I am sorry.
There are 3 threads on this forum on SD players, from EC, Koon, & Bunpei.
You posted one single picture without any technical details.
I found it a bit odd, if not rude.
If you wish to share details with other DIYer, I still suggest you start a new thread.
Currently only Bunpei's solution will do 24/192, but not USB.
Also none will do left adjusted, MSB first, left & right simultaneous.
Regards,
Patrick
There are 3 threads on this forum on SD players, from EC, Koon, & Bunpei.
You posted one single picture without any technical details.
I found it a bit odd, if not rude.
If you wish to share details with other DIYer, I still suggest you start a new thread.
Currently only Bunpei's solution will do 24/192, but not USB.
Also none will do left adjusted, MSB first, left & right simultaneous.
Regards,
Patrick
What do you exactly mean?
(sorry, I don't want to pollute this thread that, despite its name, is specific to ECDesign's SD player, but lacking a specific thread...)
Guido
Please...
Please don't ask me in this thread !
Because someone send me this notice.
I'll make a new thread asap.
Thanks
Anadigit
Please don't ask me in this thread !
Because someone send me this notice.
I'll make a new thread asap.
Thanks
Anadigit
I hope this doesnt mean proprietary code from m2tech repurposed without permission? I cannot see them letting go of it to someone unless they had a binding commercial arrangement, which you could not have unless its a product. nothing would surprise me after exposing similar technology in what is fairly obviously a thread based on a product. to not know that was uncool doesnt make sense in any language
Hi Overm,
It seems the TDA1541A-MK2 module is the final version. It performs a lot better compared to the MK1 version and I am now replacing all MK1 versions I already sold so far.
I also tested the SD-transport with a Kingston (speed rating 4) 32Gb SDHC card. It works fine and approx. 56 CDs (WAV) fitted on this card. This basically means that the SD-transport is not just an alternative to an audiophile CD player, it's also an audiophile "CD changer" that can hold up to 56 CDs until now.
When using hi-res formats, only few albums would fit on a SD(HC) card and one would have to swap the card often.
Example:
8 Gb SD-card holds up to 13 CDs in 44.1/16 wav. The same card would just barely hold 2 albums in 192/24.
I also noticed that fake SD(HC) cards are sold on ebay:
http://www.mods-n-clocks.co.uk/?p=290
I tested one of these, it acts like a 32Gb card (no reported errors) but actual capacity is far less.
Before designing alternative recording devices it would be wiser to optimize the way recordings are made and processed prior to sampling.
It seems the TDA1541A-MK2 module is the final version. It performs a lot better compared to the MK1 version and I am now replacing all MK1 versions I already sold so far.
I also tested the SD-transport with a Kingston (speed rating 4) 32Gb SDHC card. It works fine and approx. 56 CDs (WAV) fitted on this card. This basically means that the SD-transport is not just an alternative to an audiophile CD player, it's also an audiophile "CD changer" that can hold up to 56 CDs until now.
When using hi-res formats, only few albums would fit on a SD(HC) card and one would have to swap the card often.
Example:
8 Gb SD-card holds up to 13 CDs in 44.1/16 wav. The same card would just barely hold 2 albums in 192/24.
I also noticed that fake SD(HC) cards are sold on ebay:
http://www.mods-n-clocks.co.uk/?p=290
I tested one of these, it acts like a 32Gb card (no reported errors) but actual capacity is far less.
Before designing alternative recording devices it would be wiser to optimize the way recordings are made and processed prior to sampling.
Hi ANADIGIT,
My aim and the reason I started this thread was designing an almost perfect digital audio source. Meaning a straight-forward bit-perfect digital audio source with lowest possible jitter and interference levels, I2S output, and slave clock operation.
It's relatively simple to make an SD-card player with lossless (hi-res) playback capability, LCD color display and USB / SPDIF interfaces. In fact, a number of these devices are already available.
If you read this thread you will see why I built the SD-transport like this.
IF you finally achieve ultra low jitter and noise levels, the real problems start. Each and every flaw in your connected audio set will now become clearly audible as it's no longer masked by jitter and interference.
This is one of the reasons why I had to spend so much time optimizing the design of the SD-player, and later integrated the whole audio set in one box (ISD player).
One could also go the other route and produce plenty of jitter and interference for the covering up flaws elsewhere in the audio set, but I don't think this is the way to go.
My aim and the reason I started this thread was designing an almost perfect digital audio source. Meaning a straight-forward bit-perfect digital audio source with lowest possible jitter and interference levels, I2S output, and slave clock operation.
It's relatively simple to make an SD-card player with lossless (hi-res) playback capability, LCD color display and USB / SPDIF interfaces. In fact, a number of these devices are already available.
If you read this thread you will see why I built the SD-transport like this.
IF you finally achieve ultra low jitter and noise levels, the real problems start. Each and every flaw in your connected audio set will now become clearly audible as it's no longer masked by jitter and interference.
This is one of the reasons why I had to spend so much time optimizing the design of the SD-player, and later integrated the whole audio set in one box (ISD player).
One could also go the other route and produce plenty of jitter and interference for the covering up flaws elsewhere in the audio set, but I don't think this is the way to go.
That's actually a real problem. These 1543/41 devices are so sensitive to jitter and environmental conditions, that it is almost impossible to make them work and to stabilize them at 100%.
John. Ever thought about thermal control of your DACs?
Even though you covered a lot of aspects. I am wondering if the issue about thermal drifts is covered at all.
Some top-end manufacturers do have a build-in thermal control to keep the conditions constant. I do not know if this all is marketing stuff, never did
an A/B comparison thermal control vs. non thermal control. From a physical perspective it makes sense of course.
The 1543/41 are obviously pretty sensitive to thermal drifts. I am just thinking about the 20 minutes warm-up phase - on my own 1543 DAC.
If I think about it. A very sensitive clock would also be suffering of thermal drifts.
I recently read a "marketing" article about a DAC, where they put everything, DAC etc. under a mu-metal shield to care for a. EMI/RFI and temperature control. I think Lavry is also doing it on his top EQ.
Maybe it's worth to think about it.
Cheers
What do you think the thermal time constant would be of a DIP8 package ?
Are you interested in jitter below 1Hz ?
Patrick
Are you interested in jitter below 1Hz ?
Patrick
I am not quite sure if I understand you correctly, but if I read your reaction, I think you mention just what I was thinking. It's not about the SD or just the ADC 'an sich' but how to turn the analog signal (from a microphone for example) into perfect sample sets, without losing information between the microphone and the digital samples (time variation here will also cause the 'Jitter effect' I think, won't it?)
The SD could just be a convient way to store the made samples.....
Hi soundcheck,
What I meant was that when a DAC offers very low jitter and interference levels (regardless of DAC chip brand and type), a situation is created where it's much easier to hear flaws, including those in the connected audio set. These flaws are almost inaudible when masked with sufficient jitter and interference.
This leads to the situation that after reducing both jitter and interference levels, perceived sound quality may even be worse. It will only improve after dealing with the flaws that now have become audible.
I designed my housings such way that temperature inside remains fairly constant, especially around the DAC module. Power amps are cooled through the aluminum housing.
This depends on how sensitive the circuits are to temperature changes. Slow temperature changes are usually no problem for audio signal processing.
TDA1543 and TDA1541A (10 LSBs) use passive current dividers. These current dividers do just that, they divide the output current of a reference current source. Philips used emitter scaling for this, this means that it is important to maintain close matching between the transistors that are part of these passive dividers. Transistor mismatch would lead to increased bit errors.
After powering-up TDA154x, chip temperature starts rising rapidly. During this time it is not guaranteed that all transistors are at the same temperature. After the chip has warmed up and chip temperature has settled, transistor temperature should be equal. This in turn leads to best matching / lowest bit errors. This is the reason why these chips need some time before sound quality is optimal.
When using multiple TDA1543 in parallel (stack), extra cooling (heatsinks) need to be added to prevent hot spots and over-heating of the chips at the center of the stack. These increase thermal time constant and resulting settling time.
TDA1543 should usually settle within few minutes due to small DIP8 package.
Most uncompensated crystal oscillators suffer from thermal drift. Two common methods for compensating for this thermal drift are placing the oscillator in a temperature controlled oven (OCXO), or by compensating for temperature drift (TCXO).
I use a very high Q hybrid crystal and optimized oscillator circuit for reducing both, thermal drift and jitter.
Mu metal attenuates (static) magnetic fields. and electrical fields, so it makes sense to use this material for screening.
But sooner or later the signal has to leave the unit. This is the moment where interference strikes again, despite the mu metal shield in the source. The connected mains-powered equipment will create ground-loops, causing even more problems. Even speaker interlinks can pick-up EMI that in turn could affect power amp performance.
What I meant was that when a DAC offers very low jitter and interference levels (regardless of DAC chip brand and type), a situation is created where it's much easier to hear flaws, including those in the connected audio set. These flaws are almost inaudible when masked with sufficient jitter and interference.
This leads to the situation that after reducing both jitter and interference levels, perceived sound quality may even be worse. It will only improve after dealing with the flaws that now have become audible.
I designed my housings such way that temperature inside remains fairly constant, especially around the DAC module. Power amps are cooled through the aluminum housing.
This depends on how sensitive the circuits are to temperature changes. Slow temperature changes are usually no problem for audio signal processing.
TDA1543 and TDA1541A (10 LSBs) use passive current dividers. These current dividers do just that, they divide the output current of a reference current source. Philips used emitter scaling for this, this means that it is important to maintain close matching between the transistors that are part of these passive dividers. Transistor mismatch would lead to increased bit errors.
After powering-up TDA154x, chip temperature starts rising rapidly. During this time it is not guaranteed that all transistors are at the same temperature. After the chip has warmed up and chip temperature has settled, transistor temperature should be equal. This in turn leads to best matching / lowest bit errors. This is the reason why these chips need some time before sound quality is optimal.
When using multiple TDA1543 in parallel (stack), extra cooling (heatsinks) need to be added to prevent hot spots and over-heating of the chips at the center of the stack. These increase thermal time constant and resulting settling time.
TDA1543 should usually settle within few minutes due to small DIP8 package.
Most uncompensated crystal oscillators suffer from thermal drift. Two common methods for compensating for this thermal drift are placing the oscillator in a temperature controlled oven (OCXO), or by compensating for temperature drift (TCXO).
I use a very high Q hybrid crystal and optimized oscillator circuit for reducing both, thermal drift and jitter.
Mu metal attenuates (static) magnetic fields. and electrical fields, so it makes sense to use this material for screening.
But sooner or later the signal has to leave the unit. This is the moment where interference strikes again, despite the mu metal shield in the source. The connected mains-powered equipment will create ground-loops, causing even more problems. Even speaker interlinks can pick-up EMI that in turn could affect power amp performance.
Hi Overm,
If the audio signal is already degraded before sampling, even the best A/D converter won't help.
Practical example, you are listening to a symphony orchestra in a concert hall in the front row. You will hear instruments sound and how these sounds inter-act with the concert hall (room acoustics). In order to record all of this, you could place a suitable stereo microphone at the location where you were sitting, this would produce closest match to what you actually hear when sitting on that same location.
Modern recording techniques use many microphones placed near the instruments. At these locations, acoustics are completely different from the place you were listening. Then the signals of all microphones are mixed together. But since each mike picks up different acoustics, the mixed signals will result in distorted (smeared) room acoustics information.
Other problem is the hum and interference picked-up by all microphones, interlinks and connected mains powered studio equipment, and the many ground loops.
Why not use a suitable stereo microphone and perform A/D conversion and data storage (solid-state memory) as close as possible to this microphone. One could use a local extreme low jitter clock for timing. The unit can be easily battery powered, completely eliminating mains hum and interference. The unit could be placed in a suitable screen (mu-metal) for attenuating EMI and magnetic interference and it could be suspended in such way that unwanted LF and subsonic resonances (concert hall floor) are rejected.
Regarding A/D converter I am thinking of 16 bit SAR ADCs and high sample rate like used in this design:
The Altmann Creation ADC
If the audio signal is already degraded before sampling, even the best A/D converter won't help.
Practical example, you are listening to a symphony orchestra in a concert hall in the front row. You will hear instruments sound and how these sounds inter-act with the concert hall (room acoustics). In order to record all of this, you could place a suitable stereo microphone at the location where you were sitting, this would produce closest match to what you actually hear when sitting on that same location.
Modern recording techniques use many microphones placed near the instruments. At these locations, acoustics are completely different from the place you were listening. Then the signals of all microphones are mixed together. But since each mike picks up different acoustics, the mixed signals will result in distorted (smeared) room acoustics information.
Other problem is the hum and interference picked-up by all microphones, interlinks and connected mains powered studio equipment, and the many ground loops.
Why not use a suitable stereo microphone and perform A/D conversion and data storage (solid-state memory) as close as possible to this microphone. One could use a local extreme low jitter clock for timing. The unit can be easily battery powered, completely eliminating mains hum and interference. The unit could be placed in a suitable screen (mu-metal) for attenuating EMI and magnetic interference and it could be suspended in such way that unwanted LF and subsonic resonances (concert hall floor) are rejected.
Regarding A/D converter I am thinking of 16 bit SAR ADCs and high sample rate like used in this design:
The Altmann Creation ADC
This is true! I wish dear Onno was around here... 🙂
And a stereo mic that have the form of human head would be even better for stereo image 😉
Cheers,
M.
Ah, now we're talking! That's exactly what I was thinking of. I will look at the information in the link.Thanks.
I agree that the way music recordings are mostly made will not give a real reproduction.
I was thinking about a review years ago of the first 18 bit DAT recorder (philips). The reviewer had made a recording of an group of musicians. The musicians said after listening back, that it was the first time they could actually hear the room where the recording was made.
Is 16 bit enough or should we aim at 18 till maybe 20 bit? (more is completely unrealistic)
AD Conversion
I have a website
Onno Classics
Recordings is asking for a password.
That is
1Onno
Pse give your comment about the recordings. (sound quality not interpretation etc quality !)
Onno
I have a website
Onno Classics
Recordings is asking for a password.
That is
1Onno
Pse give your comment about the recordings. (sound quality not interpretation etc quality !)
Onno
- Status
- Not open for further replies.