here - TDL Tech
Contents
1. 1000 10000 20000 Frequency Hz Figure 8 Time delay com parison RIAA network to dbx231 EQ The dbx EQ does not have as good a fit to the RIAA curve as does the DOD equalizer Time delay mS 1000 10000 20000 Frequency Hz Figure 9 RIAA baseline and DEQ1024 EQ amplitude comparison The match us ing this digital equalizer is very good in the 20 Hz to 2500 Hz range Another DEQ1024 would be needed in series for comparison up to 20 kHz Amplitude dBu T 7T T 100 1000 10000 20000 Frequency Hz Digital vs Analog Equalization 241 of Disc Record Playback One listener commented as follows I first listened using my office computer s on board Realtek sound chip and the tiny speakers in my monitor I figured if I could hear a difference with a lo fi rig it would be pretty telling The difference was pretty clear to me Tracks 1 and 3 sounded dull and lifeless tracks 2 and 4 sounded like real instruments The differences were even more telling when I listened in my music room most notably in the area of bass definition All in all I would describe the sound of tracks 2 and 4 as more coherent This is apparent as a more natural sound for each instrument ease of distinguishing between instruments and a greater sense of ambiance Other listener comments followed suit the DOD EQ does not make the grade as a us2. 1000 10000 20000 Frequency Hz a nm ro a re Amplitude dBu eo a A ro 1000 10000 20000 Frequency Hz Figure 3 Low gain RIAA network for comparison baseline Only one channel was used for this project 1 resistors and matched ca pacitors produce a response that deviates no more than 0 2 dB from the defined values Photo 1 RIAA baseline cir cuit and power supply Any 5 to 8 VDC power supply can be used This low gain circuit does not need shield ing or low noise opamps Figure 4 RIAA baseline and DOD EQ amplitude compar ison The slightly wavy line is the DOD EQ response Digital vs Analog Equalization 239 of Disc Record Playback Figure 5 Time delay com parison of the baseline RIAA network with a DOD SR430Qx EQ adjusted to the RIAA amplitude re sponse using QuickSweep in TrueRTA na E i gt amp D Q E E 1000 10000 20000 Frequency Hz COMPUTER WITH SOUND CARD beste Mica THE RESTORATION PREAMP NETRONICS 350D NODE 4010 SHURE M97xE POWER AMP SPEAKERS MM CARTRIDGE AUX OUT AUX IN GROUND BREAK STEREO TO MONO DOD SR430QX EQUALIZER MODEL 414A CHANNELS IN SERIES SET GAIN AS NEEDED SET TO RIAA RESPONSE Figure 6 Equalizer test setup for listening and recording The model 4010 is set to flat Turnover and flat Rolloff Mode sw
3. after a free trial period Use your favorite search engine to look for VST EQ plug ins LinearPhase PEQ Red is a VST EQ plug in from Algorithmix GmbH You can visit their web site for product information and to re quest a link to download a trial version and the Operator s Manual a pdf file http www algorithmix com Sony Sound Forge version 10 is an audio recording and editing program available from http www sonycreativesoftware com and from many audio software distributors Galo Gary A Phase Equalization and its Importance in the Reproduction of Disc Records in ARSC Journal 2010 41 1 83 96 IIR Infinite Impulse Response and FIR Finite Impulse Response digital filter theory is beyond the scope of this paper but many texts are available For example see Williams Arthur B and Fred J Taylor Electronic Filter Design Handbook McGraw Hill Inc Second Edition 1988
4. and output an RIAA or one of several older record equalizations I chose 244 ARSC Journal QuickSweep Amplitude dBu pink noise 1000 10000 20000 Frequency Hz 0 4 0 3 0 2 0 1 0 0 1 Response in dB N N 0 4 0 54 20 1000 10000 20000 Frequency Hz Time Delay microseconds ___ e 1000 10000 20000 Frequency Hz Figure 13 PEQ Red output with TrueRTA QuickSweep and pink noise inputs All ten EQ filters were set to 0 dB gain and to the same Q bandwidth value Figure 14 Amplitude re sponse difference between the DC8 Virtual Phono Preamplifier and the calcu lated RIAA response The DC8 VPP was set to Flat input and RIAA output with the Tone Controls disabled Figure 15 RIAA time delay upper curve compared to the DC8 VPP set to Flat input and RIAA output with the Tone Controls disabled Digital vs Analog Equalization 245 of Disc Record Playback to test the VPP with a flat input and an RIAA output User setup is simple In addition to radio buttons for in and out selection it has check boxes for Tone Controls enabled or disabled and a 30 Hz Rumble Filter that can be turned on or off And there are sliders for Left Right Balance and Volume I recorded the same tracks used in the DOD EQ tests with the model 4010 pre amp set to L R flat turnover an
5. response when using digital EQ For example Gary Galo investigated this topic in a paper for the ARSC Journal 20 He too found the VPP in DC versions 6 and 7 had the correct phase response version 8 was not available at that time He found that correct phase response could be achieved with IIR digital filters but not with FIR filters which are still commonly used in digital EQ soft ware and apparently also in digital EQ hardware Some manufacturers were unable to tell him which type of filter was in use in their products so this is another case of using an item that has been proven to function correctly or else evaluating it yourself Ron Tipton is president and principal designer at TDL Technology Inc Las Cruces NM an electronics design and manufacturing company he started in 1957 He has degrees in electri cal engineering from New Mexico State University and he is retired from a design engineering position at White Sands Missile Range Ron has designed and built equipment for companies including Bose and Acoustic Research and worked as a consultant to GMS Inc Prairie Grove AR designing guitar amplifiers and effects modules for Gretsch He has authored con tributions to a number of specialist publications among them audioXpress Voice Coil Multi Media Manufacturing Poptronics Nuts amp Volts Magazine and Circuit Cellar Endnotes 1 The Recording Industry Association of 4 DOD isa manufacturer of hardware graphic America
6. were setup in Cantabile to cover the 40 dB amplitude range The GEQ81V has a digital readout to set the boost or cut for each fre quency band Figure 12 Amplitude re sponse of the KarmaFX VST EQ using QuickSweep All frequency band sliders were set to 0 dB a flat line This EQ has an amplitude range of 12 dB Digital vs Analog Equalization 248 of Disc Record Playback 2 Cantabile Solo 2 0 has a reasonable license fee and I think it s well worth it It comes with a pdf format User Guide but using it is largely intuitive Very powerful and will load several plug ins at the same time 3 Cubase LE 4 is a full featured DAW and is over kill for this project However this older version is very reasonably priced and it can be activated online for free 4 REAPER version 4 is also a full featured DAW it has a 398 page printed User Guide for an additional charge and a low license fee I used at least two host programs mostly Minihost and Cantabile for all plug in mea surements as a sanity check It s important to be able to repeat results After making measurements with several different plug ins it became apparent that the amplitude response could not be made to match the RIAA curve Fig 11 So I decided to try a flat line at OdB to find out how much of a match is possible Fig 12 shows the result of setting up this test on KarmaFX running in Cantabile A second computer is needed to supply
7. ORIGINAL ARTICLE RON TIPTON Digital vs Analog Equalization of Disc Record Playback Traditionally analog networks were used for frequency emphasis during recording of other than acoustic records It followed that analog networks be used for frequency de emphasis equalization during playback The emphasis and de emphasis networks complemented each other to produce an ideally flat frequency response Most currently produced phonograph pre amplifiers still use analog EQ However some individuals are saying that digital EQ is just as good but is it This study will hopefully help to answer this question Association of America de emphasis equalization EQ response In keeping with conventional usage I ll refer to it as the RIAA playback response But the compari son is also valid for any playback EQ that could be used for 78 or 45 rpm records I will look at four cases to do this comparison using the well defined RIAA Recording Industry 1 Commercial 2 3 octave band and 1 3 octave band analog equalizers 2 Acommercial 1 3 octave band digital equalizer 3 Several VST plug ins running in a VST host digital EQ 4 The DCS amp Virtual Phono Preamplifier and Parametric EQ digital EQ Baseline First I need an analog baseline for the comparisons The RIAA playback amplitude re sponse is defined by several time constants but these translate to amplitude frequency pairs from 20 Hz to 21 kHz or higher These values gr
8. RIAA has defined a recording and equalizers and other audio processors Product playback standard for vinyl records produced information is available from http www dod since 1954 A table of frequency vs gain or com attenuation values can be found in many places For example http www bonavolta ch 5 TrueRTA is a spectrum analysis program hobby en audio riaa html available from TrueAudio http www true audio com All testing in this article using 2 DC8is the name of an audio editor and re TrueRTA was done with the level 4 version cord restoration program from Diamond Cut Demo versions and a free version level 1 are Productions version 8 It is available from available from their web site http www diamondcut com and from http www tracertek com 6 This phase meter features automatic fre quency and input level ranging which greatly 3 Linear phase is defined as the property of simplifies making measurements This model a filter where the phase response is a linear is not in current production but pre owned function of frequency That is the phase re units may be found on ebay and from used sponse is a straight line when both phase and electronic test instrument distributors frequency are plotted on linear axes 248 10 11 12 13 14 15 ARSC Journal The Tenma 72 505 Audio Generator is avail able from MCM Electronics http www mc melectronics com The TDL model 4010 Phono pre amplifier
9. VPP and 4010 versions are very close and with different music choices I might slightly favor either one At times for example I thought that the 4010 had a somewhat warmer sound on the Pachelbel Paragraphic EQ in DC8 The Paragraphic EQ is in the Filters pull down menu It has several presets and one of them is RIAA Phono Equalizer Improved When this is selected we get the screen shown in Fig 16 The amplitude measurement procedure was the same as for the DC8 VPP with the RIAA comparison shown in Fig 17 The match is not very good so I did not bother to make time delay measurements Sony Sound Forge has a Paragraphic EQ but it has only four peak type filters each with a maximum bandwidth of 2 5 octaves so there is no way to match the RIAA curve Digital vs Analog Equalization 247 of Disc Record Playback Discussion This has been an interesting and sometimes frustrating project with the following conclu sions General purpose hardware analog EQ s are not useful for RIAA playback And at this time there seems to be only one digital RIAA EQ that s useful the Virtual Phono Preamplifier VPP in DCS My thanks to Don Walizer for being the primary music reviewer He listened to quite a few different DVD versions of Side 1 of Chet Atkin s Solo Flights RCA Victor LSP 3922 and Johann Pachelbel s Canon in D Major Musical Heritage Society MHS 1060 Others have recognized the importance of proper playback phase
10. a signal to the plug in s input In this case I used TrueRTA to generate both QuickSweep and pink noise The output is not a flat line and this shows why trying to match a smooth curve with QuickSweep or pink noise will not work A manual slow sweep using a continuously tuneable audio generator and output voltmeter produced the same result I tried this on several of the other VST EQ plug ins with much the same outcome In all the VST plug in tests the input and output voltage levels were kept well below the clipping level of both the software and the sound card The procedure for running the above tests in Cantabile is as follows Use a second computer with its sound card output connected to the Cantabile sound card input and launch TrueRTA Launch Cantabile and load an EQ plug in Click the Cantabile Record drop down menu to setup the recording choices Click the Cantabile Record button Click the TrueRTA QuickSweep button or choose pink noise and click GO I did find one exception The Algorithmix LinearPhase PEQ Red produced a nearly flat line with QuickSweep Fig 13 when all ten filters were set to OdB But this is a linear phase EQ so its phase response would be incorrect even with a perfect amplitude match Also its license fee of 999 Euros puts it out of my reach for experimentation Virtual Phono Preamplifier in DC8 This built in DCS feature is basically a re equalizer because it will accept an RIAA or flat input
11. aph to a smooth curve as shown in Fig 1 When an analog network is used the phase response Fig 2a as well as the time delay at each of the frequencies is also inherently defined Fig 2b which intentionally has a linear frequency axis shows the phase response is not linear so we can rule out EQ s with a linear phase response The RIAA time delay curve is not shown separately because it is included in each comparison graph For this project I built a low gain RIAA network Fig 3 shows the circuit diagram and its photo is in Photo 1 This network matches the defined amplitude response within 0 2 dB and has a correspondingly accu rate phase and time delay responses ARSC Journal XLIII ii 2012 Association for Recorded Sound Collections 2012 All rights reserved Printed in USA 286 Amplitude dB w o p 1 i 2 fo fo N a Phase angle degrees ARSC Journal 1000 Frequency Hz 10000 20000 1000 Frequency Hz 10000 20000 10000 Frequency Hz Figure 1 RIAA amplitude response measured on the baseline network The mea sured values differ from the defined values by a maxi mum of 0 2 dB Figure 2A Measured phase response of the baseline RIAA network log frequency axis Figure 2B Measured phase response of the baseline RIAA network with a l
12. d flat rolloff 20 Hz Rumble Filter and HF Filter set to 24 kHz I used the resulting files as the input to the VPP and recorded the output files which I sent to listeners for comments While waiting for comments I made amplitude and time delay measurements on the VPP I created an input file of 36 frequency bands from 20 Hz to 20 kHz the same frequencies produced by the Tenma audio generator for compatibility with previous measurements Each band is 20 seconds in length and separated by 4 seconds of silence from the next band and all bands have the same amplitude so it s a flat recording With the VPP set to RIAA output I recorded the equalized file Playing the output file into an RMS voltmeter produced the values used in the Fig 14 comparison graph The maximum deviation is 0 33 dB at 28 Hz The negative sign indicates the measured VPP values were lower than the calculated RIAA values The procedure for running this test is as follows Launch VPP in DC8 and setup the user parameters as mentioned above Tone Controls disabled Rumble Filter on Left Right Balance centered and Volume to maximum Open the source file in this case tones wav In DC8 click Record and setup the sample rate and bit resolution In the VPP window click Run Filter When the filter has finished running click Files in DC8 click Save Destination As and give the output file a name I then made a composite stereo file with the left track the 36 equal amplitu
13. data and user guide pdf format can be found at http www tdl tech com data4010 htm dbx Professional Products manufactures hardware graphic equalizers and other audio processing equipment Product information at http www dbxpro com Behringer manufacturers a wide range of audio products including analog and digital graphic equalizers http www behringer com Steinberg GmbH manufacturers audio hard ware and software http www steinberg com Minihost is a free VST plug in host with full ASIO support Although the download is free a donation is appreciated if you use it regu larly Information and download from http www tobybear de p_minihost html Cantabile is a VST plug in host professional and personal Solo versions from TopTen Software A free trial version is available and the Solo version registration fee is reasonable http www cantabilesoftware com Cubase is a product of Steinberg GmbH see number 11 above Reaper is a full featured Digital Audio Workstation DAW It performs as well or better than DAWs costing many times its registration fee http www reaper fm 16 17 18 19 20 21 KarmaFX is a 31 band VST EQ It s free but a donation is appreciated from frequent users Unlike many free programs it has a useful User s Manual Pdf Manual and download from http www karmafx net Many VST EQs are available online Some are free and some have a license fee usually
14. de fre quency bands and the right track the VPP output Using this file as the reference and signal inputs to the phase meter I recorded and graphed Fig 15 comparing the measured time delay to the RIAA delay Although not identical the curves are very close Even with a 96 kHz sampling rate there are some issues in making accurate time delay measurements at the higher audio frequencies only 4 8 samples per cycle at 20 kHz My primary reviewer commented as follows on the VPP equalizations I didn t have a clear preference with the Pachelbel but I did with the Atkins To my ears the VPP version had better instrument definition with clearer sound at all frequen cies The guitar sounded more like a real guitar while the 4010 version sounded slightly 246 ARSC Journal Figure 16 Screen shot of the DC8 Paragraphic EQ using pe 8208 i LS B ia the RIAA Phono Preamp 1200 H Improved preset a00 a S 400 __ 0 00 ___ 4 00 8 00 cu ___ 16 00 ____5 20 00 10 0 Hiss Amp Number of Filters Qctaves f eg oe OM 12 Eel 19820 E 3 5 o Ea RIAA Phono Equalization Curve Improved P Save Delete Reset Levels Figure 17 Comparison of the RIAA amplitude response to the DC8 Paragraphic EQ set to RIAA Improved Paragraphic EQ Amplitude dBu 1000 10000 20000 Frequency Hz veiled The
15. eful RIAA playback equalizer 1 3 Octave Band Analog EQ I used a dbx model 231S The 12 dB range of each slider does not cover the RIAA am plitude range so I used the two mono channels in series I imported the RIAA amplitude curve onto the TrueRTA Workbench and then using QuickSweep began adjusting the sliders to match the RIAA curve With 62 frequency sliders this took an hour or more Fig 7 shows the results the ripples in the dbx response pretty well correspond to the filter center frequencies The time delay comparison in Fig 8 shows a poorer match than the DOD EQ even though the frequency response match was as good if not better This supports the rule of thumb that the least amount of EQ to do the job is best 1 3 Octave Band Digital EQ A Behringer model DEQ1024 was pressed into service for this comparison I think the procedure has been sufficiently well described so I will just present the results The am plitude match Fig 9 is quite good in the 20 Hz to about 2500 Hz range With just one DEQ1024 I didn t have enough amplitude adjustment range to go to higher frequencies The time delay comparison Fig 10 shows no match at all This digital EQ is not at all suitable for trying to match the RIAA phase response Those who recommend using a phono pre amp with a flat response and then correcting the file with a digital EQ would not get good sound with the DEQ1024 VST EQ Plug Ins Virtual Studio Technology was de
16. inear frequency axis The RIAA phase response is not linear If it were the phase response would be a straight line see Reference 3 Digital vs Analog Equalization 237 of Disc Record Playback 2 3 Octave Band Analog EQ For this comparison I used a DOD model SR430Qx EQ with the two mono channels in series to achieve the nearly 40 dB RIAA amplitude range Using TrueRTA s QuickSweep I first captured the amplitude response of my baseline network and saved it to the TrueRTA Workbench Substituting the DOD EQ for the baseline network and using QuickSweep I adjusted the frequency band sliders to match the baseline curve This did take a while because I wanted a close match The result is shown in Fig 4 The match is not too good at 20 kHz because a 2 3 octave band EQ has its highest frequency adjustment at a center frequency of 16 kHz I also did the match using pink noise but QuickSweep resulted in smoother curves The next step was to measure the phase response of the DOD EQ For this I used a Dranetz model 305 phase meter mainframe with a 305 PA 3007 plug in and a Tenma model 72 505 audio generator The Tenma generator has 36 frequency steps from 20 Hz to 20 kHz and this simplifies repeated measurements because it s much faster than us ing a continuously tuned generator and frequency counter I recorded the phase readings with paper and pencil and then converted each value to time delay with a pocket calcu lator The time dela
17. itch to L R Rumble filter to 20 Hz and HF Filter to 24 kHz Where T is time delay in milliseconds 0 is the phase angle in degrees and fis the fre quency in Hz Please note that both axes Fig 5 are logarithmic but so is the amplitude re sponse in Fig 4 Although the amplitude response is linear in dB the Decibel is a loga rithmic unit I like the clarity of comparing time delays rather than phase but both are equally accurate The match in Fig 5 is not too bad so I decided on a listening test I played a selec tion of tracks from early LP records first with just a TDL model 4010 pre amp set to RIAA equalization and then with the 4010 set for flat equalization into the DOD EQ set to RIAA This setup is shown in Fig 6 The DOD EQ needs the gain of a phono pre amp to produce a useful output signal level The RIAA EQ network in the model 4010 is identical to the baseline network except the baseline network has much lower gain In addition to listening I recorded both EQ versions on the computer hard drive at 96 kHz sample rate and 24 bit depth so I could send a comparison DVD to other listeners who may have a higher musical acuity than I do Tracks 1 and 3 are from the DOD EQ and tracks 2 and 4 are from the model 4010 pre amp 240 ARSC Journal Figure 7 RIAA baseline and dbx231 EQ amplitude comparison The wavy line has more waves from this 31 band EQ than from the 15 band DOD EQ Amplitude dBu
18. veloped by Steinberg GmbH to add modular virtual instruments and effects such as EQ to their Cubase digital audio workstation DAW software By using a real time host a VST plug in such as an EQ becomes a virtual EQ with its inputs and outputs being the sound card inputs and outputs Measurements can be made on the virtual EQ just as if it were a hardware EQ There are many real time VST hosts available some free some rather pricey and some in between I have evaluated several with four useful products listed here 1 Minihost is donationware it s small in size and very easy to use Its disadvan tage is not being able to load more than one VST plug in at a time That is you can t load two instances of an EQ to achieve the 40 dB RIAA amplitude range 242 Time Delay mS gt 23 a D i v g a lt q Amplitude dBu ARSC Journal ao KR KR oa A D oO 1000 Frequency Hz 10000 20000 Se 1000 Frequency Hz 10000 20000 1000 Frequency Hz Figure 10 Comparison of the RIAA time delay toa Behringer DEQ1024 digital equalizer within the fre quency band where there is a good match between the amplitude responses Figure 11 RIAA baseline and Anwidasoft GEQ31V 31 band VST EQ amplitude response comparison using TrueRTA QuickSweep Two series instances of this EQ
19. y is just the phase angle converted to time at each frequency T 1000 360 f Table 1 RIAA Baseline Circuit Parts List Symbol Value Description Vendor Part No C1 C5 270nF 50V 5 V series polyester DigiKey P4668 C2 C6 15nF 50V 5 V series polyester DigiKey P4515 C3 C7 56nF 50V 5 V series polyester DigiKey P4522 C4 C8 22nF 50V 5 V series polyester DigiKey P4517 C9 C10 C11 C12 100nF 50V 5 V series polyester DigiKey P4525 C13 014 470 uF 16V 20 Radial electrolytic Mouser 647 UVR1E471MPD H1 H2 H3 3 pin header Mouser 538 22 03 2031 P1 P2 P3 3 pin terminal housing Mouser 538 22 01 2037 Molex terminals Mouser 538 08 50 0114 R1 R12 47K 1 4w 1 Metal film Mouser 271 47K RC R2 R3 R4 R5 R6 R8 R9 R11 R13 10K 1 4w 1 Metal film Mouser 271 10K RC R14 R15 R16 R17 R19 R20 R22 R7 R18 1100 1 4w1 Metal film Mouser 271 1100 RC R10 R21 1000 1 4w1 Metal film Mouser 271 1000 RC U1 TLC2274AIN Op amp DigiKey 296 7123 5 U2 TLC2272IP Op amp DigiKey 296 7117 5 Circuit board TDL M4041C PCB The C1 C2 parallel capacitance selected to 285nF 2 The C5 C6 parallel capacitance selected to 285nF 2 The C3 C4 parallel capacitance selected to 78nF 2 The C7 C8 parallel capacitance selected to 78nF 2 238 ARSC Journal a no i v 2 a lt
Download Pdf Manuals
Related Search