Popular Post mansr Posted March 17, 2019 Popular Post Share Posted March 17, 2019 A while ago, iFi released firmware version 5.30C for most of the their DAC models. The big news in this update is what they call the GTO filter. This filter is, according to iFi, so good that they produced a 9-page PDF espousing its virtues with much technobabble and several outright falsehoods. They even "feel" this filter is so fantastic that they decided to replace all the previously available filter choices with this one. In other words, the little filter selection switch on the DAC now does nothing at all. In order to get a closer look at this magical filter, I connected a logic analyser to the inputs on the DSD1793 DAC chip. With this new firmware, the chip is put in filter bypass mode, the XMOS now upsampling everything to 352.8/384 kHz. Only when the input already has one of these rates is it left untouched. While preparing to capture the GTO filter impulse response, I noticed that non-zero data was being sent to the DAC chip even with no input over USB. As it turns out, the firmware is truncating the samples to 18 bits with shaped dither. The figure below shows the spectrum of this dither along with TPDF dither at 16 and 22 bits for comparison. As we can see from the graph, the dynamic range is equivalent to 22 bits for frequencies below about 24 kHz. The noise level then rises, at 96 kHz reaching a level corresponding to 16 bits, then remains flat. This noise spectrum is the same whether the input is 48 kHz, 96 kHz, or 192 kHz. With 384 kHz input, no dither is applied. The analogue noise level of the DSD1793 falls somewhere between the high and low levels seen here. This means the added dither noise somewhat reduces the dynamic range at high frequencies. For some reason, no mention of this is made by iFi. The presence of the dither noise interferes somewhat with capturing the impulse response of the GTO filter, but by averaging a large number of responses, the effect of it can be minimised. Starting with a 48 kHz input, we obtain this impulse response: And the frequency response: The filter is close to minimum phase, but it does not meet the strict definition as a few of its zeros are outside the complex unit circle, not that this is important. More interestingly, it has a slow roll-off reaching only -2 dB by 24 kHz and -90 dB at 48 kHz. It then stays (mostly) below -80 dB until the vicinity of 150 kHz where it shoots up to -40 dB before falling off again towards the end of the spectrum. Besides an overall unimpressive performance, any high-frequency content in the 48 kHz original will see poor attenuation of its images, both immediately above 24 kHz as well as around 168 kHz. This can lead to problems with intermodulation distortion. To check for IMD resulting from the poor upsampling filter, we use a single 22 kHz tone at -3 dBFS. First, the spectrum on the analogue output when using a 384 kHz sample rate, bypassing the GTO filter: Not the prettiest spectrum, but at least none of the distortion products are audible. Next we send the tone using a 48 kHz sample rate. This may not be the ugliest spectrum I've seen, but it is definitely in the running. These distortion products are plainly audible as a nasty, harsh whine. On closer inspection, it turns out that the filter is clipping when presented with this input, and this contributes to the distortion. Lowering the amplitude to -4 dBFS avoids the clipping. Below, the resulting spectrum of a digital capture of the filter output. All the images are as expected from the filter frequency response. The shaped dither noise is also clearly visible. Additionally, there are some smaller spikes at 2 kHz intervals below the primary one. These are probably due to rounding errors in the filter implementation. The spectrum on the analogue output is now much cleaner: Cleaner, but still far from clean. Compared to the digital signal, harmonic and intermodulation distortion, rather than images, dominate the picture. This includes a readily audible component at 4 kHz. Moving up to 96 kHz, the impulse response looks like this: And thus the frequency response: All the characteristics remain. Minimum phase like, slow roll-off, poor stopband attenuation, and an even bigger hump towards the end. Finally, at 192 kHz, the impulse response has shrunk to a mere 5 samples: The frequency response is equally simple-looking: That's barely even a filter at all. Reaching a verdict on this update is easy. GTFO: Get This Filter Out of my DAC. Currawong, MikeyFresh, ZaphodStyle and 12 others 7 1 6 1 Link to comment
Popular Post Miska Posted March 17, 2019 Popular Post Share Posted March 17, 2019 The filter doesn't seem so far from the MQA filters. And it was introduced at the same time with MQA support. So I'm wondering how much there is actually common ground between the two... ZaphodStyle, jhwalker and semente 2 1 Signalyst - Developer of HQPlayer Pulse & Fidelity - Software Defined Amplifiers Link to comment
jhwalker Posted March 17, 2019 Share Posted March 17, 2019 Just now, Miska said: The filter doesn't seem so far from the MQA filters. And it was introduced at the same time with MQA support. So I'm wondering how much there is actually common ground between the two... I remember some discussion at the time it was released about it being (basically) an MQA-type filter; i.e., very "leaky" and little attenuation. Given we know their engineers like "bit perfect" mode (i.e., no filtering at all), not surprised they have provided this and think it sounds good. John Walker - IT Executive Headphone - SonicTransporter i9 running Roon Server > Netgear Orbi > Blue Jeans Cable Ethernet > mRendu Roon endpoint > Topping D90 > Topping A90d > Dan Clark Expanse / HiFiMan H6SE v2 / HiFiman Arya Stealth Home Theater / Music -SonicTransporter i9 running Roon Server > Netgear Orbi > Blue Jeans Cable HDMI > Denon X3700h > Anthem Amp for front channels > Revel F208-based 5.2.4 Atmos speaker system Link to comment
mansr Posted March 17, 2019 Author Share Posted March 17, 2019 16 minutes ago, Miska said: The filter doesn't seem so far from the MQA filters. And it was introduced at the same time with MQA support. So I'm wondering how much there is actually common ground between the two... They say it "was developed by iFi according to our specifications in conjunction with the MQA team." Link to comment
Popular Post asdf1000 Posted March 17, 2019 Popular Post Share Posted March 17, 2019 32 minutes ago, mansr said: They say it "was developed by iFi according to our specifications in conjunction with the MQA team." 51 minutes ago, Miska said: So I'm wondering how much there is actually common ground between the two... And they knew some of you would wonder this so they added: "We must make clear that GTO is not directly related to filter types used by MQA, it is not “MQA through the backdoor” Veri, MikeyFresh, crenca and 1 other 1 3 Link to comment
Popular Post Miska Posted March 18, 2019 Popular Post Share Posted March 18, 2019 So practically it is modified variant of the default MQA upsampling filter that for example Mytek runs for all PCM when MQA decoding is enabled... Likely on ESS-based DACs the MQA upsampling filter is just uploaded to the ESS chip, but in this case that is not possible with the TI DAC chip, so it needs to be run on XMOS. So this is functionally closer to how Meridian does in the Explorer2 DAC. crenca, ZaphodStyle, semente and 3 others 4 1 1 Signalyst - Developer of HQPlayer Pulse & Fidelity - Software Defined Amplifiers Link to comment
Popular Post mansr Posted March 18, 2019 Author Popular Post Share Posted March 18, 2019 13 hours ago, Em2016 said: And they knew some of you would wonder this so they added: "We must make clear that GTO is not directly related to filter types used by MQA, it is not “MQA through the backdoor” How is it then that the 96 kHz and 192 kHz versions are exactly the same as MQA filters found in the Bluesound firmware? I don't immediately see a direct match for the 48 kHz version, but the MQA decoding wouldn't normally need such a filter either. ZaphodStyle and MikeyFresh 2 Link to comment
asdf1000 Posted March 18, 2019 Share Posted March 18, 2019 5 minutes ago, mansr said: How is it then that the 96 kHz and 192 kHz versions are exactly the same as MQA filters found in the Bluesound firmware? It is MQA through the backdoor? 😧 The good thing about the current iFi DACs is you have options to use other non-MQA related filters, so not too big a deal. Hopefully it remains that way for future products. Link to comment
Popular Post mansr Posted March 18, 2019 Author Popular Post Share Posted March 18, 2019 12 minutes ago, Em2016 said: The good thing about the current iFi DACs is you have options to use other non-MQA related filters, so not too big a deal. By using the older firmware, yes. Will this work on their next DAC, whatever it turns out to be? I'm doubtful. Jud, MikeyFresh and Veri 2 1 Link to comment
AMR/iFi audio Posted March 18, 2019 Share Posted March 18, 2019 @mansr with regards to measurements, thank you for penning a very in-depth article. We have done many firmware updates and our 5.30C is just another one. We always like to give our customers a choice. If you like what a given FW does, please use it. If you don't, then don't. Just like Bit-Perfect, we like this but we don't insist that our customers must use this too. So if someone doesn't like what they measure, we won't judge. Truly. ZaphodStyle 1 Our PowerStation is here: click me! Check out our Tidal MQA Set-up Guides below. Android (Renderer) MobileDesktop (Decoder) via USBDesktop (Decoder) via SPDIF Link to comment
Popular Post mansr Posted March 19, 2019 Author Popular Post Share Posted March 19, 2019 It turns out the horrible distortion from the 22 kHz tone is partly due to the filter clipping. I have update the OP with some more info. In case anyone wonders, decent filter implementations do not clip from a pure tone at -3 dBFS. MikeyFresh and Veri 2 Link to comment
Popular Post FredericV Posted March 19, 2019 Popular Post Share Posted March 19, 2019 The iFI filter almost looks like the filter documented in the second link of my signature, which is the bottom plot in this image: for the aliasing this filter causes, it's somewhere in this topic TBH I don't like this filter, it makes transients artificially tight, and the decay in instruments is shortened. I like Archimago's intermediate phase a lot more, and this is the more natural sounding filter. MikeyFresh and Veri 1 1 Designer of the 432 EVO music server and Linux specialist Discoverer of the independent open source sox based mqa playback method with optional one cycle postringing. Link to comment
crenca Posted March 19, 2019 Share Posted March 19, 2019 3 hours ago, AMR/iFi audio said: @mansr with regards to measurements, thank you for penning a very in-depth article. We have done many firmware updates and our 5.30C is just another one. We always like to give our customers a choice. If you like what a given FW does, please use it. If you don't, then don't. Just like Bit-Perfect, we like this but we don't insist that our customers must use this too. So if someone doesn't like what they measure, we won't judge. Truly. Just to be clear, if you want to avoid GTO filter then you have to download 5.3 (not 5.3C). If you want to avoid GTO & MQA altogether, you have to download "5.2 & 5.2A, 5.2B ‘Limoncello’" from here: http://ifi-audio.com/downloads/ MikeyFresh 1 Hey MQA, if it is not all $voodoo$, show us the math! Link to comment
asdf1000 Posted March 19, 2019 Share Posted March 19, 2019 1 minute ago, crenca said: Just to be clear, if you want to avoid GTO filter then you have to download 5.3 (not 5.3C). If you want to avoid GTO & MQA altogether, you have to download "5.2 & 5.2A, 5.2B ‘Limoncello’" from here: http://ifi-audio.com/downloads/ Yeh but none of that applies to Pro iDSD, their newest DAC. There are non-MQA and non-GTO filters to choose from. In addition to “bit perfect” and you can do digital filtering before the DAC if you want. Link to comment
Jud Posted March 19, 2019 Share Posted March 19, 2019 2 hours ago, crenca said: Just to be clear, if you want to avoid GTO filter then you have to download 5.3 (not 5.3C). If you want to avoid GTO & MQA altogether, you have to download "5.2 & 5.2A, 5.2B ‘Limoncello’" from here: http://ifi-audio.com/downloads/ Or you can upsample non-MQA material with Audirvana+, HQPlayer, or some other program (preferably to DSD, since IIRC according to Miska, the available sigma-delta modulators aren't great). Currawong 1 One never knows, do one? - Fats Waller The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein Computer, Audirvana -> optical Ethernet to Fitlet3 -> Fibbr Alpha Optical USB -> iFi NEO iDSD DAC -> Apollon Audio 1ET400A Mini (Purifi based) -> Vandersteen 3A Signature. Link to comment
Veri Posted March 19, 2019 Share Posted March 19, 2019 11 hours ago, AMR/iFi audio said: @mansr with regards to measurements, thank you for penning a very in-depth article. We have done many firmware updates and our 5.30C is just another one. We always like to give our customers a choice. If you like what a given FW does, please use it. If you don't, then don't. Just like Bit-Perfect, we like this but we don't insist that our customers must use this too. So if someone doesn't like what they measure, we won't judge. Truly. Hi iFi thanks for chiming in. Just a small remark in context of this topic -I hope we can keep non-MQA/GTO filter on future devices. Not everyone is a fan, if new firmware becomes the norm the old filters would be gone. -I do wonder about truncating the samples to 18 bits, since this is not documented anywhere. With people feeding the iFi with high quality HQPlayer upsampled content, it is a little unexpected it gets dithered down to 18 bits! Link to comment
Popular Post mansr Posted March 19, 2019 Author Popular Post Share Posted March 19, 2019 7 hours ago, Jud said: Or you can upsample non-MQA material with Audirvana+, HQPlayer, or some other program One problem. The DSD1793 works better with 192 kHz input than 384 kHz. This is because the 192 kHz input is upsampled 8x to 1536 kHz, greatly reducing the demands of the analogue filter (external to the chip). With the weak 2nd order filter of the Nano, this makes a real difference. MikeyFresh, Currawong, crenca and 2 others 1 4 Link to comment
Jud Posted March 19, 2019 Share Posted March 19, 2019 Is upsampling to, for example, DSD256 what you would describe as putting greater demands on the analog filter? One never knows, do one? - Fats Waller The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein Computer, Audirvana -> optical Ethernet to Fitlet3 -> Fibbr Alpha Optical USB -> iFi NEO iDSD DAC -> Apollon Audio 1ET400A Mini (Purifi based) -> Vandersteen 3A Signature. Link to comment
mansr Posted March 19, 2019 Author Share Posted March 19, 2019 9 minutes ago, Jud said: Is upsampling to, for example, DSD256 what you would describe as putting greater demands on the analog filter? I was talking about PCM only. DSD is a different kettle of fish with its own issues. Link to comment
asdf1000 Posted March 19, 2019 Share Posted March 19, 2019 30 minutes ago, mansr said: One problem. The DSD1793 works better with 192 kHz input than 384 kHz. This is because the 192 kHz input is upsampled 8x to 1536 kHz, greatly reducing the demands of the analogue filter (external to the chip). And what happens with 384kHz input? Up-sampled to what rate? Or it's not up-sampled to a higher rate? Link to comment
asdf1000 Posted March 19, 2019 Share Posted March 19, 2019 7 minutes ago, mansr said: I was talking about PCM only. DSD is a different kettle of fish with its own issues. Can you share what these issues are, with feeding DSD256 to your iFi DAC? Link to comment
mansr Posted March 19, 2019 Author Share Posted March 19, 2019 7 minutes ago, Em2016 said: And what happens with 384kHz input? Up-sampled to what rate? Or it's not up-sampled to a higher rate? 384 kHz input is only possible in filter bypass mode, so that isn't upsampled at all. asdf1000 1 Link to comment
Popular Post mansr Posted March 19, 2019 Author Popular Post Share Posted March 19, 2019 6 minutes ago, Em2016 said: Can you share what these issues are, with feeding DSD256 to your iFi DAC? DSD64 has enough modulator noise at low frequencies that it can cause audible distortion artefacts. DSD256 is reasonable. asdf1000 and Jud 1 1 Link to comment
crenca Posted March 19, 2019 Share Posted March 19, 2019 3 hours ago, mansr said: One problem. The DSD1793 works better with 192 kHz input than 384 kHz. This is because the 192 kHz input is upsampled 8x to 1536 kHz, greatly reducing the demands of the analogue filter (external to the chip). With the weak 2nd order filter of the Nano, this makes a real difference. Do you or anyone else know if the analogue filter in the Nano is the same as the one in the micro iDAC2 (non black label version(s) - and for that matter do the black lable version have a different analog filter)? I have a Nano I travel with, but I use the micro iDAC2 in my HP rig when I am not using some other DAC... Hey MQA, if it is not all $voodoo$, show us the math! Link to comment
Popular Post Shadders Posted March 19, 2019 Popular Post Share Posted March 19, 2019 I tell you what, i was looking at the iFi Dac literature, and their plots look remarkably like the Stereophile ones. I reckon Stereophile borrowed them from the iFi literature. When iFi finds out, i think they ain't gonna be happy. I mean, copying someone elses drawings, is an utter disgrace. It just goes to show, you can't trust some people. rando and crenca 2 Link to comment
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now