Hello I have implemented successfully a TDA1543 and PCM56 DAC, with IV opamp filter.
The filter I have designed on LTspice have a flat response till 20kHz, and -18dB at 60kHz.
But when I measure it, it has -2.5dB at 20kHz with both DACs
Hence I have to compensate this by reducing by one order the IV filter.
Do you know if it is correct the natural HF roll-off of the two NOS DACS?
Similar roll off experience I found here (translate in google please):
Audio DA Converter SN-DAC0503
The filter I have designed on LTspice have a flat response till 20kHz, and -18dB at 60kHz.
But when I measure it, it has -2.5dB at 20kHz with both DACs
Hence I have to compensate this by reducing by one order the IV filter.
Do you know if it is correct the natural HF roll-off of the two NOS DACS?
Similar roll off experience I found here (translate in google please):
Audio DA Converter SN-DAC0503
Correct, first order S/H has sin(x)/x transfer function. I am planning to confirm this with a digital pink noise track on CD and an FFT without filtering in the analog domain.
That roll-off is normal for NOS DAC operation to exhibit. Theoretically, it should be -3.16dB @ 20KHz rather than -2.5dB, assuming a 44.1KHz NOS sample rate. The audible issue with the roll-off isn't so much the drop at 20KHz, but the fact that it colors down to 5KHz, two full octaves.
Technically, the roll-off stems from the SINC function response mask created by the sample-and-hold (stair stepped) output of the DAC chip, which then mildly suppresses the highest frequencies. This supression varies with the sample rate sent to the DAC quantizer unit. The higher the sample rate, the less audible the suppression, because it consequently gets shifted higher in frequency. By the way, digitally filtered DACs also exhibit this roll-off, but oversampling increases the sample rate, making it less audible. Any remaining roll-off is then EQ'd to flat within most digital audio oversampling filters.
Equalizing the roll-off is correct to do and improves the tonal balance. There are several way to perform the EQ in the analog stage, depending on your specific stage implementation.
Technically, the roll-off stems from the SINC function response mask created by the sample-and-hold (stair stepped) output of the DAC chip, which then mildly suppresses the highest frequencies. This supression varies with the sample rate sent to the DAC quantizer unit. The higher the sample rate, the less audible the suppression, because it consequently gets shifted higher in frequency. By the way, digitally filtered DACs also exhibit this roll-off, but oversampling increases the sample rate, making it less audible. Any remaining roll-off is then EQ'd to flat within most digital audio oversampling filters.
Equalizing the roll-off is correct to do and improves the tonal balance. There are several way to perform the EQ in the analog stage, depending on your specific stage implementation.
If you implement an LC (passive, 2nd order) low-pass filter you can adjust the damping (Q) and perhaps the corner frequency to give yourself a boost of around 3dB at 20kHz, followed by a roll-off above that.
Who can hear 20Khz?
My old ears don't respond to anything over 12K and that's from a signal generator at a power level never found in real life.
My old ears don't respond to anything over 12K and that's from a signal generator at a power level never found in real life.
Wouln't the staircase signal make the audio output rich in harmonics at higher frequencies, thus compensating the rolloff subjectively?
Who can hear 20Khz?
My old ears don't respond to anything over 12K and that's from a signal generator at a power level never found in real life.
Can they hear down to 5KHz? That's how low the roll-off extends down to.
Wouln't the staircase signal make the audio output rich in harmonics at higher frequencies, thus compensating the rolloff subjectively?
Yes, it does, but not rich enough. There needs to be more harmonic energy. Or, rather, think of the issue as an excess in lower frequency energy, due to the flat part of the staircase (the hold period), making the higher frequency energy appear relatively deficient.
Last edited:
Can they hear down to 5KHz? That's how low the roll-off extends down to.
Its -2.5db at 20K so won't be down much at 10k, let alone 5k.
I'd challenge anyone to do a blind listening test with a switchable filter. Theory does not always translate into audible differences in my experience.
Re: Do you know if it is correct the natural HF roll-off of the two NOS DACS? Yes, with measurements:
NOS vs. Digital Filtering DACs: Exploring filtering turned off, implications, fidelity and subjective audibility.
NOS vs. Digital Filtering DACs: Exploring filtering turned off, implications, fidelity and subjective audibility.
Its -2.5db at 20K so won't be down much at 10k, let alone 5k.
I'd challenge anyone to do a blind listening test with a switchable filter. Theory does not always translate into audible differences in my experience.
Okay, so you quickly moved from declaring it not audible at 20KHz, a notion with which I agree, to debating the possible audibility of an masking curve which spans and so alters two octaves, 20KHz down to 5KHz. It's your prerogative to doubt what you will. I won't waste further cyber ink trying to open you to consider the possibility that your assertion is faulty.
Cheers!
Re: Do you know if it is correct the natural HF roll-off of the two NOS DACS? Yes, with measurements:
NOS vs. Digital Filtering DACs: Exploring filtering turned off, implications, fidelity and subjective audibility.
They appear roughly correct. Down a touch more than -3dB @ 20KHz, with a smoothly rolled response (obviously not flat) spanning to around 5KHz.
Yes, to some extent. This may partly explain why the errors of NOS are nonetheless quite popular. The "harmonics" are actually images, which a reconstruction filter would remove in order to recover the signal which entered the ADC.lcsaszar said:Wouln't the staircase signal make the audio output rich in harmonics at higher frequencies, thus compensating the rolloff subjectively?
No. The -3dB point is part of the definition, apart from first-order filters which are defined by their -3dB point alone. The HF rolloff of a NOS DAC is not a first-order filter.
Warning for naive question!
I understand how the digital to analog conversion benefits from oversampling. But does the images filtration also have to happen before the conversion? My interpretation of this is that the images are byproducts of the sampling procedure. But where does this happen and how would affect the DAC chip itself? Is it possible -and meaningful- just to do oversampling prior to the DAC and then filter the analog output?
I understand how the digital to analog conversion benefits from oversampling. But does the images filtration also have to happen before the conversion? My interpretation of this is that the images are byproducts of the sampling procedure. But where does this happen and how would affect the DAC chip itself? Is it possible -and meaningful- just to do oversampling prior to the DAC and then filter the analog output?
Warning for naive question!
I understand how the digital to analog conversion benefits from oversampling. But does the images filtration also have to happen before the conversion? My interpretation of this is that the images are byproducts of the sampling procedure. But where does this happen and how would affect the DAC chip itself? Is it possible -and meaningful- just to do oversampling prior to the DAC and then filter the analog output?
If I correctly understand what you're asking, reconstruction filtering can be done in either the digital or the analog domains, or usually in both. Digital filters objectively perform much better than do analog filters, although digital filtering always leaves some residual out of band image products. These residual products can be completely removed via an relatively low order analog filter, or are sometimes simply ignored. Subjectively, there remains controversy over what some hear as the adverse affect of hardware based digital filters. Hence, the popularity of NOS (digital filter-less) DACs.
The old debate again between technically perfect but lifeless, and technically flawed but emotionally involving.
Rather than technically perfect, maybe we could say mathematically perfect, were it not for the presence of jitter. Of course, there remains controversy about the audibility of low level jitter effects.
Interestingly, Crane Song Solaris dac offers extremely low jitter (pages 13-14):
http://www.cranesong.com/solaris_manual_version_1.pdf
They say it matters and they think there will be audible benefits from going even lower. Of course they are using oversampling and digital filters. They don't sound lifeless at all at very low jitter levels, IME. Could be the dac is little too accurate for some people, but it is reported to have a sound very similar to Benchmark DAC-3.
Interestingly, Crane Song Solaris dac offers extremely low jitter (pages 13-14):
http://www.cranesong.com/solaris_manual_version_1.pdf
They say it matters and they think there will be audible benefits from going even lower. Of course they are using oversampling and digital filters. They don't sound lifeless at all at very low jitter levels, IME. Could be the dac is little too accurate for some people, but it is reported to have a sound very similar to Benchmark DAC-3.
Images are byproducts of the conversion back to analogue. Sampling produces aliases, which people often confuse with images.MagicBus said:But does the images filtration also have to happen before the conversion? My interpretation of this is that the images are byproducts of the sampling procedure. But where does this happen and how would affect the DAC chip itself? Is it possible -and meaningful- just to do oversampling prior to the DAC and then filter the analog output?
If oversampling is to happen then this must take place before the DAC; oversampling is a digital technique so the signal needs to be in the digital domain for this to happen.
The main purpose of oversampling is to achieve better reconstruction filtering; some of the filtering is done digitally, then the signal goes through the DAC, then there is a final analogue filter - which can be simpler than would be needed if it was doing the whole job itself.
NOS fans want to avoid all this precision, as they prefer the sound of rolled-off signal accompanied by images.
- Status
- Not open for further replies.
- Home
- Source & Line
- Digital Line Level
- Frequency response of NOS DAC