IMHO, supporting ALSA drivers will open for you the embedded Linux HTPC market.
I'm planning to use HTPC/Linux box as processing engine for active (FIR)x-over / delay / rms-limiter and room correction.
So, any hint on the release date?
Oh, and because we have discussed about using this interface with ESS9018, did you had the chance of using it with more mundane DAC's? Such as AK, AD, Burr-Brown parts? Any impressions on this? Thanks!
haha thats a bit harsh, although i totally agree
Oh, this was supposed to be an edit:
Since this thread is a little slow until news, I'll indulge in a bit off-topic. If I want to use the exa device with a mundane dac
, say AK4396, and I want to stream 24 bits 384Khz, can I do it? Or is this implying the use of an external digital filter?
Also, is the ESS9018 capable of sample rates over 192Khz without mods? I recall an input figure of around 1,6Mhz in a datasheet although I don;t know what was referring at exactly.
Since this thread is a little slow until news, I'll indulge in a bit off-topic. If I want to use the exa device with a mundane dac
, say AK4396, and I want to stream 24 bits 384Khz, can I do it? Or is this implying the use of an external digital filter?Also, is the ESS9018 capable of sample rates over 192Khz without mods? I recall an input figure of around 1,6Mhz in a datasheet although I don;t know what was referring at exactly.
A quick check of the AK4396 data sheet cites a maximum fs of 216 kHz, so I'd say: No, you can't do it.If I want to use the exa device with a mundane dac
An external digital filter would work, but I ask why would you bother? You could obtain superior digital filtering in non-real-time by converting any 384 kHz file you might have. But that begs the question of where you're going to get more than one or two audio files at 384 kHz that you'd want to hear with real-time SRC.
yes es901X series runs asynchronously at 1.4mhz if you can feed it 384 and your implementation is good, 384 is no problem at all, in fact it will be internally upsampled to match the higher haha bitrate and word length of the sabre internal clock. not much choice for toons though and anything you do find will simply have been upsampled, as nobody records at that speed
I have so far 22 albums that are recorded in DXD 352.8k/24bit Stereo.
The albums are in fact recorded in full DXD surround and I hope to be able to play surround in full DXD format soon.
Check out 2L co/ Lindberg Lyd 2L - the Nordic Sound as there are 11 tracks in DXD that you can download and many more in various Hi-Rez formats.
There are also other studios like Puget Sound Studio that records in DXD.
As all albums except the Nordic Sound album are sold by iTunes in lower resolutions - the DXD albums are recognized in iTunes and you get cover art etc.
Both iTunes and Pure Music plays the DXD files.
The only missing link are the exaU2I to get 352.8k out to the DAC
Currently I am using either a M2Tech USB unit and downsamples to 176.4k/32bit,
or a SDTrans192 SD card player and plays 352.8k with I2S directly to the DAC.
The ES9018 based DAC uses a 98.304MHz clock and the DAC outputs runs at 1.536MHz.
The DAC have an integrated SDTrans192 SD card player and have full IR remote control of the ES9018 chip and user control of the SD card palyer.
Attachments
Last edited:
Well the idea is to push the limits mainly.. Although I have a number of hi resolution files already (dxd, 384k) and I expect this number to grow.
Well, from what I get the internal stuff in a DAC already makes some upsampling (the digital filter etc.) So, a 192Khz pcm already gets to 384Khz when fed into the delta sigma modulator. So I guess that if the signal already is 384 or multiple, I might make it pass it through by bypassing the internal digital filter?
That's a reasonable assumption, and it might work with other chips. The AK4396 doesn't seem to have external access to the internal routing such that you could bypass anything. Also, the AK4396 operates with a 128x clock with Quad Speed rates (176.4 kHz and 192 kHz). For 384 kHz, that would be 49.152 MHz, but the AK4396 has a maximum of 41.472 MHz for that clock. Again, I don't think it's possible.
Another issue is that, technically, 384 kHz would be Octuple Speed; the AK4396 doesn't have that mode. You'd have to stretch Quad Speed to the top, and the top doesn't reach 384 kHz. It does seem to reach 324 kHz, but that's shy of 352.8 kHz.
There are nice, current-output DAC chips which can convert at 140 MHz, but the one I'm using is only 14-bit. The 24-bit converters seem to be in an area where the physics make it rather difficult to achieve actual 24-bit quality at ultra high rates.
I have so far 22 albums that are recorded in DXD 352.8k/24bit Stereo.
The albums are in fact recorded in full DXD surround and I hope to be able to play surround in full DXD format soon.
Check out 2L co/ Lindberg Lyd 2L - the Nordic Sound as there are 11 tracks in DXD that you can download and many more in various Hi-Rez formats.
There are also other studios like Puget Sound Studio that records in DXD.
As all albums except the Nordic Sound album are sold by iTunes in lower resolutions - the DXD albums are recognized in iTunes and you get cover art etc.
Both iTunes and Pure Music plays the DXD files.
The only missing link are the exaU2I to get 352.8k out to the DAC
Currently I am using either a M2Tech USB unit and downsamples to 176.4k/32bit,
or a SDTrans192 SD card player and plays 352.8k with I2S directly to the DAC.
The ES9018 based DAC uses a 98.304MHz clock and the DAC outputs runs at 1.536MHz.
The DAC have an integrated SDTrans192 SD card player and have full IR remote control of the ES9018 chip and user control of the SD card palyer.
are you sure they are actually recorded at that speed, or simply upsampled to that speed? no offense to the Nordic race, but i dont care if its recorded at 10mhz you wont catch me listening to it. i listen to music for enjoyment and do not tend to listen to all that much 'audiophile music' whatever that is; except for soe select classical and jazz; just because its high quality recording isnt enough for me. i'll take well produced 16/44 that i love over 32/384 just because i can.
ive seen the files at hdtracks etc and have a few, but there isnt enough information to gather whether or not its recorded or just supplied at 352khz? anything supplied at a multiple of 44.1 rather than 48 as any regular recording gear uses i'm automatically suspicious of and many of these artists simply were not around to be recorded like this and while analogue masters contain arguably more musical information in some areas of sound, the noise floor is high enough that the bandwidth will be effected. so remasters will likely be a bit hit and miss also.
all i'm saying is, even going by your numbers and i'm gathering you have searched high and low for music you like, 22 albums is bugger all, repeats pretty quickly, all the same i'm waiting with baited breath for the release of some more usic i enjoy at this level. i'm not arguing against the use of such hirez, just saying its not really such an advantage for me at this stage over 24-32/192, except for playback of my musical meanderings from logic audio. annoying that modern gear is capable of it, but because it represents the master the chances of release are slim. the multichannel is much bigger news to me
thanks for the catch on 1.536 vs 1.4mhz, i was still out of habit being deliberately vague with ess datasheet details, but i suppose it really doesnt matter anymore.
your dac is looking really good Raymond, but I couldnt be bothered listing out the spec of the dac i' using and its still under the knife, although kinda similar. id really like to hear your lovoltech jfet power amp one day
Last edited:
Yes I am sure!
As I wrote - the master recording are not only done in DXD, but DXD surround..
I have listened to the DXD surround masters in the 2L studio and compared to DXD stereo - due to that I would like to implement the ability to play DXD surround my self - then I can play all surround formats up to full DXD..
1.6875MHz are the highest I have been running the Sabre32, but at the time I tested I got some spurious noise issues and went back to the safe and working 1.536MHz.
Last edited:
cool, i' just never able to find that much info at hdtracks and like i said multiples of 44 make e suspicious. dxd surround is of great interest to me also eventually. but multichannel will only be used for 2 channel digital xo for me for now, thats exciting enough.
i'll have to go check out XL some more now that we upgraded the link here downloading that much is not an issue. i imagine the playback system there is pretty special
i'll have to go check out XL some more now that we upgraded the link here downloading that much is not an issue. i imagine the playback system there is pretty special
Last edited:
Hi,
Well I am basing my ramblings on some old datasheets I found regarding the Anagram upsampler. This module is used in some cambridge CD-players capable of 768Khz with AD1955 DACs as well as with some Wolfonson parts.
For example, in 8fs mode, the WM8741, seems capable of 352 or 384 Khz. (page 26)
Well I am basing my ramblings on some old datasheets I found regarding the Anagram upsampler. This module is used in some cambridge CD-players capable of 768Khz with AD1955 DACs as well as with some Wolfonson parts.
For example, in 8fs mode, the WM8741, seems capable of 352 or 384 Khz. (page 26)
Analog tape masters have a maximum bandwidth of maybe 30 kHz, so sampling at least 60 kHz will be beneficial, but sampling above 120 kHz is somewhat pointless unless you're only concerned with the anti-aliasing filter.
All 24-bit converters have noise, so the noise floor in original masters merely serves as excellent dither.
It's good to be suspicious of high sample rates, whether multiples of 44.1 kHz or 48 kHz. I have found many examples of recordings with no information other than quantization noise in the upper half of the available bandwidth, and why pay for a double-sized file when you can upconvert for yourself? What's worse, I've actually found high sample rate files with aliased frequencies above 44.1 kHz or 48 kHz, showing that they were converted with inferior technology. An aid to this sort of forensic analysis is to use linear frequency and extended dynamic range in the FFT analysis. It's much easier to detect aliasing visually when the frequency is on a linear scale.
That said, HDtracks does have some good recordings. Perhaps our tastes are different, but I went reading through the releases from artists that I already knew, and found a number of titles to purchase that were satisfying. I purchased a combination of new performances and original masters.
The one thing that bugs me is that exaU2I seems to not support 384 kHz, only 352.8 kHz. Why would the hardware go that far and not extend to 8x 48 kHz?
I would be a little suspicious of an upsampler named 'Anagram.' Does that mean that it aliases frequencies from the original content in order to fill the higher frequencies? Actually, the worst would be a hypothetical product named the 'Palindrome upsampler' - which would surely alias all of the original frequencies.
Sorry, I couldn't resist the pun ... especially since I have seen many so-called high-resolution audio files with aliased frequencies in the ultrasonic range.
Maybe it's limited by the bandwidth of the usb module?
Well yes, I see upsampling to 384 a little pointless, however some of things they did seemed to me quite clever. They are out of business, however I've found something based on the same tech, the datasheet is similar at least. Sonic Scrambler and Q5 Upsampler.
Also, on the products page I've noticed they have an USB interface based on XMOS!
Last edited: