Julf: oh hes definitely laughing, check out the special fluffy pink version on his webpage...
@ abraxilito
i'm glad you clarified your position mate, but its certainly a difficult line to walk, diagnostics are totally required to make a good dac IMO and that means numbers, you cant just do it by ear, but nor should the ear be avoided as an instrument.
its printed on the case.... and it doesnt have an asterix* and yes he has gone on record saying it can drive headphones to the fullest, yet its clear he has little knowledge of what that means... he dismisses many better headphones without having heard them in preference to his ideal headphone, which is a modified $30 Koss headphone. The Koss admittedly punches above its weight (it was the first headphone mod I did like 5 years ago), its not equal to the best available and totally useless for portable listening due to being completely open. which you can see with the pictures in his advertisement page
no Charles, hes not pushing them against his ears because he cant get enough of them, hes pushing them against his ears because its the only way you can get any bass/seal.
if you fully read the specification
and
perhaps he only found out after he got the cases printed? going by that logic my iriver supports 24/192 now that its got rockbox on it....
for me thats equivalent to including the potentially poisonous ingredients on a package for general consumption and then blaming the consumer for not reading the fine print. Certainly there are still a number of people who buy it thinking it plays hires if the user threads are anything to go by. many of them even after theyve been informed seem to think it sounds better 'playing' hi-res....
hard work? its a philips dac with an opamp IV stage and an opamp headphone 'amp' buffer. no doubt both are low power chips given the tiny supply voltage hes got. he makes a big deal about it being 'Burr Brown' very specific, wow quick let me get my cheque-book!
this sounds like a clever idea!
the funny thing is, I bet thats exactly what some of his customers do. Clearly some enjoy the sound, but I cant get past the smokescreen and the price is absurd, but it makes it seem exclusive and he will make a number of sales because of that.
anyway back to the topic, I guess I better go check out your blog, youve certainly set yourself a task =)
@ abraxilito
i'm glad you clarified your position mate, but its certainly a difficult line to walk, diagnostics are totally required to make a good dac IMO and that means numbers, you cant just do it by ear, but nor should the ear be avoided as an instrument.
its printed on the case.... and it doesnt have an asterix* and yes he has gone on record saying it can drive headphones to the fullest, yet its clear he has little knowledge of what that means... he dismisses many better headphones without having heard them in preference to his ideal headphone, which is a modified $30 Koss headphone. The Koss admittedly punches above its weight (it was the first headphone mod I did like 5 years ago), its not equal to the best available and totally useless for portable listening due to being completely open. which you can see with the pictures in his advertisement page
no Charles, hes not pushing them against his ears because he cant get enough of them, hes pushing them against his ears because its the only way you can get any bass/seal.
if you fully read the specification
nothing mentioning its only 16bit there. I dont think you can say it supports a file type it cannot actually use
modified hey? perhaps he added the extra 8 bits in there with a mod? I know the Germans are pretty good with technology....
and
hmmm. hes only recently included the dac type, didnt always mention the 16bit part, he was very secretive about it, simply saying R2R. I guess after a number of us made a bit of a deal about it he added it.
perhaps he only found out after he got the cases printed? going by that logic my iriver supports 24/192 now that its got rockbox on it....
for me thats equivalent to including the potentially poisonous ingredients on a package for general consumption and then blaming the consumer for not reading the fine print. Certainly there are still a number of people who buy it thinking it plays hires if the user threads are anything to go by. many of them even after theyve been informed seem to think it sounds better 'playing' hi-res....
hard work? its a philips dac with an opamp IV stage and an opamp headphone 'amp' buffer. no doubt both are low power chips given the tiny supply voltage hes got. he makes a big deal about it being 'Burr Brown' very specific, wow quick let me get my cheque-book!
this sounds like a clever idea!
why didnt we think of that? I can convert all my mp3 to WAV? maybe I should convert it to 24/192WAV while i'm at it?
the funny thing is, I bet thats exactly what some of his customers do. Clearly some enjoy the sound, but I cant get past the smokescreen and the price is absurd, but it makes it seem exclusive and he will make a number of sales because of that.
anyway back to the topic, I guess I better go check out your blog, youve certainly set yourself a task =)
Last edited:
50 dollar part and you're complaining
Oooh look yet another strawman. Where do you find them all - is there a field near to your home which produces all the raw material ?
Right - diagnostics is a good way of putting it. Its one of the things I mentioned to you in one of those PMs a couple of weeks back - the idea I had of closing the loop around a DAC+AMP by using an ADC to look at the output of the amp. Having an ADC there counts as a measurement doesn't it?
Well I read that Charles is saying 24bit .wav files can be read and played so I don't see there's a problem given that he clearly identifies that the DAC is 16bit. Incidentally its not R2R coz Philips to my knowledge has never made any. Its current segment.
I was actually thinking of the firmware when I wrote that, not the hardware part. Seems a fair bit of work to write it all in assembler. Has he said he's using an opamp for I/V then? Not very audiophile
You too lazy to use Google then?
Just took some time to watch this - very good, thanks.
I'm surprised they actually went to such extent WRT listening correlation.
A bare bones Sabre sans ASRC would have been nice. (yes I know you can turn
it off).
7th order elliptic anti-imaging filter
Here's something one or two of you guys might be interested in - a passive filter designed for NOS DACs. I've read a few reports and agree that NOS sounds a bit rough at the top end - I've wondered if this is due to amps or tweeters creating intermod products with the ultrasonics. So here's a design which I'm currently building to test my hypothesis - a 50dB stop-band passive anti-imaging filter. Ignore the 0.001 resistors, they're just there to see the effects of lossy inductors in running the sim.
Here's something one or two of you guys might be interested in - a passive filter designed for NOS DACs. I've read a few reports and agree that NOS sounds a bit rough at the top end - I've wondered if this is due to amps or tweeters creating intermod products with the ultrasonics. So here's a design which I'm currently building to test my hypothesis - a 50dB stop-band passive anti-imaging filter. Ignore the 0.001 resistors, they're just there to see the effects of lossy inductors in running the sim.
Attachments
Right - diagnostics is a good way of putting it. Its one of the things I mentioned to you in one of those PMs a couple of weeks back - the idea I had of closing the loop around a DAC+AMP by using an ADC to look at the output of the amp. Having an ADC there counts as a measurement doesn't it?
As part of the feedback loop? Have you looked at the time delay implications?
No, he doesn't. The owner's manual states "It will playback [sic] full resolution CD 44.1 kHz, DAT 48 kHz as well as DVD-A (high resolution) wav files up to 192 kHz with 16 bits as well as 24 bits". Nowhere does it mention that it doesn't actually play them back at anywhere near the "full resolution". It might not technically be fraud, but I would definitely consider it extremely misleading - and slimy. Is that the kind of "engineer" you want to be compared with?
I am sure you will declare this a straw man too.
No - the key word you seem to have glossed over here is 'diagnostics'. Go back and read again perhaps?
Not even a PCM1704 does, so the point is moot. Sure it accepts 24bits but the full resolution? Get outa here!
Compare me with Genghis Khan if you like, it wouldn't bother me
<edit> I agree on 2nd reading that the wording you quoted is misleading in respect of 'full resolution'. Black mark Mr Altmann.
Last edited:
OK, so by his (and your) definition even the humblest iPod "plays" high resolution music, as long as that music is converted to a suitable format. Sure. Whatever. At least we know where you stand.
Compare me with Genghis Khan if you like, it wouldn't bother me
Only reason I made the comparison was your (much) earlier question that you kept returning to:
Based on the evidence, I think I would call Charles a con man rather than an engineer. That's why I was wondering if you wanted to be compared to him, that's all. If you prefer Genghis Khan, be my guest, but to me it sounds a bit presumptuous.
Plays sure, but within the limits of its (lousy) DAC yep.
How curious - I never considered those two categories non-orthogonal.
Oh you ain't seen nothin' yet if you're at all sensitive to presumptuousness.
Just took some time to watch this - very good, thanks.
Thanks zen, gratifying to see at least one here with sufficient humility to learn something
Every project I have worked on inthe last 27 years has had to pass EMC. And the best EMC designs start by sorting it at the layout stage as much as possible.
You cant treat digital design or layout as analoge, its digital, and one of the most critical things is return current paths for digital. You say you are worried about currents going everywhere with a ground plane!!! I presume you have some inkling of how currents do flow Idf not:
Oh and a few other links that may be interesting, I have loads more info on both analogue, digital, mixed signal and RF layout if you arn't interested (pun intended)
http://www.x2y.com/filters/TechDay09... _JohnWu.pdf
Grounding of Mixes Signal Systems
http://www.ti.com/lit/ml/slyp167/slyp167.pdf
http://www.analog.com/static/importe...es/AN-1142.pdf .
How are you goiung to distribute your clock signal (this is a serious request) to all the devices. I recently worked on somthing similar to your Blog design, several tens of ADC (in this case) for a phased array thingy (cant say more, nod nod) so \i have some idea of the challenges you face.
And finaly a brief introduction to the ground plane:
www.elmac.co.uk/pdfs/Lord_of_the_board.pdf
You cant treat digital design or layout as analoge, its digital, and one of the most critical things is return current paths for digital. You say you are worried about currents going everywhere with a ground plane!!! I presume you have some inkling of how currents do flow
Idf not:Oh and a few other links that may be interesting, I have loads more info on both analogue, digital, mixed signal and RF layout if you arn't interested (pun intended)
http://www.x2y.com/filters/TechDay09... _JohnWu.pdf
Grounding of Mixes Signal Systems
http://www.ti.com/lit/ml/slyp167/slyp167.pdf
http://www.analog.com/static/importe...es/AN-1142.pdf .
How are you goiung to distribute your clock signal (this is a serious request) to all the devices. I recently worked on somthing similar to your Blog design, several tens of ADC (in this case) for a phased array thingy (cant say more, nod nod) so \i have some idea of the challenges you face.
And finaly a brief introduction to the ground plane:
www.elmac.co.uk/pdfs/Lord_of_the_board.pdf
16-bits
Hi there,
I am using a 16-bit Philips DAC in the Tera-Player, because I like the sound so much
I could have chosen any DAC chip.
Except some Hifi-man players, all existing DAPs are using sigma-delta DAC chips, and inherently they all fail, when it comes to sound quality.
The reason they fail, is because of the low-bit DAC at the output of a sigma-delta chip.
A sigma-delta DAC may be marketed as 24- or 32-bit but at the output there actually is a one-bit or very low-bit DAC only.
For example, the ESS Sabre chip as found in some newer DAPs only has a 6-bit DAC at the output.
http://www.esstech.com/PDF/ES9016_PF_091102.pdf
What about 32 bit DACs
In the Tera-Player we use a true 16-bit DAC, and that is the reason why it can sound so damn good, because it can playback a CD recording with full resolution without any tricks and gimmicks, and especially without the switching noise that is unavoidable with all sigma-delta DACs.
The Tera-Player can play wav-files up to 192 kHz sample-rate. It will playback 16-bit wav files, and it will playback 24-bit wav files through its 16-bit DAC.
When a 24-bit wav file is played, only the upper 16 bit are fed to the DAC. Now you could say, then it will only play 2/3 of the music, only 16 from 24bit, but that would be a wrong understanding: The lower 8 bit in a 24-bit wav-file are at -96dB and lower. What do you expect to be there ? Which ADC do you expect to give any significant resolution there ?
The volume control in the Tera-Player is based on the 32-bit division instruction of the ARM-processor in order to give a complete 16-bit result.
if you turn the volume of the Tera-Player to its very lowest setting, the output will still have 4 bit resolution. This will be extremely silent in your headphones, however then it is comparable to the resolution of a sigma-delta DAC running full-scale, and still without the switching noise.
Charles
Hi there,
I am using a 16-bit Philips DAC in the Tera-Player, because I like the sound so much
I could have chosen any DAC chip.
Except some Hifi-man players, all existing DAPs are using sigma-delta DAC chips, and inherently they all fail, when it comes to sound quality.
The reason they fail, is because of the low-bit DAC at the output of a sigma-delta chip.
A sigma-delta DAC may be marketed as 24- or 32-bit but at the output there actually is a one-bit or very low-bit DAC only.
For example, the ESS Sabre chip as found in some newer DAPs only has a 6-bit DAC at the output.
http://www.esstech.com/PDF/ES9016_PF_091102.pdf
What about 32 bit DACs
In the Tera-Player we use a true 16-bit DAC, and that is the reason why it can sound so damn good, because it can playback a CD recording with full resolution without any tricks and gimmicks, and especially without the switching noise that is unavoidable with all sigma-delta DACs.
The Tera-Player can play wav-files up to 192 kHz sample-rate. It will playback 16-bit wav files, and it will playback 24-bit wav files through its 16-bit DAC.
When a 24-bit wav file is played, only the upper 16 bit are fed to the DAC. Now you could say, then it will only play 2/3 of the music, only 16 from 24bit, but that would be a wrong understanding: The lower 8 bit in a 24-bit wav-file are at -96dB and lower. What do you expect to be there ? Which ADC do you expect to give any significant resolution there ?
The volume control in the Tera-Player is based on the 32-bit division instruction of the ARM-processor in order to give a complete 16-bit result.
if you turn the volume of the Tera-Player to its very lowest setting, the output will still have 4 bit resolution. This will be extremely silent in your headphones, however then it is comparable to the resolution of a sigma-delta DAC running full-scale, and still without the switching noise.
Charles
Charles,
Do you know how delta-sigma actually works? Have you ever heard of pulse density modulation?
24-bit means "with 24 bit resolution". 24-bit delta-sigma DACs actually achieve at least 20 bit resolution. Your 16-bit DAC doesn't.
And will it actually reproduce them at that sample rate?
So it will play back 16-bit files, and a 16-bit approximation of the 24-bit wav file. It will not reproduce the full 24 bits.
Sure, you can claim that the stuff beyond 16 bits is not very important, but that is a different issue than claiming you can play 24-bit files.
Any 24-bit delta-sigma DAC.
And what DSP doesn't use higher precision for intermediate results? If you do a division on 16 bit data, and then convert the result to 16 bits, you will have less than 16 bits of resolution, even if you use 1000000 bits in the intermediate calculation.
So it will operate as a 4-bit DAC. Great.
No, it won't. You clearly don't understand how delta-sigma and pulse density modulation works.
This thread is reaching new levels of absurdity.
Do you know how delta-sigma actually works? Have you ever heard of pulse density modulation?
24-bit means "with 24 bit resolution". 24-bit delta-sigma DACs actually achieve at least 20 bit resolution. Your 16-bit DAC doesn't.
And will it actually reproduce them at that sample rate?
So it will play back 16-bit files, and a 16-bit approximation of the 24-bit wav file. It will not reproduce the full 24 bits.
Sure, you can claim that the stuff beyond 16 bits is not very important, but that is a different issue than claiming you can play 24-bit files.
Any 24-bit delta-sigma DAC.
And what DSP doesn't use higher precision for intermediate results? If you do a division on 16 bit data, and then convert the result to 16 bits, you will have less than 16 bits of resolution, even if you use 1000000 bits in the intermediate calculation.
So it will operate as a 4-bit DAC. Great.
No, it won't. You clearly don't understand how delta-sigma and pulse density modulation works.
This thread is reaching new levels of absurdity.
Since that's not a reply to the question I asked I take it the original question gets the reply 'no'.
Yes, I agree the best designs have EMC considerations baked in from the get-go. I remain unconvinced by your 'you can't do X' claims though because you don't support them with any reasoning. You strike me as hot on the theory and the writings of Henry Ott (who I hold in very high regard) but perhaps your experience isn't so practical.
Hi Charles, welcome to the fray - you noticed your detractors were having a go at you then?
I am using a 16-bit Philips DAC in the Tera-Player, because I like the sound so much
You and me both Charles. Can't get enough of that multibit sound
There's one HiFi-man player which uses the TDA1543 - that fails at least because the layout's poor. I've helped one guy on here improve it though. Also the TDA1543 is a battery drainer.
You could make it a bit clearer though - from the quote that Julf posted earlier it was a bit misleading that 8 bits get lopped off when the file's 24bits. Do you dither incidentally when you truncate from 24 down to 16bits?
Great SOH btw with the fluffy pink dice, keep up the good work and hope the turntable project is going well
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.