windows handles audio. A 16 bit flac would spit out of the exa as 16bit +16 zeros (per channel).
Now maybe after the 16th bit the TDA1541 would truncate the zeros spilling in until wordclock shifts, I don't know but the datasheet is pretty clear about a max 16 bit. And I would be concerned with how the LSB is handled if it is followed by 16 zeros. The TDA also has a max data input rate of 6Mbits/s.
Can't believe no one has tried this Exa with the grand father of I2S DAC's
Screw home theater, and digital crossovers, speakers too, I want jitter free TDA1541 conversion for my Grados
( I have a family that is alergic to loud music.)
Now maybe after the 16th bit the TDA1541 would truncate the zeros spilling in until wordclock shifts, I don't know but the datasheet is pretty clear about a max 16 bit. And I would be concerned with how the LSB is handled if it is followed by 16 zeros. The TDA also has a max data input rate of 6Mbits/s.
Can't believe no one has tried this Exa with the grand father of I2S DAC's
Screw home theater, and digital crossovers, speakers too, I want jitter free TDA1541 conversion for my Grados
( I have a family that is alergic to loud music.)Hopefully I am missing something, but my copy of the TDA1541A datasheet says data must be 16 bit. Reading thru this thread and the EXA webpage says that the exa will only output 32 bit data, this makes sense as this is how windows handles audio. A 16 bit flac would spit out of the exa as 16bit +16 zeros (per channel).
Now maybe after the 16th bit the TDA1541 would truncate the zeros spilling in until wordclock shifts, I don't know but the datasheet is pretty clear about a max 16 bit. And I would be concerned with how the LSB is handled if it is followed by 16 zeros. The TDA also has a max data input rate of 6Mbits/s.
Can't believe no one has tried this Exa with the grand father of I2S DAC's
Screw home theater, and digital crossovers, speakers too, I want jitter free TDA1541 conversion for my Grados
You play a 16bit file, then the output is 16bit, all depend on your file, except you set your player to up-bit/upsampling.
By the way, if you want 24bit to 32bit and over 44.1K, it is already nothing should related to TDA1541 or any famous old chips like also PCM63, it is non-sense to compare current HD CAS requirement together with typical CD (16/44,1) specifications.
The exa data is always 64fs, so the word length is always 32 bit package per channel, it may contain only 16 bits of data but as I said earlier there will be 16 bits of zeros that I am not sure how the IS2 input TDA1541 will handle, probably just waits till the next latch and has no harmful effects as data isn't timing. However it would be nice to hear experience of success from someone or Exa before purchasing.
Also you mentioned the PCM63K, this will not interface directly with the PCM63K unless you accept setting up a separate L and R channel and the resulting upsampling since the EXA is sending 32 bit Lchannel-32-bits L channel-... and 32 bit Rchannel-32-bits Rchannel, instead of the non-upsampled method of inversion of wordclock and required datashift (much preferred for NOS builders.) I would think this could be fixed with the software driver.
Also you mentioned the PCM63K, this will not interface directly with the PCM63K unless you accept setting up a separate L and R channel and the resulting upsampling since the EXA is sending 32 bit Lchannel-32-bits L channel-... and 32 bit Rchannel-32-bits Rchannel, instead of the non-upsampled method of inversion of wordclock and required datashift (much preferred for NOS builders.) I would think this could be fixed with the software driver.
The I2S standard allows for the use of different bit depth on the transmitting and the receiving end. The standard takes care of maximizing the amount of information delivered.
Specifically for exaU2I, the transmitting end is always 32bit. A 16 bit source will be extended with zeros. (assuming pure source - no volume control or other software processing is used and the LSB extension feature of the driver is off) On the receiving end, the 16 bit DAC will receive the first 16 bits, and than it will wait for the next word.
The 32bit envelope size has nothing to do with Windows. With the ASIO driver we bypass the entire Windows sound system. It is simply one size fits all - and without compromising anything.
exa065
Specifically for exaU2I, the transmitting end is always 32bit. A 16 bit source will be extended with zeros. (assuming pure source - no volume control or other software processing is used and the LSB extension feature of the driver is off) On the receiving end, the 16 bit DAC will receive the first 16 bits, and than it will wait for the next word.
The 32bit envelope size has nothing to do with Windows. With the ASIO driver we bypass the entire Windows sound system. It is simply one size fits all - and without compromising anything.
exa065
64*fs
It is 64*fs. You shouls be able to go up to 96kHz. 96000*64=6.14MHz. Sorry, I haven't tested TDA1541.
Cheers,
exa065
It is 64*fs. You shouls be able to go up to 96kHz. 96000*64=6.14MHz. Sorry, I haven't tested TDA1541.
Cheers,
exa065
Isn't the data format from the exaU2I the same as any other USB->I2S interface? The TDA1541 supports the standard 24-bit I2S signal, and converts only the first 16 bits of the data stream, so there's no need for glue logic. Other R2R dacs like the PCM63 or AD1862 would need additional cirquitry.
I know people are running the TDA1541 at 192 kHz sample rate (Audial, ECdesigns). The TDA1541 is more sensitive to jitter at higher sampling rates though. It's hard to predict how good it will be. You could be a pioneer. Be the first to try out the exaU2I + TDA1541 combo.
I'm running the Wolfson WM8741 myself, which is less jitter tolerant than the Sabre chip. The exaU2I + WM8741 combination is fantastic.
I know people are running the TDA1541 at 192 kHz sample rate (Audial, ECdesigns). The TDA1541 is more sensitive to jitter at higher sampling rates though. It's hard to predict how good it will be. You could be a pioneer. Be the first to try out the exaU2I + TDA1541 combo.
I'm running the Wolfson WM8741 myself, which is less jitter tolerant than the Sabre chip. The exaU2I + WM8741 combination is fantastic.
Sounds like it should work.
I have a clue from another Asynchrous USB converter on how to deal with the jittery galvanic isolation issue for you guys using the WM8741 (which I will try when I can select all the filters via software mode without needing a course in programming) and the ESS dac.
I have a few clues on how to get rid of the isolators and deal with the grounded by looking at my higher end commercial asynch USB-SPDIF transport.
I guess what I am saying is the EXA has a lot of potential. Obviosly I don't want to hack up my 3ps USB device, but the EXA looks like a nice platform to build from, I am hoping that software/firware support goes beyond the WM and ESS dac's, with the programming knowledge of the designer the flexibility is there to have compatibility with many more well thought of DAC's, just a suggestion. As an engineer (in an entirely different industry) never rest on your laurels Can you imagine all the AudioNote devote's who would jump on this if it supported the AD1865?
I will probably take one for the team with the TDA1541, in shear magnitude there are probably 10,000 to 1 TDA154x DIY audiphile owners to ESS owners.
Hoping that once the OSX driver is made there will be support for more DAC's. Features like true Left and Right channel data output (not Lx2 Rx2), little things like that could make this thing sell off the charts.
Another would be a selectible USB2.0 12Mbits/s mode, wherein you lose a the highrez but you can easily galvanically isolate on the USB side. You could literally build a NOS DAC with as low of Jitter as the SD players yet still retain all the wonderful luxuries of computer playback (VST's, dither, decoding HDCD, instant playback from 3TB HDD archives.) This is the future, not some SD card player that is limited to spitting out a few flac's or a wav's.
Isn't it ironic how we have two camps: one with near perfect timing SD transports feeding old DAC's and only RBCD but with practically no jitter, the other camp where low jitter/rf digital design fundamentals are sacrificed for ultra hirez but it has all the computer luxuries and flexibility. This product could do both and be a real advancement for a lot of DAC builders. It would be a mistake let the design stagnate with only ESS and WM (basically TP modules.)
I have a clue from another Asynchrous USB converter on how to deal with the jittery galvanic isolation issue for you guys using the WM8741 (which I will try when I can select all the filters via software mode without needing a course in programming) and the ESS dac.
I have a few clues on how to get rid of the isolators and deal with the grounded by looking at my higher end commercial asynch USB-SPDIF transport.
I guess what I am saying is the EXA has a lot of potential. Obviosly I don't want to hack up my 3ps USB device, but the EXA looks like a nice platform to build from, I am hoping that software/firware support goes beyond the WM and ESS dac's, with the programming knowledge of the designer the flexibility is there to have compatibility with many more well thought of DAC's, just a suggestion. As an engineer (in an entirely different industry) never rest on your laurels
Can you imagine all the AudioNote devote's who would jump on this if it supported the AD1865? I will probably take one for the team with the TDA1541, in shear magnitude there are probably 10,000 to 1 TDA154x DIY audiphile owners to ESS owners.
Hoping that once the OSX driver is made there will be support for more DAC's. Features like true Left and Right channel data output (not Lx2 Rx2), little things like that could make this thing sell off the charts.
Another would be a selectible USB2.0 12Mbits/s mode, wherein you lose a the highrez but you can easily galvanically isolate on the USB side. You could literally build a NOS DAC with as low of Jitter as the SD players yet still retain all the wonderful luxuries of computer playback (VST's, dither, decoding HDCD, instant playback from 3TB HDD archives.) This is the future, not some SD card player that is limited to spitting out a few flac's or a wav's.
Isn't it ironic how we have two camps: one with near perfect timing SD transports feeding old DAC's and only RBCD but with practically no jitter, the other camp where low jitter/rf digital design fundamentals are sacrificed for ultra hirez but it has all the computer luxuries and flexibility. This product could do both and be a real advancement for a lot of DAC builders. It would be a mistake let the design stagnate with only ESS and WM (basically TP modules.)
Hi regal,
Let's talk about the sacrificed fundamentals. I am interested to participate in discussion backed by facts. For example what is the 3ps USB-SPDIF device? How was 3 ps measured? Was that 3 ps on the device output or on the input of the DAC with the jitter caused by the cable and the receiver end taken into account?
Let's also make the jitter discussion factual. What else is out there? We need both independent measurements and subjective tests.
Jitter is one of many dimensions. It cannot be taken out of context, except in marketing talk. A third party not affiliated with vendors should compare the overall performance of the competing devices.
Let's also be factual about what is sacrificed for ultra high-resolution. I often hear that multichannel is sacrifice, hi-res is sacrifice... Please give me the facts and please base your subjective evaluation on first-hand experience with exaU2I.
Regards,
exa065
Just got a friend's feedback on the ExaU2I and he says it is better than all the other USB converters he has used, including his own 2707 designs, hiface and hiface Evo and some unnamed ones. Not much comment on the sound, just saying it has a much wider soundstage, better separation between instruments and music sounds flowing. I do not know what that means, he just says he is not giving it back to me 'cause "it is in a different league than all the rest". But what he was most excited about was how stable that device is - plug/unplug/start/stop - it just works which for him is quite impressive. he had all sort of trouble with the other devices and my Young DAC and the software playing tricks and crushing - he loves the ExaU2I and says that now he believes me that computer audio is the future And one thing that is important for him - he says that the Exa is very easy on the processor which for him has two big positive sides - his silent computer does feel better and requires less energy and his software is more stable, personal thanks from him on that. Just sharing, I still expect to hear it at home with the ESS9018, hopefully that will happen soon as I am dying to hear it myself
And one thing that is important for him - he says that the Exa is very easy on the processor which for him has two big positive sides - his silent computer does feel better and requires less energy and his software is more stable, personal thanks from him on that. Just sharing, I still expect to hear it at home with the ESS9018, hopefully that will happen soon as I am dying to hear it myselfHi tuyen,
There are several stories about powering the exaU2I and they are all true. The relaxed and careless attitude of the developers left the board with only two power sources by default. The USB, FPGA core and the clocks are powered from the USB cable with applying purification using linear regulators and capacitors. The exaU2I output must be powered by the DAC power supply. We thought two power sources with galvanic isolation is stare of the art.
Demanding DIY skeptics questioned the conventional wisdom and removed the linear regulators to provide independent power to the FPGA core and the clocks. These mods are illustrated here in this thread and on our blog on www.exaDevices.com.
One should start with the default configuration, develop his own experiments and hopefully share his experience. I think it is important to compare the impact of more independent power sources with the baseline configuration. In my own environment I use and like the default configuration.
Maybe I am missing the action
.exa065
Last edited:
Stable I2S USB digital playback & DSD playback (Watch out Mytek!)
Audio Friends:
I have posted a performance review following a seven week evaluation of the ExaU2l I2S USB interface for I2S USB playback. During my evaluation ExaDevices introduced DSD USBDAC playback utilizing ExaSound loTech DSDDIFF (proprietary) player which I also review. The beta version is a free upgrade.
The full review is posted here:
ExaDevices exaU2l USB-I2S interface: stable next generation I2S USB digital playback and DSD playback (Watch out Mytek!)
Summary:
The ExaU2l I2S USB interface is my new reference for digital playback, providing stable I2S USB playback with a meaningful performance improvement. ExaU2l plays WAV 32 bit 192Hz files (and much higher) with I2S. (Note to current users!) My experimentation confirms DSD playback provides additional musical enjoyment. With the ExaU2l audio enthusiasts are closer than ever to embracing digital playback without a need to continually compare Analog playback to Digital playback.
(BTW, Nikola I agree with your friend, "it is in a different league than all the rest"; I wouldn't give it back either.)
Best,
Bob
(drspence)
Audio Friends:
I have posted a performance review following a seven week evaluation of the ExaU2l I2S USB interface for I2S USB playback. During my evaluation ExaDevices introduced DSD USBDAC playback utilizing ExaSound loTech DSDDIFF (proprietary) player which I also review. The beta version is a free upgrade.
The full review is posted here:
ExaDevices exaU2l USB-I2S interface: stable next generation I2S USB digital playback and DSD playback (Watch out Mytek!)
Summary:
The ExaU2l I2S USB interface is my new reference for digital playback, providing stable I2S USB playback with a meaningful performance improvement. ExaU2l plays WAV 32 bit 192Hz files (and much higher) with I2S. (Note to current users!) My experimentation confirms DSD playback provides additional musical enjoyment. With the ExaU2l audio enthusiasts are closer than ever to embracing digital playback without a need to continually compare Analog playback to Digital playback.
(BTW, Nikola I agree with your friend, "it is in a different league than all the rest"; I wouldn't give it back either.)
Best,
Bob
(drspence)
